- Email: [email protected]
No-moving-part commutation of gas flows in generating plasma by cumulative detonations (survey)
Energy 157 (2018) 493e502
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
No-moving-part commutation of gas flows in generating plasma by cumulative detonations (survey) clav Tesar*, Jirí Sonský Va Department of Thermodynamics, Institute of Thermomechanics V.V.I., Czech Academy of Sciences, Dolejskova 5, 182 00 Prague, Czech Republic
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 October 2017 Received in revised form 26 April 2018 Accepted 24 May 2018 Available online 25 May 2018
This paper surveys controlling, without use of mechanical valves, of flows of three gases: a combustible mixture (fuel and oxidant), combustion products, and a neutral scavenging gas. The gases are delivered periodically in an alternating way into a pair of pulsed MHD electricity generators. Essential problem of this type of electricity generation is insufficient ionisation level of combustion products in the usual deflagration type burning. Solution was found in achieving momentarily extremely high temperatures by cumulative detonations. In the 1st phase the cavities leading into the 1st MHD generator are filled with the combustible mixture, which is then ignited, generating a high-temperature high-speed detonation wave moving into the generator. In the subsequent 2nd phase the combustion product are scavenged in the cavities by the neutral gas. In the other generator the phases are in reversed order. Because of the extreme temperature, the flows are switched not by mechanical valves but by no-moving-part devices, in various configuration that are systematically surveyed. © 2018 Elsevier Ltd. All rights reserved.
Contents 1. 2.
3.
4.
5.
6.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494 Generating plasma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495 2.1. Cumulative effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495 2.2. Single-shot model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495 Control of gas flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496 3.1. Two-phase operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496 3.2. Pulsation frequency and size of detonation tube . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496 Commutation by fluidic diverter valves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496 4.1. Jet-deflection amplification effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496 4.2. Coanda-effect bistable diverters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497 4.3. Commutator with Coanda-effect valves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 498 4.4. Fluidic commutator operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 498 Alternative commutator configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499 5.1. Fluidic oscillators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499 5.2. Autonomous commutator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500 5.3. “Master & Slave” valve configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500 5.4. Use of vortex diodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501 5.5. A faster configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502
* Corresponding author. E-mail address: [email protected] (V. Tesar). https://doi.org/10.1016/j.energy.2018.05.165 0360-5442/© 2018 Elsevier Ltd. All rights reserved.
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
494
1. Introduction Despite the contemporary large progress in renewable electricity production, the largest percentage of the generation processes still relies on burning various fuels. The generation consists of conversion steps with the typical sequence as follows: (1) (2) (3) (4) (5)
fuel combustion, steam generation in a boiler, accelerating the steam in turbine, generating mechanical power on turbine shaft, and electric current generation in induction coils.
Each of these steps is inevitably associated with some energy loss - so that, together with the Carnot cycle limit, the overall efficiency is low. Typically it is not above 35e40%. There is a generally known but practically not used alternative d the magnetohydrodynamic (MHD) generators. They avoid the above processes (2), (3), and (4), thus offering in principle a substantially better conversion efficiency. The original idea is described in Ref. [9] and the present state of art is surveyed in Ref. [11]. The principle requires heating the gas to temperature levels so high that the it is ionised and thus becomes electrically charged. Unfortunately, those few MHD facilities so far built for feasibility tests [11], all fail to reach the necessary ionisation temperatures. Various attempts at efficiency improvement, such as supplying pure oxygen or alkali metal particles to the combustion process, are highly un-economical. A substantially higher temperatures than now achievable - and consequently much higher degree of gas ionisation - may be reached by replacing the standard deflagration combustion by detonations. It is an idea that has not been so far followed because of the danger associated with very high mechanical stresses of the generator components. Moreover, even the degrees of ionisation achieved in the investigated detonations so far remain below the necessary level. A solution is offered in Ref. [17]. The basic idea is the cumulative implosion [4,14] inside the combustion chamber, Fig. 1. The temperature in the focal point of the colliding detonation waves is extraordinarily high, as is demonstrated by the military applications in the hollow charge weapons d and yet the stress levels in the chamber walls are not particularly high because of mutual cancellation of oppositely acting forces in the focal point [18]. The military hollow charge concept cannot be simply taken over for electricity generation. It is suitable for a single detonation of solid
Fig. 1. Schematic representation of the plasma generation by implosion detonations in the combustion chamber. Generated in the tube at left, the detonation wave is distributed through the large number of equal-length channels into wavelets all colliding mutually in the focal point.
explosives d while for the use in a MHD generator the detonations are to be in the gaseous phase and periodically repeated. This requires a rather complex system (seen schematically represented in Fig. 1) ensuring the synchronisation of colliding detonation waves coming towards the focus from all directions e and also scavenging the combustion products from the cavities after each detonation followed by re-filling by a fresh combustible mixture. To achieve the mutual cancellation of the opposing forces, the implosive wave inside the combustion chamber has to be generated in a spherical configuration very accurately e apart from the exit into the MHD generator. The necessary very small deviations from the sphericity are practically impossible to achieve with a large number of independent detonation sources on the outer wall of the combustion chamber. There are two reasons: (a) the extremely high speed of detonation waves propagation (always much faster than local speed of sound) and (b) uncertainties associated with the initiation of detonation waves. The latter fact is due to the character of deflagration-to-detonation transition. It depends on turbulence, which is a stochastic phenomenon with substantial uncertainty. Experience has shown that the required accuracy of the spherical symmetry is practically achievable only with a single detonation source e the detonation tube in Fig. 1 - with its output distributed into the large number of parallel equal-length channels. As a result, the necessary geometry of the cavities is rather complex. The removal by neutral gas of the combustion products from these complex cavities and their replacement by a fresh combustible mixture is achieved by gas flows - combustible mixture and neutral gas e each available from an external pressurised source. The order of their flows and their directing into the combustion chamber must be controlled by a commutator system. Because of the high temperature environment, mechanical commutator valves with moving components are out of question, especially due to preferable operation at the standard 50 Hz repetition frequency (Fig. 4) of the generated alternating electric current. The solution discussed in this paper is based on commutation by no-movingpart devices, made of a refractory material. 2. Generating plasma 2.1. Cumulative effect The cumulative effect of imploding detonation waves is capable of generating a plasma flow with high degree of ionisation at very high, hypersonic velocities. Both properties are desirable in the MHD generator. The high velocity is achieved by the axial imbalance of detonation waves, with the distributor channels missing on the right-hand side in Fig. 1, where there is the entry into the MHD generator. Existing literature discusses Mach numbers of the detonation wave propagation of the order Ma z 4e8 [14,16,24].
Fig. 2. Principle of generating the alternating current in a pair of MDH generators. Detonations are controlled by two synchronised diverter valves. The 1st detonation tube is here shown filled by combustible mixture through the node A while the 2nd tube is scavenged by air flow delivered through B.
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
Fig. 3. The other phase of the gas flows in the same circuit as in Fig. 2. While the 1st detonation tube is scavenged by air after the detonation, the 2nd tube is re-filled with fresh combustible gas mixture.
Fig. 4. Detonations in the pair of mutually phase-shifted MHD generators. The output from the 2nd generator is connected to generate the negative part of the AC period.
Also the high temperatures, measured spectroscopically (evaluated from changes in spectra lines of emitted light [26]) was already demonstrated in present authors' reference [18]. The implosive phenomenon is in literature known as Munroe or Neumann effect (although the priority of the basic idea is now known to belong to von Baader [2] who suggested e but failed to demonstrate in practice - the use of the effect in mining). The term “Munroe effect” - common in the USA - is named after the author of reference [12], who in 1883e1886 performed in US naval torpedo laboratory a series of demonstrations of solid-state hollow charges for armour penetration. 2.2. Single-shot model Correctness of the present authors' assumptions about the achievable performance parameters were demonstrated by performing experiments using a laboratory model, described in Ref. [18]. Because of the feasibility study purposes, it was operated in the regime of single shots, with long time periods between the detonations. The combustion chamber internal diameter was 95 mm. The combustible mixture used was the stoichiometric mixture of hydrogen and oxygen, produced prior to the experiments by water electrolysis and accumulated. The product gas of this combustion is steam in various contents of ions. It has a tendency to condense as water drops on the less hot locations of the internal cavity walls. The removal of the drops was one of the reasons for the waiting time between the individual shots d the air blowing was left to act for several minutes to ensure the cavities were dry. The detonation wave was generated in detonation tube of constant 8 mm 8 mm square internal cross section (which means a cross-section area 64 106 m2) and 180 mm long. After the drying of
495
the cavities and then re-filling them with the O2þ H2 mixture, the detonation was started by the electric high-voltage spark (using for the purpose a standard internal combustion engine type spark plug) near the entrance into the detonation tube. In the initial stages of the combustion process, the wave that progressed along the tube length from the ignition was in the ordinary mode of deflagration burning. Its transition into the detonation was caused by the turbuliser [7], a local configuration in the tube generating turbulent vortices. Instead of the popular Shchelkin's spiral [15] used as the turbuliser by a number of other researchers [5,8], in the authors' test model the turbulence was generated by a row of local tube excrescences [6]. As is apparent in Fig. 1, the part of the model most difficult to make was the distributor d a system of channels converting the planar detonation wave from the detonation tube into a large number of confocal wavelets. The conversion is performed by the channels increasing their number in the flow direction by progressive bifurcations. There were altogether 168 channels with exits directed towards the focal point inside the combustion chamber. The channels were all of the same 6 mm 6 mm cross sections (i.e. with total exits cross-section area 0.36 106 m2) and the same length 72.2 mm. For the design of the scavenging and re-filling systems are important factors the internal volumes in the cavities. The spherical combustion chamber volume was 450 106 m3, the volume of the detonation tube was 11.5 106 m3 and the total volume of the channels in the distributor was 440 106 m3 (almost the same as the combustion chamber). Considering the initial trifurcation at the inlet into the distributor, the total volume of all cavities was
Vz1:0 103 m3
(1)
In the experiments, for the scavenging was used a low-power blower, delivering approximate volume flow rate 83 106 m3/s. This means the minimum time between the shots would be 2 12 ¼ 24 s. Actual times in the tests (because of the need to dry the surfaces) was much longer, of the order of minutes. Of course, there was no intention to run the model in a periodic regime. The most important aspects of the tests were the achieved values of implosion temperature and gas velocity at the exit from the combustion chamber. The data obtained were quite promising. Velocity of the plasma leaving the combustion chamber was determined by analysis of high-speed camera images to be approximately 2.7 km/s. This corresponds to hypersonic Mach number Ma z 5.6 [13]. Temperature measurement was performed by analysis of hydrogen lines in the spectrum of emitted light [26]. The results were between 10 000 K and 15 000 K (may be compared with the temperature 6000 K of solar surface). 3. Control of gas flows 3.1. Two-phase operation Because of the requirements of standard electricity consumption, the ideal pulsed MHD electricity generator would operate at pulse repetition frequency 50 Hz. Simplified calculation of the ideal scavenging and re-filling times Dts and Dtf respectively, shows very short durations
Dts ¼ Dtf ¼ ½:1=50 ¼ 0:01s
(2)
This would necessitate choosing a very small size of both combustion chamber and generator. It is inconvenient from many points of view. Avoiding the very short times is possible with a multi-phase configuration. This, of course, means to have a number of MHD sequentially operating generators and other components
496
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
d which tends to be too expensive. As a reasonable compromise is seen two-phase operation, with two alternatively fired generators, as shown in Figs. 2e4. The repetition frequency of the detonation pulses is thus halved, ideally to 25 Hz. The scavenging and re-filling times are then
Dts ¼ Dtf ¼ ½:1=25 ¼ 0:02 s
(3)
It is a quite small but acceptable value. Of course, it means a fast operation of the valves used to control the gases flows. With mechanical valves it would be rather difficult to move so fast the solid components opening and closing the flowpaths into the nodes A and B. (Fig. 2). The critical factor is the inertia of these components. Also, the mechanical valves are unlikely to endure for a reasonable lifetime the high temperature environment and the detonation stresses. This paper suggests a different approach: applying the idea of no-moving-part fluidic commutator from Ref. [19]. Instead of opening and closing the entry into the gas flow pipes, this design operates by diverting the flows. Two jets generated in a pair of nozzles are alternatively captured and led into one of the two exits and the two combustion chambers connected to them.
3.2. Pulsation frequency and size of detonation tube Although the demonstration laboratory model described in Sect. 2.2 was not intended for actual AC electricity production, it is useful to estimate how it would behave in a periodic operation regime. It is assumed that there are, under pressure, external sources of the two commutated gases d the combustible gas mixture and air used for the scavenging. The flows of these gases have to from the detonation cavity its whole contents and to re-fill it. According to the equations (1) and (3) the volume flow rates needed for the task are
. . V_ ¼ V=Dts ¼ 1:0 103 0:02 ¼ 0:50 103 m3 s
4. Commutation by fluidic diverter valves 4.1. Jet-deflection amplification effect The basic components of the now discussed fluidic commutator is the synchronised pair of fast acting flow diverter valves, schematically represented at the left side of Figs. 2 and 3. The fluidic commutator as suggested in Ref. [17] consists of the two jet deflection devices, based on the ideas shown in Figs. 5a, 5b, 6 and 7. Solid and not moving, the fluidic valves made of a refractory material can withstand the adverse high-temperature conditions for very long periods of time without significant erosion. Their task is to direct the supplied gas alternatively between the nodes A and B (Figs. 2 and 3) which lead to the two detonation tubes. Fluidic valves have also already demonstrated the capability to operate at high switching frequencies. For example, ref. [22] has shown a fluidic jet-type valve oscillating at the supplied air flow rate V_ ¼
0.324 103 m3/s d which is merely 1.5-times less than the value in eq. (4) and may be easily increased d at the frequency f ¼ 8230 Hz. This is a value 330-times higher than what is needed in the commutator. The operation cycle of the commutator consists of the two phases shown in Figs. 2 and 3. At the end of each phase is generated the detonation wave. In the first phase, presented schematically in Fig. 2, the 1st detonation cavities (on top of the picture) are filled with the combustible gas mixture through the node A d while the other, bottom diverter in this picture despatches into the node B the scavenging gas. This makes the 2nd (bottom) cavities prepared for the next filling by the combustible mixture. The first phase finishes by ignition of the mixture in the 1st detonation tube and the subsequent (extremely brief) detonation. The detonation wave progresses through 1st detonation tube into the combustion chamber and from there to the MHD electricity generator. The fast moving
(4)
This is a large value. In the test model the smallest cross section was 8 mm 8 mm of the detonation tube, of cross section area F ¼ 64 106 m2. This means that the flow of the air in the detonation tube of the original size move during the scavenging through the tube at the velocity 780 m/s, i.e. at the supersonic Mach number Ma z 2.4. This is out of question. The situation may be improved upon by choosing a larger detonation tube cross section F. A reasonable velocity value with substantially lower driving power would be obtained by selecting approximately 10-times larger area 25 mm 25 mm of the detonation tube cross section. The value is, of course, the reasonable Ma z 0.24.
Fig. 6. Bistable Coanda-effect attachments alternatively to one of the two walls. Switching between the two regimes is achieved by short flow pulses applied to the control nozzle.
Fig. 5. (a). The most popular among the no-moving-part fluidic flow control principles is jet deflection of the powerful main jet by small control flow. (b). Coanda effect: wall presence l prevents the jet-pumping entrainment on the top side of the jet creating the transverse pressure difference between its sides. This deflects the jet towards the wall and keeps it attached.
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
Fig. 7. Typical bistable fluidic diverter valve and its schematic representation.
flow of the charge carriers passing through the induction coil generates in the MHD generator segmented electrodes exposed to the flow an electric current pulse. Immediately thereafter begins the second phase, presented schematically in Fig. 3. The top diverter valve directs the combustible gas mixture into the node B while the other, bottom diverter despatches the scavenging gas into the other node A. This results in removing from the 1st detonation tube e as well as from the combustion chamber connected to it e the combustion products. At the same time, the 2nd detonation tube and the connected combustion chamber d after their being scavenged in the previous first phase d are now, as is shown in Fig. 3, filled with the combustible mixture. Again, when the cavities are full, this filling ends by ignition and generation of detonation wave. It progresses through the mixture to the combustion chamber and to the MHD generator. There the inversely connected MHD coil generates in the common exit terminal the negative current pulse - symbolically shown at right in Fig. 3. As a result, the two pulses (positive and then negative) represent a full cycle of an AC output current. Its waveshapes are not likely to be harmonic, but this is easy to improve d in the simplest way it may be done by a passive LC filter and a transformer.
497
nozzle (in A of Fig. 5a) to generate a control flow. By the momentum interaction between the main jet and the control flow the main jet is deflected (upwards in Fig. 5a). Opposite to the exit from the main nozzle may be then placed a collector capturing the main jet and leading it into the output terminal. Magnitude of the output flow rate captured in this terminal thus depends on the deflection angle and this, in turn, on the magnitude of the control flow rate. The device operates as a fluidic amplifier. Early designers of the fluidic jet-deflection amplifiers were surprised by encountering a tendency of the jet to remain deflected d as shown in B of Fig. 4 d even if the corresponding control flow was absent. The phenomenon is known as the Coanda effect of jet attachment to a wall. It is caused by the jet entraining the surrounding gas and carrying it away. If there is a nearby solid wall (B in Fig. 4), the entrainment on that side is prevented. This develops between the jet and the wall a very low pressure - resulting in a side force that deflects the jet. This jet attachment phenomenon, initially considered a nuisance, later became very useful d especially in the symmetric configuration with two attachment walls, shown in Fig. 5. There is no continuous jet deflection control e instead there are only two stable regimes. In the discussed application, however, this is exacly what is needed. The configuration of the valve, as shown in Fig. 7, is planar, with everywhere constant depth h (Fig. 6) of the cavities. The jet may be switched between the two attachment walls by short flow pulses applied to the control nozzle on the wall attachment side (the top nozzle inactive in Fig. 5b). Once the jet is switched, it remains deflected and the control flow is no more needed. This results in saving of the control power. The bistable configuration is symmetric and provided with the two collectors, as shown in Fig. 7. A continuation of the flowpath from the collectors is on each side of the valve a diffuser. Its task is the opposite to the action in a nozzle: it converts kinetic energy of the jet in the pressure energy (note that the hydraulic losses depend on the kinetic energy). Also note that the apex angle of a diffuser has to be quite small to avoid flow separation from its walls and consequent loss of efficiency. The small apex angle results in long diffusers, which tend to dominate the whole valve body in Fig. 7. In the bottom part of Fig. 7 is a schematic representation of the valve as it is used in circuit diagrams. The nozzles there are represented by black triangle symbol while the diffusers are shown as white (i.e. not filled) triangles. Also schematically represented in Fig. 7 is the pair of attachment walls.
4.3. Commutator with Coanda-effect valves Shown in two identical circuit diagrams presented in Figs. 8 and 9 are the two phases of the cycle in the commutator according to [19]. In particular, drawn by the thick curved grey line, is indicated
4.2. Coanda-effect bistable diverters The pair of fluidic gas flow diverting valves described as the essential parts of the commutator in Ref. [19] operate on the principle of deflecting a gas jet leaving a nozzle, shown in Fig. 5a and 5b. The jet is formed from the supplied gas flow in the main nozzle (at left in the pictures). A jet-pumping action e entrainment of the surrounding gas - generates near the nozzle exit a region of low pressure. This makes easy for perpendicularly oriented control
Fig. 8. Circuit diagram of the fluidic commutator according to [19]. After the control pulse a had switched the combustible gas mixture flow into the node A, the 1st detonation tube is re-filled by the combustible gas mixture.
498
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
4.4. Fluidic commutator operation
Fig. 9. The other phase of the fluidic commutator from Fig. 8. After the detonation and the switching pulse b it is the 1st tube that is scavenged e while combustible mixture fills the 2nd detonation tube.
in both pictures the character of the flow of combustible gas mixture. The role of the two schematic flow diverters that was shown above as a block diagrams in Figs. 2 and 3 is here performed by the two jet-deflection Coanda effect valves e both corresponding to the Fig. 7. Both pictures Figs. 8 and 9 show also the control lines that transfer into the valves the flow-switching pulses a (in Fig. 8) and b (in Fig. 9). The anti-parallel operation of the valves is as follows: if one valve delivers the combustible gas through the node A into the 1st detonation tube (Fig. 8), the other valve simultaneously directs the scavenging air through the node B into the 2nd tube d and vice versa. The commutator alternates between two phases, the PHASE 1 and PHASE 2 which together form the operation cycle. The corresponding time duration diagrams of the gas flows through the node A are presented on the top of Fig. 10. The spatial character of the flow drawn by thick grey curved line in Fig. 8 occurs after arrival of input flow pulse a e and lasts then until the arrival of the pulse b. This latter pulse comes at a the instant of time (Fig. 10) when the 1st detonation tube and cavities connected to it are already re-filled with the combustible mixture d and, in fact, when there is already some inflow of protective neutral gas. This suppresses occurrence of a return motion of detonation wave in the tube. This difference between the two phases results in the slight asymmetry of the two phases in Fig. 10. After the detonation, which is started by the electric ignition pulse (Fig. 10), the corresponding detonation tube is cleaned and the combustion products are removed being displaced by air. In the PHASE 2 the combustible gas mixture is switched e as shown by the thick grey line in Fig. 9 to fill the 2nd detonation tube through the node B.
Because of the geometric symmetry of both phase flowpaths, it is sufficient for explaining the operation of the jet-deflection commutator to describe only the processes by which is serviced through the node A - the 1st detonation tube. The time dependences of flow and switching control pulses are presented in Fig. 10 while the spatial aspects are presented in Fig. 11 for PHASE 1 and in the next Fig. 12 for the PHASE 2. The operation cycle begins by arrival of the fluid flow pulse a (Fig. 10) into the upper control nozzle (Fig. 11) of the diverter D1, which switches the air flow away from the node A and e since the pulse is simultaneously admitted into the lower control nozzle of the other diverter D2 e directing into A the combustible mixture O2þ H2. From D2 the mixture continues into the 1st detonation tube (which was in the previous cycle scavenged). The filling of the 1st tube and also of all the channels leading from it into the combustion chamber lasts for the whole PHASE 1. This filling of cavities stops after arrival of the pulse b (Figs. 10 and 12) which determines the beginning of PHASE 2. The diverter D1 now admits into the node A the scavenging air which progresses from there into the inlet of the 1st detonation tube. The combustible mixture in this tube is pushed forward to the combustion chamber and upstream from the spark plug is formed the non-combustible air “cushion”. This is the exact time for the combustion started by the spark. Note the position of the electric impulse at the bottom of Fig. 10. No combustion can take place in the “cushion” because it now contains only non-combustible air and no fuel. There is therefore no retrograde flow into the left-hand inlet part of the 1st detonation tube. The combustion progresses in the detonation tube towards the combustion chamber, but very soon it reaches the turbuliser (cf. also Fig. 1). This converts the deflagration into the detonation. This reaches the very high supersonic velocity and progresses through the rest of the detonation tube as well as through all the other cavities that were previously filled with the O2þH2 mixture. This all happens at the small initial fraction of PHASE 2. By the cumulative effect the colliding waves [1,3,10,27] generate ionisation in the combustion chamber and directs the resultant charge carriers into the 1st MHD generator. Clean air continues flowing into the 1st detonation tube for the rest of the PHASE 2 removing the combustion products that were left there by the detonation wave. This scavenging action ends by the arrival of the flow pulse a, which starts the next cycle.
5. Alternative commutator configurations 5.1. Fluidic oscillators Apart from the basic version of the fluidic commutator described in Ref. [19], there are several other alternatives. Those discussed here retain the jet-deflection type of the fluidic diverter valves - but with some changes in the circuits. Especially interesting
Fig. 10. Time dependence of flows in the node A (and in the 1st detonation tube connected to it) of the commutator in Fig. 8. Also shown is the timing of the control pulses a (Fig. 8) and b (Fig. 9) and the ignition.
Fig. 11. Schematic representation of the filling process in the 1st detonation tube by the fluidic switching circuit shown in Fig. 8.
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
499
Fig. 12. Schematic representation of the air flow in the commutator cavities at the instant of igniting the combustible gas mixture.
are configurations employing the amplification capability of the fluidic jet-deflection valves. This property may be used to generate self-excited periodic oscillation. The advantage thus gained is the autonomous character of the commutator, not needing the external switching pulses from external source. The commutator becomes a self-contained unit consisting of nothing more than cavities made inside a solid body. The control by external flow pulses - as shown in Figs. 10e12 above e usually assumes a timing in an electric controller and an electro/fluidic (E/F) transducer [24] converting the electric pulses into fluidic ones. These devices have to be located near the detonation tubes and combustion chambers. Considering the extremely high temperatures inside the combustion chamber and the quite short distances from it, it is necessary to expect elevated temperatures also in the whole commutator. This brings no problem for the fluidic commutator itself e the thermal robustness is one of the reasons for its choice, as it is nothing more than cavities in a solid material of refractory character. Unfortunately, similar resilience to heat without complications of added cooling to heat cannot be expected in usual electronics and transducers. The generation also of the switching signals by fluidics is therefore an interesting possibility d the more so because the fluidic pulse generator may be very simple and in fact necessitates just an addition of a feedback flow channels. An example of fluidic oscillator is presented in Fig. 13. Similar in principle to the feedback-type oscillators in electronics, typical fluidic oscillators also consist of two components: an amplifier and a feedback loop. It should be noted that apart from the configuration from Foig. 13 there are many other oscillator circuits, as reviewed in Ref. [20]. The important fact is the role of the fluidic amplifier in the oscillator may be played by one of the diverter valves in the commutator e simultaneously with its flow switching role. All that is necessary is to connect by the feedback loop channel the two control terminals X1 and X2 (Fig. 13). Length of the connecting loop - which may be also shaped as a cavity in the refractory body defines the oscillation frequency. The switching of the jet inside the valve from one attachment wall to the opposite one is caused by the flow in the feedback loop and the control nozzles. The flow is driven by the pressure difference DPX (Fig. 13) e the difference between the two sides of deflected jet. The pressure in X1, at the side to which the jet is attached, is lower than the pressure on the opposite side, in the location X2. The fluid in the loop thus flows under the action of this pressure difference from X2 to X1. Because of the amplification capability of the fluidic valve (i.e. its ability to switch the powerful main jet by a small flow issuing from the control nozzle) the flow in the loop suffices for switching the deflected jet. After this switching to the other side, the situation e direction of the pressure difference force - is reversed. This reversal takes a certain time and this represents the phase shift necessary for generating the oscillation. Appearance of one example of the robust commutator with the integral oscillator,
Fig. 13. Fluidic oscillator set up from the diverter by added feedback loop in which the flow is driven by the pressure difference DPX between the control terminals. Oscillation frequency is adjusted by varying the loop length [20].
insensitive to high temperature, is presented in the next Fig. 14. The cavities shown there were made by laser cutting in the flat plate of ceramic material. From top are the cavities closed by another flat plate (not shown) of the same material e in a similar manner as is done on the bottom side. The cavities actually consist of a whole stack of plates. The lines in Fig. 14 indicate the connecting (much wider) cavities made in some of these plates. 5.2. Autonomous commutator The circuit diagram of a commutator with integral generator of switching signals, is presented in Fig. 15. There is a pair of switched diverter valves, D1 (handling air) and D2 (handling the combustible mixture). They periodically vary their diverted gas flows into the 1st and 2nd detonation tubes. The valve D1 has the feedback loop
Fig. 14. An example of the commutator cavities made by the refractory material removal. Also shown are idealised channel connections ensuring both feedback and mutual synchronisation of the diverters.
500
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
connecting, like the oscillator in Fig. 13, its control nozzles. The alternating air flows from D1, apart from their ordinary role of the scavenging of the internal spaces of the commutator, also deliver the air into the control nozzles of the other valve D2. This valve is periodically switched by the air flows through the control inlets a and b. The next two illustrations, Figs. 16 and 17, explain the internal processes in the commutator from Fig. 15. In the Fig. 16 is shown by the superposed wide grey colour lines, what happens in this commutator the PHASE 1. The next Fig. 17 then presents the flow phenomena occurring in the other PHASE 2. The air flow from the output of the oscillator in Fig. 16 is seen divided into two branches. The more substantial flow enters the node A and continues into the 1st detonation tube. It is shown arriving there at the instant of the air cushion formation. The smaller part of the air flow is branched into a and there it controls the switching of the gas mixture. It deflects the mixture jet and directs it into the node B, continuing from there into the 2nd (bottom) detonation tube. When the air flow in the feedback loop of the oscillator is reversed, it generates an outflow from the other control nozzle, as shown in Fig. 17. It is seen directing the main air flow into the node B. This air flow leaving the oscillator also bifurcates. The much more powerful air flow gets into the 2nd detonation tube and scavenges it. The smaller part of the flow from the bifurcation passes through b. It enters the gas mixture valve and diverts its flow into the 1st tube. For switching the valves would suffice in the control inlets a and b only very short flow pulses. This, however, does not happen in the circuit shown in Figs. 15e17. The control flows there last for the whole duration of each of the respective phases. In some commutator designs this causes no problems e especially if the scavenging role is played by air, the oxygen contents of which may be used in the combustion process. Nevertheless, the resultant combustion is
Fig. 17. Gas flows after the detonation in the 2nd phase d in the circuit presented above in Fig. 15.
likely to be less efficient than burning the pure oxygen. There may be circumstances that require for the scavenging a gas wholly neutral, perhaps pure nitrogen. Its inflow through the control nozzle for the whole duration of each phase may make the combustion process less efficient due to the dilution of the combustible mixture. It is useful to consider present alternatives suppressing this dilution effect. First of all, it is obvious that the dilution is the less intensive the higher is the flow gain of the gas-mixture diverter valve. Higher flow gain (i.e. the ratio of the main nozzle flow to the switching control flow) means less scavenging gas needed for switching the gas mixture flow rate. 5.3. “Master & Slave” valve configurations Intensity of the dilution of the combustion gas mixture, caused by the prolonged control flow into the diverter handling the combustion gas mixture may be reduced substantially by the circuit layout presented in Fig. 18. The layout [21] is very similar to the one shown in Fig. 15. The difference is a single valve of Fig. 15 is in Fig. 18 replaced by an amplifying cascade. The overall flow gain value G of this cascade is the product of the two gain values G1 and G2 of the two diverter valves. It is thus possible to obtain the desirable small dilution. 5.4. Use of vortex diodes
Fig. 15. An example of fluidic autonomous commutator with one air flow diverter (here the diverter D1) operating as the generator of control pulses.
Fig. 16. Gas flows in the circuit presented in Fig. 15. The flow lines suggests the conditions just prior to spark ignition in the 1st detonation tube. Note the air forming the non-combusting “cushion” upstream from the spark plug in this tube.
An alternative commutator design with autonomous selfexcited oscillation e but suppressed dilution of the combustible gas mixture - is presented in Fig. 19. Its basic idea is use of dynamic properties of vortex diodes d their capability to shorten the pulse
Fig. 18. Reducing the dilution of the combustible mixture by the scavenging gas may be achieved as shown here using the “Master & Slave” flow-diverting circuit discussed in Ref. [21].
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
duration to mere fraction of the half-period. The circuit diagram of this configuration is presented in Fig. 19. It is very similar to Fig. 15, the only difference is the addition of the two diodes into the control flow entrances a and b. Both diodes are oriented in the highdissipance direction, i.e. with the internal air rotation suppressing the flows to the control nozzles. The property employed in the present case is the growth of dissipance (resistance to the flow) after the beginning of the input flow pulse. This shapes the transmitted flow pulses, as was the subject of ref. [25] (differing from the configuration shown here by an unimportant detail) or the parts of the oscillator based on the same principle [23]. The behaviour of fluidic diodes is explained in Figs. 20e22. In literature are described fluidic diodes of quite complex configuration, but for the explanation purposes it is here useful to consider a quite simple example of diodes consists of a shallow circular vortex chamber, tangentially oriented inlet, and central outlet. The depth of the chamber is small so that if the inflow (indicated by the wide white arrow at the top of the picture) stops the motion inside the chamber is rapidly stopped by friction on the bottom surface - as well as the top surface which in these pictures is removed. In Fig. 22 there are two diagrams of the flow rate dependences on time. Let us consider the beginning of the cycle in which the oscillator in Fig. 19 has just switched the air flow into the node A and to the control inlet a. Prior to this incoming flow pulse both flow rates are zero. There is no air rotation inside the vortex chamber of the diode d1 because it was stopped by the friction on the top and bottom surface. The air admitted into the tangential nozzle does not meet any large opposition and passes through the diode rather easily. It has only to overcome in the leading part of the control-flow pulse to overcome the inertia of the accelerated air. The character of this easy flow is indicated in Fig. 20. Because of the tangential orientation of the nozzle in the diode, however, the fluid inside the chamber begins to rotate as shown in Fig. 21. There is the general tendency of this rotation to keep more or less constant the moment of fluid momentum. As the radial distance r of the gas decreases e since it has to leave through the central exit d this conservation tendency leads to increasing the tangential velocity w. As the time progresses, this rotation velocity actually rapidly grows and soon becomes so high that the centrifugal force acting on the gas almost blocks the flow. The output flow rate diagram in Fig. 22 shows what is practically an end of the pulse in the diode exit location Y (only very small air flow manages to pass through the boundary layers). The diode thus effectively shortens the duration of the control flow pulse. It lasts only during the period of attaining the full rotation speed. During the rest of the phase the flow rate through the diode and through the commutator valve control nozzle - is negligible.
501
Fig. 20. (Left) Vortex diode at the beginning of the flow through the vortex chamber. The fluid passes to the central exit without any significant hydraulic resistance.
Fig. 21. (Right) Later the fluid in the vortex chamber becomes spinning. The tangential velocity w increases rapidly with diminishing flow radius r. The flow through the diode is now effectively opposed by the centrifugal force.
5.5. A faster configuration The generation of the plasma pulses at the repetition frequency corresponding to the AC electricity 50 Hz would simplify the uses of the electricity generator. The problem is the scavenging and refilling times in the cavity may tend to be rather long. This calls for increasing the scavenging and re-filling flow velocities, but this is not a really good idea as it would mean higher pressure to be overcome e and may even encounter the Mach number compressibility limitation. Much better approach is the shortening of the flowpaths. The detonation tube may be made shorter e and as shown in Fig. 23, the gases may enter the cavities roughly in the location at the middle of distance that has to be travelled.
6. Conclusions
Fig. 19. An alternative possibility how to suppress the combustible gas mixture dilution in the commutator by using the dynamic properties of vortex diodes [22,23].
The capability of achieving the ionisation temperatures and also the high detonation-wave velocities in an implosion-type combustor was already earlier demonstrated in a single-shot laboratory model. The next stage, discussed in this paper, concerns the
502
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
in the internal cavities. There are alternative configurations, in the paper surveyed and discussed. Acknowledgements Authors' research was supported by institutional support research RVO:61388998 and Czech Republic Grant Agency GACR grant 17/08218S. References
Fig. 22. Graphical presentation of vortex diode response to the pulse of the input flow _ The initial part of the pulse passes quite easily from X to Y d but once the gas rate M X _ Y is opposed and rotation speed inside the vortex chamber increases, the diode flow M practically stops.
Fig. 23. Flows in the commutator may be not fast enough for the desirable 50 Hz frequency of pulses. Faster scavenging and re-filling may be obtained by the shorter travelled distances iby the commutator (here shown with monostable switching valves) delivering the gases into the middle location, between the detonation tube and the combustion chamber.
auxiliary facility necessary for achieving the reciprocating operation of the plasma MHD generation d the delivering alternatively the combustible gas mixture and a scavenging gas into and away from the combustion chamber. The configuration considered is the two-phase version, with two systems (combustion chamber and generator) operating in anti-parallel. The gas flow commutator [19] for handling the gases is the no-moving-part fluidic unit e effectively a solid body of refractory material with the actions performed
[1] Axford RA, Holm DD. Converging finite-strength shocks. Phys Nonlinear Phenom 1981;2:194. [2] v. Baader FX. An attempt at a theory of working with explosives. Bergmannisches J 1792;1:193. [3] Bilbao LE, Gratton J. Spherical and cylindrical convergent shocks. Il Nuovo Cimento 1996;18 D:1041. [4] Bolonkin AA. Cumulative thermonuclear inertial reactor. Energy Sustain Soc 2016;6:8. [5] Frolov SM. Detonation initiation techniques for pulse detonation propulsion. Progress Propuls Phys 2009;1:321. [6] Gamezo VN, Ogawa T, Oran ES. Flame acceleration and DDT in channels with obstacles: effect of obstacle spacing. Combust Flame 2008;155:302. [7] Gelfand BE, et al. Detonation and deflagration initiation at the focusing of shock waves in combustible gaseous mixture. Shock Waves 2000;10:197. [8] Han W, Gao Y, Law CK. Flame acceleration and deflagration-to-detonation transition in micro- and macro-channels: an integrated mechanistic study. Combust Flame 2017-10-02;176:285. [9] Karlovitz B., Halasz D.: Process for the conversion of energy and apparatus for carrying out the process, US Patent No. 2 210 918, filed August 1935. [10] Lee BHK. Non-uniform propagation of imploding shocks and detonation. AIAA J 1967;5:1997. [11] Messerle HK. Magnetohydrodynamic power generation. John Wiley; 1994. [12] Munroe C. E.: On certain phenomena produced by the detonation of gun cotton, Proceedings of the newport natural history society, Newport, R. I. Report no. vol. 6, 1883e1886. [13] Nikitin VF, et al. Pulse detonation engines: technical approaches. Acta Astronaut 2009;64:281. [14] Qin Y, Yu J, Gao G. Detonation initiation induced by flame implosion and shock wave focusing. Trans Nan Jing Univ Aeronaut Astronaut 2011;28:57. [15] Shchelkin KI, Troshin YK. Gasdynamics of combustion. Mono Book Corp; 1965. [16] Singh VP. Effect of thermal pressure in converging detonation waves. Pramana 1985;24:527. [17] Sonský J., Tesar V., Gruber J.: Ways and means for converting energy in gaseous mixtures into electric energy, Czech Repub Patent Appl PV2017eP2117. ni A. Implosive thermal plasma source for [18] Sonský J, Tesar V, Gruber J, Masla energy conversion. Plasma Phys Technol 2017;4:87. [19] Tesar V.: Circuit connection of a fluidic commutator, Czechoslovak Patent Nr. 218285. filed Oct. 1980. [20] Tesar V. Taxonomic trees of fluidic oscillators. Eur Phys J 2017;143:02128. [21] Tesar V. Master and slave’ fluidic amplifier cascade. Eur Phys J 2012;25:01093. [22] Tesar V. High-frequency fluidic oscillator. Sensor Actuator Phys 2015;234:158. [23] Tesar V.: Fluidic oscillator, Czechoslovak Patent Nr. 306064 filed Dec. 2014. [24] Tesar V. No-moving-part electro/fluidic transducer based on plasma discharge effect. Sensor Actuator Phys 2015;232:20. [25] Tesar V.: Device for adjusting time dependence of fluid flow, Czechoslovak Patent Nr. 202898 filed Jan. 1979. [26] Makino T, Wakabayashi H. New spectrofotometer system for measuring thermal radiation characteristics of real surfaces of thermal engineering entirely. J Therm Sci Technol 2011;Vol. 6:80. [27] Wang B, He H, Yu S-TJ. Direct calculation of wave implosion for detonation initiation in pulsed detonation engines. In: 43rd AIAA aerospace sciences meeting and exhibit, Reno, Nevada; January 2005.
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
No-moving-part commutation of gas flows in generating plasma by cumulative detonations (survey) clav Tesar*, Jirí Sonský Va Department of Thermodynamics, Institute of Thermomechanics V.V.I., Czech Academy of Sciences, Dolejskova 5, 182 00 Prague, Czech Republic
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 October 2017 Received in revised form 26 April 2018 Accepted 24 May 2018 Available online 25 May 2018
This paper surveys controlling, without use of mechanical valves, of flows of three gases: a combustible mixture (fuel and oxidant), combustion products, and a neutral scavenging gas. The gases are delivered periodically in an alternating way into a pair of pulsed MHD electricity generators. Essential problem of this type of electricity generation is insufficient ionisation level of combustion products in the usual deflagration type burning. Solution was found in achieving momentarily extremely high temperatures by cumulative detonations. In the 1st phase the cavities leading into the 1st MHD generator are filled with the combustible mixture, which is then ignited, generating a high-temperature high-speed detonation wave moving into the generator. In the subsequent 2nd phase the combustion product are scavenged in the cavities by the neutral gas. In the other generator the phases are in reversed order. Because of the extreme temperature, the flows are switched not by mechanical valves but by no-moving-part devices, in various configuration that are systematically surveyed. © 2018 Elsevier Ltd. All rights reserved.
Contents 1. 2.
3.
4.
5.
6.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494 Generating plasma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495 2.1. Cumulative effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495 2.2. Single-shot model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495 Control of gas flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496 3.1. Two-phase operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496 3.2. Pulsation frequency and size of detonation tube . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496 Commutation by fluidic diverter valves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496 4.1. Jet-deflection amplification effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496 4.2. Coanda-effect bistable diverters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497 4.3. Commutator with Coanda-effect valves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 498 4.4. Fluidic commutator operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 498 Alternative commutator configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499 5.1. Fluidic oscillators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499 5.2. Autonomous commutator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500 5.3. “Master & Slave” valve configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500 5.4. Use of vortex diodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501 5.5. A faster configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502
* Corresponding author. E-mail address: [email protected] (V. Tesar). https://doi.org/10.1016/j.energy.2018.05.165 0360-5442/© 2018 Elsevier Ltd. All rights reserved.
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
494
1. Introduction Despite the contemporary large progress in renewable electricity production, the largest percentage of the generation processes still relies on burning various fuels. The generation consists of conversion steps with the typical sequence as follows: (1) (2) (3) (4) (5)
fuel combustion, steam generation in a boiler, accelerating the steam in turbine, generating mechanical power on turbine shaft, and electric current generation in induction coils.
Each of these steps is inevitably associated with some energy loss - so that, together with the Carnot cycle limit, the overall efficiency is low. Typically it is not above 35e40%. There is a generally known but practically not used alternative d the magnetohydrodynamic (MHD) generators. They avoid the above processes (2), (3), and (4), thus offering in principle a substantially better conversion efficiency. The original idea is described in Ref. [9] and the present state of art is surveyed in Ref. [11]. The principle requires heating the gas to temperature levels so high that the it is ionised and thus becomes electrically charged. Unfortunately, those few MHD facilities so far built for feasibility tests [11], all fail to reach the necessary ionisation temperatures. Various attempts at efficiency improvement, such as supplying pure oxygen or alkali metal particles to the combustion process, are highly un-economical. A substantially higher temperatures than now achievable - and consequently much higher degree of gas ionisation - may be reached by replacing the standard deflagration combustion by detonations. It is an idea that has not been so far followed because of the danger associated with very high mechanical stresses of the generator components. Moreover, even the degrees of ionisation achieved in the investigated detonations so far remain below the necessary level. A solution is offered in Ref. [17]. The basic idea is the cumulative implosion [4,14] inside the combustion chamber, Fig. 1. The temperature in the focal point of the colliding detonation waves is extraordinarily high, as is demonstrated by the military applications in the hollow charge weapons d and yet the stress levels in the chamber walls are not particularly high because of mutual cancellation of oppositely acting forces in the focal point [18]. The military hollow charge concept cannot be simply taken over for electricity generation. It is suitable for a single detonation of solid
Fig. 1. Schematic representation of the plasma generation by implosion detonations in the combustion chamber. Generated in the tube at left, the detonation wave is distributed through the large number of equal-length channels into wavelets all colliding mutually in the focal point.
explosives d while for the use in a MHD generator the detonations are to be in the gaseous phase and periodically repeated. This requires a rather complex system (seen schematically represented in Fig. 1) ensuring the synchronisation of colliding detonation waves coming towards the focus from all directions e and also scavenging the combustion products from the cavities after each detonation followed by re-filling by a fresh combustible mixture. To achieve the mutual cancellation of the opposing forces, the implosive wave inside the combustion chamber has to be generated in a spherical configuration very accurately e apart from the exit into the MHD generator. The necessary very small deviations from the sphericity are practically impossible to achieve with a large number of independent detonation sources on the outer wall of the combustion chamber. There are two reasons: (a) the extremely high speed of detonation waves propagation (always much faster than local speed of sound) and (b) uncertainties associated with the initiation of detonation waves. The latter fact is due to the character of deflagration-to-detonation transition. It depends on turbulence, which is a stochastic phenomenon with substantial uncertainty. Experience has shown that the required accuracy of the spherical symmetry is practically achievable only with a single detonation source e the detonation tube in Fig. 1 - with its output distributed into the large number of parallel equal-length channels. As a result, the necessary geometry of the cavities is rather complex. The removal by neutral gas of the combustion products from these complex cavities and their replacement by a fresh combustible mixture is achieved by gas flows - combustible mixture and neutral gas e each available from an external pressurised source. The order of their flows and their directing into the combustion chamber must be controlled by a commutator system. Because of the high temperature environment, mechanical commutator valves with moving components are out of question, especially due to preferable operation at the standard 50 Hz repetition frequency (Fig. 4) of the generated alternating electric current. The solution discussed in this paper is based on commutation by no-movingpart devices, made of a refractory material. 2. Generating plasma 2.1. Cumulative effect The cumulative effect of imploding detonation waves is capable of generating a plasma flow with high degree of ionisation at very high, hypersonic velocities. Both properties are desirable in the MHD generator. The high velocity is achieved by the axial imbalance of detonation waves, with the distributor channels missing on the right-hand side in Fig. 1, where there is the entry into the MHD generator. Existing literature discusses Mach numbers of the detonation wave propagation of the order Ma z 4e8 [14,16,24].
Fig. 2. Principle of generating the alternating current in a pair of MDH generators. Detonations are controlled by two synchronised diverter valves. The 1st detonation tube is here shown filled by combustible mixture through the node A while the 2nd tube is scavenged by air flow delivered through B.
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
Fig. 3. The other phase of the gas flows in the same circuit as in Fig. 2. While the 1st detonation tube is scavenged by air after the detonation, the 2nd tube is re-filled with fresh combustible gas mixture.
Fig. 4. Detonations in the pair of mutually phase-shifted MHD generators. The output from the 2nd generator is connected to generate the negative part of the AC period.
Also the high temperatures, measured spectroscopically (evaluated from changes in spectra lines of emitted light [26]) was already demonstrated in present authors' reference [18]. The implosive phenomenon is in literature known as Munroe or Neumann effect (although the priority of the basic idea is now known to belong to von Baader [2] who suggested e but failed to demonstrate in practice - the use of the effect in mining). The term “Munroe effect” - common in the USA - is named after the author of reference [12], who in 1883e1886 performed in US naval torpedo laboratory a series of demonstrations of solid-state hollow charges for armour penetration. 2.2. Single-shot model Correctness of the present authors' assumptions about the achievable performance parameters were demonstrated by performing experiments using a laboratory model, described in Ref. [18]. Because of the feasibility study purposes, it was operated in the regime of single shots, with long time periods between the detonations. The combustion chamber internal diameter was 95 mm. The combustible mixture used was the stoichiometric mixture of hydrogen and oxygen, produced prior to the experiments by water electrolysis and accumulated. The product gas of this combustion is steam in various contents of ions. It has a tendency to condense as water drops on the less hot locations of the internal cavity walls. The removal of the drops was one of the reasons for the waiting time between the individual shots d the air blowing was left to act for several minutes to ensure the cavities were dry. The detonation wave was generated in detonation tube of constant 8 mm 8 mm square internal cross section (which means a cross-section area 64 106 m2) and 180 mm long. After the drying of
495
the cavities and then re-filling them with the O2þ H2 mixture, the detonation was started by the electric high-voltage spark (using for the purpose a standard internal combustion engine type spark plug) near the entrance into the detonation tube. In the initial stages of the combustion process, the wave that progressed along the tube length from the ignition was in the ordinary mode of deflagration burning. Its transition into the detonation was caused by the turbuliser [7], a local configuration in the tube generating turbulent vortices. Instead of the popular Shchelkin's spiral [15] used as the turbuliser by a number of other researchers [5,8], in the authors' test model the turbulence was generated by a row of local tube excrescences [6]. As is apparent in Fig. 1, the part of the model most difficult to make was the distributor d a system of channels converting the planar detonation wave from the detonation tube into a large number of confocal wavelets. The conversion is performed by the channels increasing their number in the flow direction by progressive bifurcations. There were altogether 168 channels with exits directed towards the focal point inside the combustion chamber. The channels were all of the same 6 mm 6 mm cross sections (i.e. with total exits cross-section area 0.36 106 m2) and the same length 72.2 mm. For the design of the scavenging and re-filling systems are important factors the internal volumes in the cavities. The spherical combustion chamber volume was 450 106 m3, the volume of the detonation tube was 11.5 106 m3 and the total volume of the channels in the distributor was 440 106 m3 (almost the same as the combustion chamber). Considering the initial trifurcation at the inlet into the distributor, the total volume of all cavities was
Vz1:0 103 m3
(1)
In the experiments, for the scavenging was used a low-power blower, delivering approximate volume flow rate 83 106 m3/s. This means the minimum time between the shots would be 2 12 ¼ 24 s. Actual times in the tests (because of the need to dry the surfaces) was much longer, of the order of minutes. Of course, there was no intention to run the model in a periodic regime. The most important aspects of the tests were the achieved values of implosion temperature and gas velocity at the exit from the combustion chamber. The data obtained were quite promising. Velocity of the plasma leaving the combustion chamber was determined by analysis of high-speed camera images to be approximately 2.7 km/s. This corresponds to hypersonic Mach number Ma z 5.6 [13]. Temperature measurement was performed by analysis of hydrogen lines in the spectrum of emitted light [26]. The results were between 10 000 K and 15 000 K (may be compared with the temperature 6000 K of solar surface). 3. Control of gas flows 3.1. Two-phase operation Because of the requirements of standard electricity consumption, the ideal pulsed MHD electricity generator would operate at pulse repetition frequency 50 Hz. Simplified calculation of the ideal scavenging and re-filling times Dts and Dtf respectively, shows very short durations
Dts ¼ Dtf ¼ ½:1=50 ¼ 0:01s
(2)
This would necessitate choosing a very small size of both combustion chamber and generator. It is inconvenient from many points of view. Avoiding the very short times is possible with a multi-phase configuration. This, of course, means to have a number of MHD sequentially operating generators and other components
496
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
d which tends to be too expensive. As a reasonable compromise is seen two-phase operation, with two alternatively fired generators, as shown in Figs. 2e4. The repetition frequency of the detonation pulses is thus halved, ideally to 25 Hz. The scavenging and re-filling times are then
Dts ¼ Dtf ¼ ½:1=25 ¼ 0:02 s
(3)
It is a quite small but acceptable value. Of course, it means a fast operation of the valves used to control the gases flows. With mechanical valves it would be rather difficult to move so fast the solid components opening and closing the flowpaths into the nodes A and B. (Fig. 2). The critical factor is the inertia of these components. Also, the mechanical valves are unlikely to endure for a reasonable lifetime the high temperature environment and the detonation stresses. This paper suggests a different approach: applying the idea of no-moving-part fluidic commutator from Ref. [19]. Instead of opening and closing the entry into the gas flow pipes, this design operates by diverting the flows. Two jets generated in a pair of nozzles are alternatively captured and led into one of the two exits and the two combustion chambers connected to them.
3.2. Pulsation frequency and size of detonation tube Although the demonstration laboratory model described in Sect. 2.2 was not intended for actual AC electricity production, it is useful to estimate how it would behave in a periodic operation regime. It is assumed that there are, under pressure, external sources of the two commutated gases d the combustible gas mixture and air used for the scavenging. The flows of these gases have to from the detonation cavity its whole contents and to re-fill it. According to the equations (1) and (3) the volume flow rates needed for the task are
. . V_ ¼ V=Dts ¼ 1:0 103 0:02 ¼ 0:50 103 m3 s
4. Commutation by fluidic diverter valves 4.1. Jet-deflection amplification effect The basic components of the now discussed fluidic commutator is the synchronised pair of fast acting flow diverter valves, schematically represented at the left side of Figs. 2 and 3. The fluidic commutator as suggested in Ref. [17] consists of the two jet deflection devices, based on the ideas shown in Figs. 5a, 5b, 6 and 7. Solid and not moving, the fluidic valves made of a refractory material can withstand the adverse high-temperature conditions for very long periods of time without significant erosion. Their task is to direct the supplied gas alternatively between the nodes A and B (Figs. 2 and 3) which lead to the two detonation tubes. Fluidic valves have also already demonstrated the capability to operate at high switching frequencies. For example, ref. [22] has shown a fluidic jet-type valve oscillating at the supplied air flow rate V_ ¼
0.324 103 m3/s d which is merely 1.5-times less than the value in eq. (4) and may be easily increased d at the frequency f ¼ 8230 Hz. This is a value 330-times higher than what is needed in the commutator. The operation cycle of the commutator consists of the two phases shown in Figs. 2 and 3. At the end of each phase is generated the detonation wave. In the first phase, presented schematically in Fig. 2, the 1st detonation cavities (on top of the picture) are filled with the combustible gas mixture through the node A d while the other, bottom diverter in this picture despatches into the node B the scavenging gas. This makes the 2nd (bottom) cavities prepared for the next filling by the combustible mixture. The first phase finishes by ignition of the mixture in the 1st detonation tube and the subsequent (extremely brief) detonation. The detonation wave progresses through 1st detonation tube into the combustion chamber and from there to the MHD electricity generator. The fast moving
(4)
This is a large value. In the test model the smallest cross section was 8 mm 8 mm of the detonation tube, of cross section area F ¼ 64 106 m2. This means that the flow of the air in the detonation tube of the original size move during the scavenging through the tube at the velocity 780 m/s, i.e. at the supersonic Mach number Ma z 2.4. This is out of question. The situation may be improved upon by choosing a larger detonation tube cross section F. A reasonable velocity value with substantially lower driving power would be obtained by selecting approximately 10-times larger area 25 mm 25 mm of the detonation tube cross section. The value is, of course, the reasonable Ma z 0.24.
Fig. 6. Bistable Coanda-effect attachments alternatively to one of the two walls. Switching between the two regimes is achieved by short flow pulses applied to the control nozzle.
Fig. 5. (a). The most popular among the no-moving-part fluidic flow control principles is jet deflection of the powerful main jet by small control flow. (b). Coanda effect: wall presence l prevents the jet-pumping entrainment on the top side of the jet creating the transverse pressure difference between its sides. This deflects the jet towards the wall and keeps it attached.
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
Fig. 7. Typical bistable fluidic diverter valve and its schematic representation.
flow of the charge carriers passing through the induction coil generates in the MHD generator segmented electrodes exposed to the flow an electric current pulse. Immediately thereafter begins the second phase, presented schematically in Fig. 3. The top diverter valve directs the combustible gas mixture into the node B while the other, bottom diverter despatches the scavenging gas into the other node A. This results in removing from the 1st detonation tube e as well as from the combustion chamber connected to it e the combustion products. At the same time, the 2nd detonation tube and the connected combustion chamber d after their being scavenged in the previous first phase d are now, as is shown in Fig. 3, filled with the combustible mixture. Again, when the cavities are full, this filling ends by ignition and generation of detonation wave. It progresses through the mixture to the combustion chamber and to the MHD generator. There the inversely connected MHD coil generates in the common exit terminal the negative current pulse - symbolically shown at right in Fig. 3. As a result, the two pulses (positive and then negative) represent a full cycle of an AC output current. Its waveshapes are not likely to be harmonic, but this is easy to improve d in the simplest way it may be done by a passive LC filter and a transformer.
497
nozzle (in A of Fig. 5a) to generate a control flow. By the momentum interaction between the main jet and the control flow the main jet is deflected (upwards in Fig. 5a). Opposite to the exit from the main nozzle may be then placed a collector capturing the main jet and leading it into the output terminal. Magnitude of the output flow rate captured in this terminal thus depends on the deflection angle and this, in turn, on the magnitude of the control flow rate. The device operates as a fluidic amplifier. Early designers of the fluidic jet-deflection amplifiers were surprised by encountering a tendency of the jet to remain deflected d as shown in B of Fig. 4 d even if the corresponding control flow was absent. The phenomenon is known as the Coanda effect of jet attachment to a wall. It is caused by the jet entraining the surrounding gas and carrying it away. If there is a nearby solid wall (B in Fig. 4), the entrainment on that side is prevented. This develops between the jet and the wall a very low pressure - resulting in a side force that deflects the jet. This jet attachment phenomenon, initially considered a nuisance, later became very useful d especially in the symmetric configuration with two attachment walls, shown in Fig. 5. There is no continuous jet deflection control e instead there are only two stable regimes. In the discussed application, however, this is exacly what is needed. The configuration of the valve, as shown in Fig. 7, is planar, with everywhere constant depth h (Fig. 6) of the cavities. The jet may be switched between the two attachment walls by short flow pulses applied to the control nozzle on the wall attachment side (the top nozzle inactive in Fig. 5b). Once the jet is switched, it remains deflected and the control flow is no more needed. This results in saving of the control power. The bistable configuration is symmetric and provided with the two collectors, as shown in Fig. 7. A continuation of the flowpath from the collectors is on each side of the valve a diffuser. Its task is the opposite to the action in a nozzle: it converts kinetic energy of the jet in the pressure energy (note that the hydraulic losses depend on the kinetic energy). Also note that the apex angle of a diffuser has to be quite small to avoid flow separation from its walls and consequent loss of efficiency. The small apex angle results in long diffusers, which tend to dominate the whole valve body in Fig. 7. In the bottom part of Fig. 7 is a schematic representation of the valve as it is used in circuit diagrams. The nozzles there are represented by black triangle symbol while the diffusers are shown as white (i.e. not filled) triangles. Also schematically represented in Fig. 7 is the pair of attachment walls.
4.3. Commutator with Coanda-effect valves Shown in two identical circuit diagrams presented in Figs. 8 and 9 are the two phases of the cycle in the commutator according to [19]. In particular, drawn by the thick curved grey line, is indicated
4.2. Coanda-effect bistable diverters The pair of fluidic gas flow diverting valves described as the essential parts of the commutator in Ref. [19] operate on the principle of deflecting a gas jet leaving a nozzle, shown in Fig. 5a and 5b. The jet is formed from the supplied gas flow in the main nozzle (at left in the pictures). A jet-pumping action e entrainment of the surrounding gas - generates near the nozzle exit a region of low pressure. This makes easy for perpendicularly oriented control
Fig. 8. Circuit diagram of the fluidic commutator according to [19]. After the control pulse a had switched the combustible gas mixture flow into the node A, the 1st detonation tube is re-filled by the combustible gas mixture.
498
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
4.4. Fluidic commutator operation
Fig. 9. The other phase of the fluidic commutator from Fig. 8. After the detonation and the switching pulse b it is the 1st tube that is scavenged e while combustible mixture fills the 2nd detonation tube.
in both pictures the character of the flow of combustible gas mixture. The role of the two schematic flow diverters that was shown above as a block diagrams in Figs. 2 and 3 is here performed by the two jet-deflection Coanda effect valves e both corresponding to the Fig. 7. Both pictures Figs. 8 and 9 show also the control lines that transfer into the valves the flow-switching pulses a (in Fig. 8) and b (in Fig. 9). The anti-parallel operation of the valves is as follows: if one valve delivers the combustible gas through the node A into the 1st detonation tube (Fig. 8), the other valve simultaneously directs the scavenging air through the node B into the 2nd tube d and vice versa. The commutator alternates between two phases, the PHASE 1 and PHASE 2 which together form the operation cycle. The corresponding time duration diagrams of the gas flows through the node A are presented on the top of Fig. 10. The spatial character of the flow drawn by thick grey curved line in Fig. 8 occurs after arrival of input flow pulse a e and lasts then until the arrival of the pulse b. This latter pulse comes at a the instant of time (Fig. 10) when the 1st detonation tube and cavities connected to it are already re-filled with the combustible mixture d and, in fact, when there is already some inflow of protective neutral gas. This suppresses occurrence of a return motion of detonation wave in the tube. This difference between the two phases results in the slight asymmetry of the two phases in Fig. 10. After the detonation, which is started by the electric ignition pulse (Fig. 10), the corresponding detonation tube is cleaned and the combustion products are removed being displaced by air. In the PHASE 2 the combustible gas mixture is switched e as shown by the thick grey line in Fig. 9 to fill the 2nd detonation tube through the node B.
Because of the geometric symmetry of both phase flowpaths, it is sufficient for explaining the operation of the jet-deflection commutator to describe only the processes by which is serviced through the node A - the 1st detonation tube. The time dependences of flow and switching control pulses are presented in Fig. 10 while the spatial aspects are presented in Fig. 11 for PHASE 1 and in the next Fig. 12 for the PHASE 2. The operation cycle begins by arrival of the fluid flow pulse a (Fig. 10) into the upper control nozzle (Fig. 11) of the diverter D1, which switches the air flow away from the node A and e since the pulse is simultaneously admitted into the lower control nozzle of the other diverter D2 e directing into A the combustible mixture O2þ H2. From D2 the mixture continues into the 1st detonation tube (which was in the previous cycle scavenged). The filling of the 1st tube and also of all the channels leading from it into the combustion chamber lasts for the whole PHASE 1. This filling of cavities stops after arrival of the pulse b (Figs. 10 and 12) which determines the beginning of PHASE 2. The diverter D1 now admits into the node A the scavenging air which progresses from there into the inlet of the 1st detonation tube. The combustible mixture in this tube is pushed forward to the combustion chamber and upstream from the spark plug is formed the non-combustible air “cushion”. This is the exact time for the combustion started by the spark. Note the position of the electric impulse at the bottom of Fig. 10. No combustion can take place in the “cushion” because it now contains only non-combustible air and no fuel. There is therefore no retrograde flow into the left-hand inlet part of the 1st detonation tube. The combustion progresses in the detonation tube towards the combustion chamber, but very soon it reaches the turbuliser (cf. also Fig. 1). This converts the deflagration into the detonation. This reaches the very high supersonic velocity and progresses through the rest of the detonation tube as well as through all the other cavities that were previously filled with the O2þH2 mixture. This all happens at the small initial fraction of PHASE 2. By the cumulative effect the colliding waves [1,3,10,27] generate ionisation in the combustion chamber and directs the resultant charge carriers into the 1st MHD generator. Clean air continues flowing into the 1st detonation tube for the rest of the PHASE 2 removing the combustion products that were left there by the detonation wave. This scavenging action ends by the arrival of the flow pulse a, which starts the next cycle.
5. Alternative commutator configurations 5.1. Fluidic oscillators Apart from the basic version of the fluidic commutator described in Ref. [19], there are several other alternatives. Those discussed here retain the jet-deflection type of the fluidic diverter valves - but with some changes in the circuits. Especially interesting
Fig. 10. Time dependence of flows in the node A (and in the 1st detonation tube connected to it) of the commutator in Fig. 8. Also shown is the timing of the control pulses a (Fig. 8) and b (Fig. 9) and the ignition.
Fig. 11. Schematic representation of the filling process in the 1st detonation tube by the fluidic switching circuit shown in Fig. 8.
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
499
Fig. 12. Schematic representation of the air flow in the commutator cavities at the instant of igniting the combustible gas mixture.
are configurations employing the amplification capability of the fluidic jet-deflection valves. This property may be used to generate self-excited periodic oscillation. The advantage thus gained is the autonomous character of the commutator, not needing the external switching pulses from external source. The commutator becomes a self-contained unit consisting of nothing more than cavities made inside a solid body. The control by external flow pulses - as shown in Figs. 10e12 above e usually assumes a timing in an electric controller and an electro/fluidic (E/F) transducer [24] converting the electric pulses into fluidic ones. These devices have to be located near the detonation tubes and combustion chambers. Considering the extremely high temperatures inside the combustion chamber and the quite short distances from it, it is necessary to expect elevated temperatures also in the whole commutator. This brings no problem for the fluidic commutator itself e the thermal robustness is one of the reasons for its choice, as it is nothing more than cavities in a solid material of refractory character. Unfortunately, similar resilience to heat without complications of added cooling to heat cannot be expected in usual electronics and transducers. The generation also of the switching signals by fluidics is therefore an interesting possibility d the more so because the fluidic pulse generator may be very simple and in fact necessitates just an addition of a feedback flow channels. An example of fluidic oscillator is presented in Fig. 13. Similar in principle to the feedback-type oscillators in electronics, typical fluidic oscillators also consist of two components: an amplifier and a feedback loop. It should be noted that apart from the configuration from Foig. 13 there are many other oscillator circuits, as reviewed in Ref. [20]. The important fact is the role of the fluidic amplifier in the oscillator may be played by one of the diverter valves in the commutator e simultaneously with its flow switching role. All that is necessary is to connect by the feedback loop channel the two control terminals X1 and X2 (Fig. 13). Length of the connecting loop - which may be also shaped as a cavity in the refractory body defines the oscillation frequency. The switching of the jet inside the valve from one attachment wall to the opposite one is caused by the flow in the feedback loop and the control nozzles. The flow is driven by the pressure difference DPX (Fig. 13) e the difference between the two sides of deflected jet. The pressure in X1, at the side to which the jet is attached, is lower than the pressure on the opposite side, in the location X2. The fluid in the loop thus flows under the action of this pressure difference from X2 to X1. Because of the amplification capability of the fluidic valve (i.e. its ability to switch the powerful main jet by a small flow issuing from the control nozzle) the flow in the loop suffices for switching the deflected jet. After this switching to the other side, the situation e direction of the pressure difference force - is reversed. This reversal takes a certain time and this represents the phase shift necessary for generating the oscillation. Appearance of one example of the robust commutator with the integral oscillator,
Fig. 13. Fluidic oscillator set up from the diverter by added feedback loop in which the flow is driven by the pressure difference DPX between the control terminals. Oscillation frequency is adjusted by varying the loop length [20].
insensitive to high temperature, is presented in the next Fig. 14. The cavities shown there were made by laser cutting in the flat plate of ceramic material. From top are the cavities closed by another flat plate (not shown) of the same material e in a similar manner as is done on the bottom side. The cavities actually consist of a whole stack of plates. The lines in Fig. 14 indicate the connecting (much wider) cavities made in some of these plates. 5.2. Autonomous commutator The circuit diagram of a commutator with integral generator of switching signals, is presented in Fig. 15. There is a pair of switched diverter valves, D1 (handling air) and D2 (handling the combustible mixture). They periodically vary their diverted gas flows into the 1st and 2nd detonation tubes. The valve D1 has the feedback loop
Fig. 14. An example of the commutator cavities made by the refractory material removal. Also shown are idealised channel connections ensuring both feedback and mutual synchronisation of the diverters.
500
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
connecting, like the oscillator in Fig. 13, its control nozzles. The alternating air flows from D1, apart from their ordinary role of the scavenging of the internal spaces of the commutator, also deliver the air into the control nozzles of the other valve D2. This valve is periodically switched by the air flows through the control inlets a and b. The next two illustrations, Figs. 16 and 17, explain the internal processes in the commutator from Fig. 15. In the Fig. 16 is shown by the superposed wide grey colour lines, what happens in this commutator the PHASE 1. The next Fig. 17 then presents the flow phenomena occurring in the other PHASE 2. The air flow from the output of the oscillator in Fig. 16 is seen divided into two branches. The more substantial flow enters the node A and continues into the 1st detonation tube. It is shown arriving there at the instant of the air cushion formation. The smaller part of the air flow is branched into a and there it controls the switching of the gas mixture. It deflects the mixture jet and directs it into the node B, continuing from there into the 2nd (bottom) detonation tube. When the air flow in the feedback loop of the oscillator is reversed, it generates an outflow from the other control nozzle, as shown in Fig. 17. It is seen directing the main air flow into the node B. This air flow leaving the oscillator also bifurcates. The much more powerful air flow gets into the 2nd detonation tube and scavenges it. The smaller part of the flow from the bifurcation passes through b. It enters the gas mixture valve and diverts its flow into the 1st tube. For switching the valves would suffice in the control inlets a and b only very short flow pulses. This, however, does not happen in the circuit shown in Figs. 15e17. The control flows there last for the whole duration of each of the respective phases. In some commutator designs this causes no problems e especially if the scavenging role is played by air, the oxygen contents of which may be used in the combustion process. Nevertheless, the resultant combustion is
Fig. 17. Gas flows after the detonation in the 2nd phase d in the circuit presented above in Fig. 15.
likely to be less efficient than burning the pure oxygen. There may be circumstances that require for the scavenging a gas wholly neutral, perhaps pure nitrogen. Its inflow through the control nozzle for the whole duration of each phase may make the combustion process less efficient due to the dilution of the combustible mixture. It is useful to consider present alternatives suppressing this dilution effect. First of all, it is obvious that the dilution is the less intensive the higher is the flow gain of the gas-mixture diverter valve. Higher flow gain (i.e. the ratio of the main nozzle flow to the switching control flow) means less scavenging gas needed for switching the gas mixture flow rate. 5.3. “Master & Slave” valve configurations Intensity of the dilution of the combustion gas mixture, caused by the prolonged control flow into the diverter handling the combustion gas mixture may be reduced substantially by the circuit layout presented in Fig. 18. The layout [21] is very similar to the one shown in Fig. 15. The difference is a single valve of Fig. 15 is in Fig. 18 replaced by an amplifying cascade. The overall flow gain value G of this cascade is the product of the two gain values G1 and G2 of the two diverter valves. It is thus possible to obtain the desirable small dilution. 5.4. Use of vortex diodes
Fig. 15. An example of fluidic autonomous commutator with one air flow diverter (here the diverter D1) operating as the generator of control pulses.
Fig. 16. Gas flows in the circuit presented in Fig. 15. The flow lines suggests the conditions just prior to spark ignition in the 1st detonation tube. Note the air forming the non-combusting “cushion” upstream from the spark plug in this tube.
An alternative commutator design with autonomous selfexcited oscillation e but suppressed dilution of the combustible gas mixture - is presented in Fig. 19. Its basic idea is use of dynamic properties of vortex diodes d their capability to shorten the pulse
Fig. 18. Reducing the dilution of the combustible mixture by the scavenging gas may be achieved as shown here using the “Master & Slave” flow-diverting circuit discussed in Ref. [21].
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
duration to mere fraction of the half-period. The circuit diagram of this configuration is presented in Fig. 19. It is very similar to Fig. 15, the only difference is the addition of the two diodes into the control flow entrances a and b. Both diodes are oriented in the highdissipance direction, i.e. with the internal air rotation suppressing the flows to the control nozzles. The property employed in the present case is the growth of dissipance (resistance to the flow) after the beginning of the input flow pulse. This shapes the transmitted flow pulses, as was the subject of ref. [25] (differing from the configuration shown here by an unimportant detail) or the parts of the oscillator based on the same principle [23]. The behaviour of fluidic diodes is explained in Figs. 20e22. In literature are described fluidic diodes of quite complex configuration, but for the explanation purposes it is here useful to consider a quite simple example of diodes consists of a shallow circular vortex chamber, tangentially oriented inlet, and central outlet. The depth of the chamber is small so that if the inflow (indicated by the wide white arrow at the top of the picture) stops the motion inside the chamber is rapidly stopped by friction on the bottom surface - as well as the top surface which in these pictures is removed. In Fig. 22 there are two diagrams of the flow rate dependences on time. Let us consider the beginning of the cycle in which the oscillator in Fig. 19 has just switched the air flow into the node A and to the control inlet a. Prior to this incoming flow pulse both flow rates are zero. There is no air rotation inside the vortex chamber of the diode d1 because it was stopped by the friction on the top and bottom surface. The air admitted into the tangential nozzle does not meet any large opposition and passes through the diode rather easily. It has only to overcome in the leading part of the control-flow pulse to overcome the inertia of the accelerated air. The character of this easy flow is indicated in Fig. 20. Because of the tangential orientation of the nozzle in the diode, however, the fluid inside the chamber begins to rotate as shown in Fig. 21. There is the general tendency of this rotation to keep more or less constant the moment of fluid momentum. As the radial distance r of the gas decreases e since it has to leave through the central exit d this conservation tendency leads to increasing the tangential velocity w. As the time progresses, this rotation velocity actually rapidly grows and soon becomes so high that the centrifugal force acting on the gas almost blocks the flow. The output flow rate diagram in Fig. 22 shows what is practically an end of the pulse in the diode exit location Y (only very small air flow manages to pass through the boundary layers). The diode thus effectively shortens the duration of the control flow pulse. It lasts only during the period of attaining the full rotation speed. During the rest of the phase the flow rate through the diode and through the commutator valve control nozzle - is negligible.
501
Fig. 20. (Left) Vortex diode at the beginning of the flow through the vortex chamber. The fluid passes to the central exit without any significant hydraulic resistance.
Fig. 21. (Right) Later the fluid in the vortex chamber becomes spinning. The tangential velocity w increases rapidly with diminishing flow radius r. The flow through the diode is now effectively opposed by the centrifugal force.
5.5. A faster configuration The generation of the plasma pulses at the repetition frequency corresponding to the AC electricity 50 Hz would simplify the uses of the electricity generator. The problem is the scavenging and refilling times in the cavity may tend to be rather long. This calls for increasing the scavenging and re-filling flow velocities, but this is not a really good idea as it would mean higher pressure to be overcome e and may even encounter the Mach number compressibility limitation. Much better approach is the shortening of the flowpaths. The detonation tube may be made shorter e and as shown in Fig. 23, the gases may enter the cavities roughly in the location at the middle of distance that has to be travelled.
6. Conclusions
Fig. 19. An alternative possibility how to suppress the combustible gas mixture dilution in the commutator by using the dynamic properties of vortex diodes [22,23].
The capability of achieving the ionisation temperatures and also the high detonation-wave velocities in an implosion-type combustor was already earlier demonstrated in a single-shot laboratory model. The next stage, discussed in this paper, concerns the
502
V. Tesar , J. Sonský / Energy 157 (2018) 493e502
in the internal cavities. There are alternative configurations, in the paper surveyed and discussed. Acknowledgements Authors' research was supported by institutional support research RVO:61388998 and Czech Republic Grant Agency GACR grant 17/08218S. References
Fig. 22. Graphical presentation of vortex diode response to the pulse of the input flow _ The initial part of the pulse passes quite easily from X to Y d but once the gas rate M X _ Y is opposed and rotation speed inside the vortex chamber increases, the diode flow M practically stops.
Fig. 23. Flows in the commutator may be not fast enough for the desirable 50 Hz frequency of pulses. Faster scavenging and re-filling may be obtained by the shorter travelled distances iby the commutator (here shown with monostable switching valves) delivering the gases into the middle location, between the detonation tube and the combustion chamber.
auxiliary facility necessary for achieving the reciprocating operation of the plasma MHD generation d the delivering alternatively the combustible gas mixture and a scavenging gas into and away from the combustion chamber. The configuration considered is the two-phase version, with two systems (combustion chamber and generator) operating in anti-parallel. The gas flow commutator [19] for handling the gases is the no-moving-part fluidic unit e effectively a solid body of refractory material with the actions performed
[1] Axford RA, Holm DD. Converging finite-strength shocks. Phys Nonlinear Phenom 1981;2:194. [2] v. Baader FX. An attempt at a theory of working with explosives. Bergmannisches J 1792;1:193. [3] Bilbao LE, Gratton J. Spherical and cylindrical convergent shocks. Il Nuovo Cimento 1996;18 D:1041. [4] Bolonkin AA. Cumulative thermonuclear inertial reactor. Energy Sustain Soc 2016;6:8. [5] Frolov SM. Detonation initiation techniques for pulse detonation propulsion. Progress Propuls Phys 2009;1:321. [6] Gamezo VN, Ogawa T, Oran ES. Flame acceleration and DDT in channels with obstacles: effect of obstacle spacing. Combust Flame 2008;155:302. [7] Gelfand BE, et al. Detonation and deflagration initiation at the focusing of shock waves in combustible gaseous mixture. Shock Waves 2000;10:197. [8] Han W, Gao Y, Law CK. Flame acceleration and deflagration-to-detonation transition in micro- and macro-channels: an integrated mechanistic study. Combust Flame 2017-10-02;176:285. [9] Karlovitz B., Halasz D.: Process for the conversion of energy and apparatus for carrying out the process, US Patent No. 2 210 918, filed August 1935. [10] Lee BHK. Non-uniform propagation of imploding shocks and detonation. AIAA J 1967;5:1997. [11] Messerle HK. Magnetohydrodynamic power generation. John Wiley; 1994. [12] Munroe C. E.: On certain phenomena produced by the detonation of gun cotton, Proceedings of the newport natural history society, Newport, R. I. Report no. vol. 6, 1883e1886. [13] Nikitin VF, et al. Pulse detonation engines: technical approaches. Acta Astronaut 2009;64:281. [14] Qin Y, Yu J, Gao G. Detonation initiation induced by flame implosion and shock wave focusing. Trans Nan Jing Univ Aeronaut Astronaut 2011;28:57. [15] Shchelkin KI, Troshin YK. Gasdynamics of combustion. Mono Book Corp; 1965. [16] Singh VP. Effect of thermal pressure in converging detonation waves. Pramana 1985;24:527. [17] Sonský J., Tesar V., Gruber J.: Ways and means for converting energy in gaseous mixtures into electric energy, Czech Repub Patent Appl PV2017eP2117. ni A. Implosive thermal plasma source for [18] Sonský J, Tesar V, Gruber J, Masla energy conversion. Plasma Phys Technol 2017;4:87. [19] Tesar V.: Circuit connection of a fluidic commutator, Czechoslovak Patent Nr. 218285. filed Oct. 1980. [20] Tesar V. Taxonomic trees of fluidic oscillators. Eur Phys J 2017;143:02128. [21] Tesar V. Master and slave’ fluidic amplifier cascade. Eur Phys J 2012;25:01093. [22] Tesar V. High-frequency fluidic oscillator. Sensor Actuator Phys 2015;234:158. [23] Tesar V.: Fluidic oscillator, Czechoslovak Patent Nr. 306064 filed Dec. 2014. [24] Tesar V. No-moving-part electro/fluidic transducer based on plasma discharge effect. Sensor Actuator Phys 2015;232:20. [25] Tesar V.: Device for adjusting time dependence of fluid flow, Czechoslovak Patent Nr. 202898 filed Jan. 1979. [26] Makino T, Wakabayashi H. New spectrofotometer system for measuring thermal radiation characteristics of real surfaces of thermal engineering entirely. J Therm Sci Technol 2011;Vol. 6:80. [27] Wang B, He H, Yu S-TJ. Direct calculation of wave implosion for detonation initiation in pulsed detonation engines. In: 43rd AIAA aerospace sciences meeting and exhibit, Reno, Nevada; January 2005.