- Email: [email protected]
A Low-Cost Test Bench for Underwater Thruster Identification
Available online at www.sciencedirect.com
ScienceDirect A A A A
IFAC PapersOnLine 52-21 (2019) 254–259
Low-Cost Low-Cost Test Test Bench Bench for for Underwater Underwater Low-Cost Test Bench for Underwater Thruster Identification Low-Cost Test Bench for Underwater Thruster Identification Thruster Identification Thruster Identification ∗ Abdelmalek LAIDANI ∗,∗∗ ∗,∗∗ Mohamed BOUHAMIDA ∗
Abdelmalek LAIDANI Mohamed BOUHAMIDA ∗,∗∗ ∗ ∗ ∗∗∗ ∗,∗∗ Mohamed ∗ Abdelmalek LAIDANI BOUHAMIDA Mustapha BENGHANEM ∗ Karl SAMMUT ∗∗∗ Abdelmalek LAIDANI Mohamed BOUHAMIDA Mustapha BENGHANEM Karl SAMMUT ∗,∗∗ ∗ ∗ ∗∗∗ ∗∗,∗∗∗ ∗ Karl ∗∗∗ Abdelmalek LAIDANI Mohamed BOUHAMIDA Mustapha BENGHANEM SAMMUT Benoit CLEMENT ∗∗,∗∗∗ Mustapha BENGHANEM Karl SAMMUT ∗∗∗ Benoit CLEMENT ∗ ∗∗,∗∗∗ ∗∗,∗∗∗ Mustapha BENGHANEM Karl SAMMUT Benoit CLEMENT CLEMENT Benoit ∗ Benoit CLEMENT ∗∗,∗∗∗ ∗ AVCIS Research Laboratory, Department of Automatics, USTO-MB, ∗ AVCIS Research Laboratory, Department of Automatics, USTO-MB, ∗ AVCIS Research Laboratory, Department Oran, Algeria, (e-mail: [email protected]). AVCIS Research Laboratory, Department of of Automatics, Automatics, USTO-MB, USTO-MB, ∗ ∗∗ Oran, Algeria, (e-mail: [email protected]). AVCIS Research Laboratory, Department of Automatics, USTO-MB, Oran, Algeria, (e-mail: [email protected]). Lab-STICC, UMRCNRS 6285, ENSTA-Bretagne, 29806 Brest ∗∗ Oran, Algeria, (e-mail: [email protected]). Lab-STICC, UMRCNRS 6285, ENSTA-Bretagne, 29806 Brest ∗∗ ∗∗ Cedex Oran, Algeria, (e-mail: [email protected]). Lab-STICC, UMRCNRS 6285, ENSTA-Bretagne, 29806 9, France (e-mail: [email protected]). Lab-STICC, UMRCNRS 6285, ENSTA-Bretagne, 29806 Brest Brest Cedex 9, France (e-mail: [email protected]). ∗∗∗∗∗ Lab-STICC, UMRCNRS 6285, ENSTA-Bretagne, 29806 Brest Cedex 9, France France (e-mail: [email protected]). Flinders University, Adelaide, SA 5042, Australia (e-mail: ∗∗∗ Cedex 9, (e-mail: [email protected]). Flinders University, Adelaide, SA 5042, Australia (e-mail: ∗∗∗ ∗∗∗ Cedex 9, France (e-mail: [email protected]). Flinders University, Adelaide, SA 5042, Australia (e-mail: [email protected]). Flinders University, Adelaide, SA 5042, Australia (e-mail: [email protected]). ∗∗∗ Flinders University, Adelaide, SA 5042, Australia (e-mail: [email protected]). [email protected]). [email protected]). Abstract: Underwater vehicles are increasingly being used for the inspection of marine strucAbstract: Underwater vehicles are increasingly being used for the inspection of marine strucAbstract: Underwater vehicles are being used the of structures, bathymetry and underwater intervention applications. manoeuvring performance Abstract: Underwater vehicles are increasingly increasingly being used for for The the inspection inspection of marine marine structures, bathymetry and underwater intervention applications. The manoeuvring performance Abstract: Underwater vehicles are increasingly usedoffor thethruster inspection of marine tures, bathymetry and underwater intervention applications. The manoeuvring performance of these vehicles isand critically dependant on thebeing quality the model which strucmust tures, bathymetry underwater intervention applications. The manoeuvring performance of these vehicles is critically dependant on the quality of the thruster model which must tures, bathymetry and underwater intervention applications. Thethruster manoeuvring performance of these vehicles critically dependant on the quality of the model which must faithfully representis the relationship between thruster inputs and corresponding outputs. of these vehicles is critically dependant on the quality of the thruster model which must corresponding outputs. faithfully represent the relationship between the thruster and of these vehicles critically dependant on quality ofinputs the thruster model which must faithfully represent relationship between inputs and outputs. Thruster manufacturers typically provide onlythe thethruster essential electrical, mechanical and thrust faithfully representis the the relationship between the thruster inputs and corresponding corresponding outputs. Thruster manufacturers typically provide only the essential electrical, mechanical and thrust faithfully represent the relationship between the thruster inputs and corresponding outputs. Thruster manufacturers typically provide only the essential electrical, mechanical and thrust performance characteristics, for this reason the identification of the thruster’s parameters is Thruster manufacturers typically provide only essential electrical, mechanical and thrust performance characteristics, for this this reason thethe identification of the the thruster’s thruster’s parameters is Thruster provide only the essential electrical, mechanical andvehicle. thrust performance characteristics, for reason the identification of parameters is especially manufacturers important for typically constructing a reliable control system for thruster’s the underwater performance characteristics, for this reason the identification of the parameters is especially important for constructing a reliable control system for the underwater vehicle. performance characteristics, for the is identification of the parameters is especially important constructing aa reliable control system for the underwater vehicle. Physical testing usingfor some formthis of reason test bench normally used for thruster’s identifying the thruster especially important for constructing reliable control system for the underwater vehicle. the thruster Physical testing using some form of test bench is normally used for identifying especially important for constructing a reliable control system for the underwater vehicle. Physical testing using some form of test bench is normally used for identifying the thruster parameters. These test benches usually consist of a water tank and the thruster which is clamped Physical testing using some form of test bench normally usedthe forthruster identifying the thruster parameters. These test benches usually consist of ais tank and which is clamped Physical testing using some form of test bench iswater normally used forthruster identifying the thruster parameters. These test usually consist of water tank and the which is to some form of instrumented reaction frame for model parameters. The cost of parameters. These test benches benches usually consist of aameasuring water tankthe and the thruster which is clamped clamped to some form of instrumented reaction frame for measuring the model parameters. The cost of of parameters. test benchesto usually consist of ameasuring water and the is clamped to some formThese of instrumented instrumented reaction framefactors for measuring the model parameters. The cost a test bench varies according two major whichtank arethe themodel sizethruster and thewhich instrumentation to some form of reaction frame for parameters. The cost of a test bench varies according to two major factors which are the size and the instrumentation to some formvaries ofdepends instrumented reaction frame for measuring the model parameters. of a test bench according two factors which the size the which strongly on theto physical quantities that we are are trying toand measure. InThe this cost article a test bench varies according to two major major factors which are the sizeto and the instrumentation instrumentation which strongly depends on the physical quantities that we are trying measure. In this article a test bench varies according to two major factors which are the size and the instrumentation which strongly depends on the physical quantities that we are trying to measure. In this article we will make a state of the art survey of existing test benches in order to provide a technical which strongly depends on the physical quantities that we are trying to measure. In athis article to provide technical we will make a state of the art survey of existing test benches in order which strongly the quantities that are trying to measure. In aathis article we will make a state of the art survey of existing test benches in to provide technical basis that will allow to build a low-cost test bench. A thruster will be using the we will make a depends state us of on the artphysical survey of existing test we benches in order order to modelled provide technical basis that will allow us to build a low-cost test bench. A thruster will be modelled using the we will make a allow stateand of the art survey of existing test benches in order topaper. provide a using technical basis that willbench allow us to build a low-cost low-cost test bench. A thruster will be modelled using the resulting test thebuild outcomes will be presented at thruster the end of this basis that will us to a test bench. A will be modelled the resulting test bench and the outcomes will be presented at the end of this paper. basis thattest willbench allowand us to a low-cost test bench. A will bepaper. modelled using the resulting the outcomes will presented at the this resulting test bench and thebuild outcomes will be be presented at thruster the end end of of this paper. Copyright ©test 2019. The Authors. Published by Elsevier Ltd. All rights reserved. resulting bench and the outcomes will be presented at the end of this paper. Keywords: System Identification, Test Bench, Thruster Model. Keywords: Keywords: System System Identification, Identification, Test Test Bench, Bench, Thruster Thruster Model. Model. Keywords: System Identification, Test Bench, Thruster Model. Keywords: System Identification, Test Bench, Thruster Model. 1. INTRODUCTION the thruster and the measurement of the different physical 1. INTRODUCTION the thruster and the measurement of the different physical 1. INTRODUCTION the thruster and measurement of the different physical quantities needed. size of the test will depend 1. INTRODUCTION the thruster and the theThe measurement of thebench different physical quantities needed. The size of the test bench will depend 1. INTRODUCTION the thruster and the measurement of the different physical A thruster is an fundamental element in an autonomous quantities needed. The size of the test bench will depend on the size of the thruster being identified, its realization is quantities needed. The size of the test bench will depend A thruster is an fundamental element in an autonomous on the size of the thruster being identified, its realization is A thruster is fundamental element in an autonomous quantities needed. The size of the test bench will depend underwater oran remotely operated vehicle (AUV/ROV), it on the size of the thruster being identified, its realization is multidisciplinary because it requires skills in electrical and A thruster is an fundamental element in an autonomous on the size of the thruster being identified, its realization is underwater oran remotely operated vehicle (AUV/ROV), it multidisciplinary because it requires skills in electrical and A thruster fundamental element an autonomous underwater remotely vehicle (AUV/ROV), it on the size ofengineering. the thruster identified, its realization is consists of aisor motor and operated a propeller to in generate a thrust because it requires skills electrical and mechanical Abeing literature studyin made by Guibunderwater ormotor remotely operated vehicle (AUV/ROV), it multidisciplinary multidisciplinary becauseA itliterature requires skills in electrical and consists of a and a propeller to generate a thrust mechanical engineering. study made by Guibunderwater or remotely operated vehicle (AUV/ROV), it consists aa motor and aa(Aras propeller generate aa thrust multidisciplinary because itliterature requires skills in electrical and that can of move the vehicle et al. to (2013a)). Given that mechanical engineering. A study made by Guibert (2005), shows that several laboratories have developed consists of motor and propeller to generate thrust mechanical engineering. A literature study made by Guibthat can move the vehicle et al. to (2013a)). Given that ert (2005), shows that several laboratories have consists a ismotor andwithin a(Aras propeller generate a thrust that can move the vehicle (Aras et (2013a)). Given that mechanical engineering. A literature study madedeveloped by Guibthe thruster located the innermost control loop (2005), that laboratories have developed various testshows benches inseveral order to model thrusters. that can of move the vehicle (Aras et al. al. (2013a)). Given that ert ert (2005), shows thatin several laboratories have developed the thruster is located within the innermost control loop various test benches order to model thrusters. that can move the vehicle (Aras et al. (2013a)). Given that the thruster is located within the innermost control loop ert (2005), shows that several laboratories have developed of the vehicle control system, then it is essential in to order various test benches in order to model thrusters. the thruster located withinthen the itinnermost loop various test benches in order to model thrusters. of the the vehicleis control system, is essentialcontrol in to to order The organization of this article will be thrusters. as follows: a state the thruster iscontrol located withinthen the control loop The of vehicle system, it is in various test benches in order to will model to the achieve good manoeuvring control that the thruster organization of this article be as follows: a of vehicle control system, then itinnermost is essential essential in thruster to order order to achieve good manoeuvring control that the The organization of this article will be as follows: aa state state of the art review of the literature on the test benches The organization of this article will be as follows: state of the vehicle control system, then it is essential in to order to achieve good manoeuvring control that the thruster model accurately represents thecontrol thruster characteristics. of the art review of the literature on the test benches to achieve good manoeuvring that the thruster model accurately represents the thruster characteristics. The organization of this article will be as follows: a state of the art review of the literature on the test benches already realized will be introduced in Section 2 while of the art review of the literature on the test benches to achieve good manoeuvring control that the thruster model accurately represents the thruster characteristics. The subject of thruster modelling and control are still already realized will be introduced in Section 2 while modelsubject accurately represents the thruster characteristics. The of thruster modelling and control are still of the art review of the literature on the test benches already realized will be introduced in Section 2 while mentioning the advantages and disadvantages of each of already realized will be introduced in Sectionof 2each while model accurately represents thestudies thruster characteristics. The subject of thruster modelling and control are still the object of research in the mentioning the advantages advantages and disadvantages disadvantages of The subject ofvarious thruster modelling and published control are still the object of various research studies published in the already realized will be introduced in Section 2 while mentioning the and of each of them. In Section 3 we will make a comparison between the mentioning the advantages and disadvantages of each of The subject ofvarious thruster and control still the object of research studies published in the literature (Muljowidodo etmodelling al. (2009); Yang et al. are (2015)). them. In Section 3 we will make a comparison between the the object of various research studies published in the literature (Muljowidodo et al. (2009); Yang et al. (2015)). mentioning the advantages and aadisadvantages of and eachthe of In 33 we comparison between test benches based onwill themake measurement sensors them. In Section Section we will make comparison between the the of various studies published in the them. literature (Muljowidodo et Yang et (2015)). The object composition of anresearch underwater thruster isal. generally test benches based on the measurement sensors and the literature (Muljowidodo et al. al. (2009); (2009); Yang et is al.generally (2015)). The composition of an underwater thruster them. In Section 3 we will make a comparison between test benches based on the measurement sensors and the mechanical reaction frame. The description of the realized test benchesreaction based on the The measurement sensors and the literature (Muljowidodo et al.and (2009); Yang et (2015)). mechanical The of underwater is basedcomposition on an electrical motor the thruster thrust force depends frame. description of the realized The composition of an an underwater thruster isal.generally generally based on an electrical motor and the thrust force depends test benches based on the The measurement sensors and the mechanical reaction frame. description the realized bench will be discussed in Section 44of by detailing mechanical reaction frame. The description of the realized The composition of an underwater thruster is generally based on an electrical motor and the thrust force depends on the model of the motor, the propeller and other test bench will be discussed in Section by detailing based onmodel an electrical motor andthe the propeller thrust force depends on the of the motor, and other mechanical reaction frame. The description of the realized test bench will be discussed in Section 4 by detailing the mechanical part (tank, mechanical reaction frame and test bench will be discussed in Section 4 by detailing based an make electrical andmathematical the propeller thrust force depends on the model of the motor, the and other factors that obtaining the model more the mechanical part (tank, mechanical reaction frame and on theon model ofobtaining themotor motor, the propeller and other factors make the mathematical more test bench be (tank, discussed in Section 4 byacquisition detailing the mechanical part reaction frame and thruster) andwill electronic partmechanical (sensors and data the mechanical part (tank, mechanical reaction frame and on the that model ofobtaining the motor, the propeller model and other factors that make the mathematical model more complex. thruster) and electronic part (sensors and data acquisition factors that make obtaining the mathematical model more complex. the mechanical part (tank, mechanical reaction frame thruster) and electronic part (sensors and data acquisition system). In Section 5 we present the results obtained and thruster) and electronic part (sensors and data acquisition factors that make obtaining the mathematical model more complex. In Section 55 we present the results obtained and complex. thruster) and electronic (sensors and data acquisition To obtain a model of a thruster without going through system). system). In present the obtained and we make aa Section comparison with aa commercial thruster. In system). In Section 5 we wepart present the results results obtained and complex. To obtain a model of a thruster without going through we make comparison with commercial thruster. In To obtain a model of a thruster without going through system). In 5conclusions we with present the results obtained and a complex mathematical development, it isgoing necessary to we make aa Section comparison aa commercial thruster. In the final section the and perspectives will be To obtain a model of a thruster without through we make comparison with commercial thruster. In a complex mathematical development, it is necessary to the final section the conclusions and perspectives will be obtain amathematical modelto ofconstruct a thruster through aTo development, it is necessary to we make aand comparison with a commercial thruster. In identification thewithout model from measured final the and will presented discussed. ausecomplex complex mathematical development, it from isgoing necessary to the the final section section the conclusions conclusions and perspectives perspectives will be be use identification to construct the model measured presented and discussed. a complex mathematical development, it is necessary to use identification to construct the model from measured the final section the conclusions and perspectives will be input-output data that is obtained from a test bench. A presented and discussed. use identification construct the model input-output datato that is obtained from afrom test measured bench. A presented and discussed. use construct the model measured input-output that is from test bench. and discussed. test identification bench for data the to identification of underwater thrusters is presented 2. STATE OF ART ON THE TEST BENCHES input-output data that is obtained obtained from aafrom testthrusters bench. A A test bench for the identification of underwater is 2. STATE OF ART ON THE TEST BENCHES input-output data that is obtained from a test testthrusters bench. A test bench for the identification of underwater is 2. STATE OF ART generally constituted by (Guibert (2005)): tank filled test bench for the identification of underwater thrusters is 2. STATE OF ART ON ON THE THE TEST TEST BENCHES BENCHES generally constituted by (Guibert (2005)): a test tank filled test bench for the identification of underwater thrusters is 2. STATE OF ART ON THE TEST BENCHES generally constituted by (Guibert (2005)): a test tank filled with water, the system of propulsion (motor + propeller) The first underwater thruster test bench realized by Dygenerally constituted by (Guibert (2005)): a test tank filled with water, the system of propulsion (motor + propeller) The first underwater thruster test bench realized by Dygenerally constituted by (Guibert (2005)): a test tank filled with water, the system of propulsion (motor + propeller) The first underwater thruster test bench realized by and instrumentation which will make it possible to control namical System and Control Laboratory (DSCL) is Dydewithinstrumentation water, the system of propulsion (motor + propeller) The firstSystem underwater thruster test bench realized byis Dyand which will make it possible to control namical and Control Laboratory (DSCL) dewith water, the system of propulsion (motor + propeller) The first underwater thruster test bench realized by and deand instrumentation instrumentation which which will will make make it it possible possible to to control control namical namical System System and and Control Control Laboratory Laboratory (DSCL) (DSCL) is is Dydeand instrumentation which makePublished it possible to control System and Control Laboratory (DSCL) is de2405-8963 Copyright © 2019. Thewill Authors. by Elsevier Ltd. All namical rights reserved. Peer review under responsibility of International Federation of Automatic Control. 10.1016/j.ifacol.2019.12.316
Abdelmalek Laidani et al. / IFAC PapersOnLine 52-21 (2019) 254–259
scribed in Whitcom and Yoerger (1999), it consists of a cylinder tank with a depth of 3.5 m and a diameter of 5 m, a support structure composed of a fixed horizontal beam that overhangs the tank and a vertical beam with pivot connection that keeps the thruster in the tank. A force sensor is located in the other non-submerged end of the pivoting beam at the same distance as the thruster, therefore the sensor directly returns thrust measurements generated by the thruster. The following Fig. 1 shows the test bench:
255
Fig. 3. Test bench of DOE: testing SQUID AUV
Fig. 4. Test bench of CAUVR Fig. 1. DSCL test bench This test bench was made to test a thruster of 1 kWatts with a voltage of 120 Volts. This thruster is based on a brushless electrical motor and a propeller of 246 mm diameter. In the instrumentation part, a force sensor ranging from -960 to +960 Newton was used as well as an encoder with a resolution of 4096 Pulses/Revolution. This first test bench of DSCL could measure the force generated by the thruster in one direction only, to measure the force in the other direction it was necessary to reverse the disposition of the thruster. A second test bench was made to correct the deficiencies of the first (Bachmayer et al. (1999)). This test bench is composed of a vertical beam fixed by four brackets. A 6-axis force sensor is inserted between the thruster and the vertical beam (Fig. 2). This test stand is designed to support and test a wide variety of thrusters. At Florida Atlantic University’s Department of Ocean Engineering (DOE) (Guibert (2005)), tests on the SQUID Autonomous Uunderwater Vehicule (AUV) were conducted in a 26,500 liters test tank with a section of 1,7 m2 using an elastic element equipped with a strain gauge to measure the thrust produced (Fig. 3). Fig. 4 shows the test bench of the Centre of AUV Research (CAUVR) as used in Healey et al. (1995) and Cody (2003) to measure the thrust of a thruster to obtain its model.
Fig. 2. Second test bench of DSCL
Fig. 5. Test bench of DSL
Fig. 6. The THL 404 Thruster mounted on the LDTN’s Elastic Structure A simple triangular shaped structure with a force sensor placed in the arm is exposed to traction/compression from the thruster. In Cooke (1989) the Deep Submergence Laboratory (DSL) test bench is presented, it consists of a submerged structure in a tank of approximately 3758 liters. The thruster is mounted and compensated so that only the thrust can be exerted on the force sensor (see Fig. 5). To measure the thrust of the THL 404 thruster Deniellou et al. (1998), the Laboratoire de D´eveloppement des Technologies Nouvelles of ENSTA-Bretagne (LDTN) opted for the production of an elastic structure equipped with a displacement sensor instead of a strain gauge. The Fig. 6 shows the principle of thrust measurement.
256
Abdelmalek Laidani et al. / IFAC PapersOnLine 52-21 (2019) 254–259
Fig. 7. The IREENA’s test bench
Fig. 9. The CUMS’s test bench
Fig. 8. The BlueRobotics’s test bench for testing T100 Thruster The Institut de Recherche en Energie Electrique of Nantes Atlantique (IREENA) in collaboration with the Institut de Cr´eativit´e Industrielle (ICI) have made the realization of a test bench to allow the validation of models and control laws of thruster in the context of Guibert (2005) thesis. The Fig. 7 shows an isometric view of the realized structure. This test bench consists of a tank of 3 x 3 x 1.5 m for Length, Width and Height (L,W,H) respectively, a thruster based on a brushless IP67 motor which can be submerged in water, a fixed structure holding the thruster and a force sensor integrated in the mounting structure, as well as other sensors for measuring water velocity and hydrodynamic torque. BlueRobotics, a company specializing in the design and manufacture of components for underwater robotics, has developed a test bench to identify its T100 thruster (Thruster based brushless motor M100). The structure is L-shaped with one degree of freedom allowing rotation as showing in Fig. 8 (BlueRobotics (2019)). A weight placed on the horizontal part just above the force sensor makes it possible to measure the thrust in both directions of rotation without making any changes to the positioning of the thruster. At the Center for Unmanned System Studies (CUMS), a thruster has been designed to be implemented on their underwater robot. The thruster parameters were identified by the test bench shown in Fig. 9 (Muljowidodo et al. (2009)). In Juca et al. (2012) a test bench was presented and used for the identification of a thruster model mounted in the Remotely Operated Vehicle (LAURS ROV), the same bench was also used for identifying other thruster models from Tecnadyne Inc, see Fig. 10. 3. COMPARISON OF TESTS BENCHES Now we will summarize and compare all the characteristics of the various test benches described above. This comparison will provide us with a technical basis for the design selection and realization of our test bench. The points on which we are going to base these comparisons are given in Table 1 and Table 2.
Fig. 10. The LAURS ROV’s test bench We can notice from Table 1 that most of the test benches are equipped with strain gauge sensors for measuring the thrust, this is related to their simplicity of implementation and their cost which is less expensive then a 6-axis force sensor (DSL2 and IREENA) or displacement sensor (LDTN). Also, most test benches immerse the force sensor in the water in order to make a direct measurement of the thrust force, this choice requires some means of sealing the sensor which may inadvertently affect the sensor performance. Some test benches are equipped with additional sensors (e.g: velocimeter) for measuring the velocity of the fluid (DSCL2,DSL and IREENA), this makes it possible to identify other parameters possibly influencing the thrust measurements and thus improve the accuracy of the thruster model, but this will significantly increase the cost of the test bench. The structure comparison in Table 2 shows that there are two approaches, the first is the use of a fixed structure Table 1. Comparison of test bench measurement sensors Test Bench DSCL1 DSCL2 DoE CAUVR DSL LDTN IREENA BlueRobotics CUMS LAURS
Thrust Measurement Strain gauge 6-Axis force sensor Strain gauge Strain gauge Strain gauge Displacementsensor 6-Axis force sensor Strain gauge
Immersed Sensor No Yes
Velocimeter
No Yes Yes Yes
No No Yes No
Yes
Yes
No
No
Strain gauge Strain gauge
No No
No No
No Yes
Abdelmalek Laidani et al. / IFAC PapersOnLine 52-21 (2019) 254–259
257
that does not allow any degree of freedom, however the advantage of this solution is the measurement of the thrust force in both directions using a 6-axis load cell without changing the thruster configuration. The second is to use a structure that allows movement in one degree of freedom, the force sensor is removed out of the water so as to avoid the necessity of sealing the sensor, however the measurement of thrust is only in one direction. The cost of a test bench depends on the following parameters:
Fig. 11. Test bench used for identification
• The type and variety of sensors employed (strain gauge, 6 axis, displacement, tachometer and velocimeter). • The type of material used in the structure part. • The size of the thruster under test. • The dimensions of the test tank.
The objective of this comparison is to have a general overview of the technical choices and criteria that must be taken into consideration for the realization of a test bench destined for the identification or control of an underwater thruster. 4. DESCRIPTION OF THE TEST BENCH REALIZED Our goal is to build a low-cost test bench that will allow us to identify and control an underwater thruster that will also be low-cost, this thruster is designed and realized by us and will be integrated in our future underwater vehicle. The desired requirements for the test bench to be designed and realised are as follows: • The test bench will be used for the identification of small thrusters (<1kWatt), the tank must be of rectangular form to allow for the inclusion of a side observation window, • The test structure must be adapted to the tank, with low drag so as not to impact on the water flow, simple to construct, and able to measure the thrust in both directions, • A low-cost thruster with a ducted propeller and a sealed brushless motor, that is easy to build and to control, will be mounted to the test structure, • The use of a force and current sensor that can be interfaced to an acquisition card should be used for digitizing measurements. According to the specifications mentioned above, we were inspired by the BlueRobotics test bench, that is simple Table 2. Structure comparison of test benches Test Bench DSCL1 DSCL2 DoE CAUVR DSL LDTN IREENA BlueRobotics CUMS LAURS
Structure Type 1DoF Fixed Fixed 1DoF Fixed Fixed Fixed 1DoF
Thrust Direction 1 2 2 1 1 2 2 2
Thruster Power 1.000 kW 1.000 kW 1.000 kW 0.333 kW 0.360 kW THL 404 THL 404 T100/T200
Overall Cost >100 $ >1000 $ <100 $ <100 $ >100 $ >100 $ >1000 $ <100 $
1DoF 1DoF
1 1
0.400 kW 1.000 kW
<100 $ <100 $
Fig. 12. The thruster assembly and easy to build, destined for testing small thrusters, the supporting structure is (L) shaped form, and the entire structure is easily transportable. In the following paragraphs we will present each constituent element: 4.1 Test Tank The dimensions of the tank are: 800x325x415 mm for L, W and H, respectively. An opening on the side allows to observe the thruster during its immersion in water as shown in Fig. 11. The tank is designed for small thruster tests, therefore, its dimensions are sufficient to center the thruster so as not to impose a boundary wall effect (force feedback) on the measurement of the thrust. 4.2 Support Structure It is composed of a support of L-shaped form, with a vertical section of 320 mm immersed on which the thruster will be fixed and another horizontal section of 220 mm that will transfer the force onto the force sensor located outside the tank. This support allows the rotational movement around an axis of 8 mm and fixed by bearings. 4.3 Propulsion The thruster that we want to identify was designed by us. It is based on an integrated brushless motor, reference A2212. The fairing of the thruster and its 3 bladed propeller are made by a 3D printer. Table 3 shows the dimensions of the thruster and Fig. 12 shows its assembly. Table 3. Thruster Dimensions Diameter Length Blades Material
Propeller 66 mm 20 mm 3 ABS
Fairing/Duct 70 mm 98 mm – ABS
4.4 Electronic Components A 5 kg strain gauge load cell is used to measure the thrust force generated. The principle of measurement with
Abdelmalek Laidani et al. / IFAC PapersOnLine 52-21 (2019) 254–259
258
Fig. 14. Thrust force generated by the thruster.
Fig. 13. Schematic of the developed test bench. a strain gauge is based on a Wheatstone bridge with variable resistors that are sensitive to deformation of the load cell. The variation of the voltage at the bridge is very small, so it is necessary to use an amplifier. A circuit which integrates an amplifier and a converter with a resolution of 24 bits (HX711) was used to make the sampling and digitization of the thrust force, then transmit it in numerical format to an acquisition card. The current absorbed by the thruster is measured by an ACS712 Hall effect current sensor, it produces an analog output that is measured by the Analog to Digital Converters (ADC) of the acquisition card. An Electronic Speed Controller (ESC) of 30 A is used to control the thruster. Fig. 13 shows the schematic diagram that explains the acquisition system. Table 4 shows the components used in the realization of the test bench and their approximate prices. The total price of the bench does not exceed 100$, which makes it low-cost and the results that will be presented in the next section will show its efficiency. Table 4. Itemised list and costs of constituent parts used to build the test bench Items Power Supply Acquisition Card Force Sensor 24 bits ADC Current Sensor Support Structure Other accessories Water Tank Overall Cost
Type ATX Arduino 5 kg Strain gauge HX711 ACS712 30A Stainless Steel Wires Recovery Article
Approx. Price 20 $ 10 $ 05 $ 03 $ 05 $ 10 $ 10 $ 00 $ 70 $
5. RESULTS AND COMPARISONS 5.1 Results To characterise the thruster we need to obtain the measurements concerning the thrust force generated and the corresponding current consumed by the the thruster,
Fig. 15. Current consumed by thruster for forward/reverse directions. across the speed range of the thruster from 0 to 100% which corresponds to a Pulse Width Modulation (PWM) from 1000 to 2000 µs. Fig. 14 shows the thrust force curve generated in the forward and reverse directions. We notice that there are 3 operating zones: the dead zone in which the thruster doesn’t produce force and is of the order of 50 us, then a linear zone which is the operating zone between 50 and 1800 us, but beyond this zone the thruster enters into a saturation zone where the thrust force remains constant despite the increase of the PWM signal. The thrust force generated in the forward and reverse direction is approximately 12.25 and 10.78 Newton respectively which is equivalent to 1.25 and 1.10 Kgf. The current consumed by the thruster is shown in Fig. 15, we note that the maximum current reaches 21 A which is equivalent to a power of about 250 Watts under a supply voltage of 12 Volts. In Fig. 16 the curve shows the efficiency of the thruster, where efficiency can be determined from the thrust force generated with respect to the power consumed. We note that the efficiency in the forward direction is 0.28 Newton/Watt which is greater than in the reverse direction which is equal to 0.15 Newton/Watt. Fig. 17 shows that the relationship between current and thrust force has a quadratic form and this implies that the Torque-Thrust relationship is also quadratic because the torque is proportional to the current.
Abdelmalek Laidani et al. / IFAC PapersOnLine 52-21 (2019) 254–259
259
constructed by us using a 3D printing technique. The curve presented in Fig. 14 can be transformed into a polynomial of N degrees by neglecting the higher order dynamics of the thruster. This polynomial will be integrated in the mathematical model of an underwater vehicle in order to get a simulation behavior closer to the reality so that the control laws can be validated before they are implemented on the vehicle. The comparison in Table 5 shows that the thruster that we have realized and tested using the test bench described above, does not perform as well as the T100 but its performance is nevertheless satisfactory given its cost. It must be taken into account that the respective tests are made using different test benches, and that the T100 values used in the comparison are values obtained from the manufacturer’s data sheet (BlueRobotics (2019)).
Fig. 16. Thruster efficiency.
REFERENCES
Fig. 17. Current vs. Thrust. 5.2 Comparison In order to validate the thruster that we have built and analyzed with the test bench, we will make a comparison with another thruster that is supposed to have similar performances. This thruster is the T100 from BlueRobotics (BlueRobotics (2019)), The Table 5 shows the main features of the comparison. Table 5. Comparison with T100 Thruster Electrical/Physical /Performance Max Forward Thrust Max Reverse Thrust Min Thrust Rotational Speed Operating Voltage Max Current Max Power Length Diameter Weight Propeller Diameter Cost(without ESC) Cost(with ESC)
T100
Our Thruster
2.36 kgf 1.85 kgf 0.01 kgf 300-4200 rev/min 6-16 Volts 12.5 Amps 136 Watts 102 mm 100 mm 295 gr 76 mm 119 $ 144 $
1.25 kgf (12.25 N) 1.10 kgf (10.78 N) 0.05 kgf (0.049 N) 700-15000 rev/min 12 Volts 21.0 Amps 250 Watts 98 mm 70 mm 152 gr 66 mm 45 $ 65 $
6. CONCLUSION We have presented in this article a low-cost test bench destined for the identification of an underwater thruster, this bench was realized by taking into account the advantages and disadvantages of existing test benches. The thruster that we have tested is a thruster designed and
Aras, M., Abdullah, S., A.Rahman, and Aziz, M. (2013a). Thruster modeling for underwater vehicle using system identification method. International Journal of Advanced Robotic Systems, 10, 252–264. Bachmayer, R., Whitcomb, L., and Grosenbaugh, M. (1999). Development, comparison, and preliminary experimental validation of nonlinear dynamic thruster models. IEEE Journal of Oceanic Engineering, 24, 481– 494. BlueRobotics (2019). BlueRobotics test bench. https:// www.bluerobotics.com. Cody, S. (2003). An Experimental Study of the Response of Small Tunnel Thrusters to Triangular and Square Wave Inputs. Master’s thesis, Naval Postgraduate School, Monterey, Canada. Cooke, J. (1989). Incorporating Thruster Dynamics in the Control of an Underwater Vehicle. Master’s thesis, MIT, USA. Deniellou, L., Gallou, Y., Gourmelen, P., and Seube, N. (1998). Force control of underwater thrusters with application to auv motion control. IEEE Oceanic Engineering Society, 2. Guibert, C. (2005). Modelisation et commande en pous´ee de moteurs ` a courant alternatifs en propulsion navale. Ph.D. thesis, Universit´e de Nantes. Healey, A., Rock, S., Cody, S., Miles, D., and Brown, J. (1995). Toward an improved understanding of thruster dynamics for underwater vehicles. IEEE Journal of Oceanic Engineering, 20, 354–361. Juca, J., Saito, M., CezarAdamowski, J., Takase, F., and Maruyama, N. (2012). Modelling and identification of yaw motion of an open-frame underwater robot. Journal of Intelligent and Robotic Systems, 66, 37–56. Muljowidodo, K., Adi, N., Prayogo, N., and Budiyono, A. (2009). Design and testing of underwater thruster for shrimp rov-itb. Indian Journal of Marine Sciences, 38, 232–237. Whitcom, L. and Yoerger, D. (1999). Development, comparison, and preliminary experimental validation of nonlinear dynamic thruster models. IEEE Journal of Oceanic Engineering, 24, 481–494. Yang, R., Clement, B., Mansour, A., Li, M., and Wu, N. (2015). Modeling of a complex-shaped underwater vehicle for robust control scheme. Journal of Intelligent and Robotic Systems.
ScienceDirect A A A A
IFAC PapersOnLine 52-21 (2019) 254–259
Low-Cost Low-Cost Test Test Bench Bench for for Underwater Underwater Low-Cost Test Bench for Underwater Thruster Identification Low-Cost Test Bench for Underwater Thruster Identification Thruster Identification Thruster Identification ∗ Abdelmalek LAIDANI ∗,∗∗ ∗,∗∗ Mohamed BOUHAMIDA ∗
Abdelmalek LAIDANI Mohamed BOUHAMIDA ∗,∗∗ ∗ ∗ ∗∗∗ ∗,∗∗ Mohamed ∗ Abdelmalek LAIDANI BOUHAMIDA Mustapha BENGHANEM ∗ Karl SAMMUT ∗∗∗ Abdelmalek LAIDANI Mohamed BOUHAMIDA Mustapha BENGHANEM Karl SAMMUT ∗,∗∗ ∗ ∗ ∗∗∗ ∗∗,∗∗∗ ∗ Karl ∗∗∗ Abdelmalek LAIDANI Mohamed BOUHAMIDA Mustapha BENGHANEM SAMMUT Benoit CLEMENT ∗∗,∗∗∗ Mustapha BENGHANEM Karl SAMMUT ∗∗∗ Benoit CLEMENT ∗ ∗∗,∗∗∗ ∗∗,∗∗∗ Mustapha BENGHANEM Karl SAMMUT Benoit CLEMENT CLEMENT Benoit ∗ Benoit CLEMENT ∗∗,∗∗∗ ∗ AVCIS Research Laboratory, Department of Automatics, USTO-MB, ∗ AVCIS Research Laboratory, Department of Automatics, USTO-MB, ∗ AVCIS Research Laboratory, Department Oran, Algeria, (e-mail: [email protected]). AVCIS Research Laboratory, Department of of Automatics, Automatics, USTO-MB, USTO-MB, ∗ ∗∗ Oran, Algeria, (e-mail: [email protected]). AVCIS Research Laboratory, Department of Automatics, USTO-MB, Oran, Algeria, (e-mail: [email protected]). Lab-STICC, UMRCNRS 6285, ENSTA-Bretagne, 29806 Brest ∗∗ Oran, Algeria, (e-mail: [email protected]). Lab-STICC, UMRCNRS 6285, ENSTA-Bretagne, 29806 Brest ∗∗ ∗∗ Cedex Oran, Algeria, (e-mail: [email protected]). Lab-STICC, UMRCNRS 6285, ENSTA-Bretagne, 29806 9, France (e-mail: [email protected]). Lab-STICC, UMRCNRS 6285, ENSTA-Bretagne, 29806 Brest Brest Cedex 9, France (e-mail: [email protected]). ∗∗∗∗∗ Lab-STICC, UMRCNRS 6285, ENSTA-Bretagne, 29806 Brest Cedex 9, France France (e-mail: [email protected]). Flinders University, Adelaide, SA 5042, Australia (e-mail: ∗∗∗ Cedex 9, (e-mail: [email protected]). Flinders University, Adelaide, SA 5042, Australia (e-mail: ∗∗∗ ∗∗∗ Cedex 9, France (e-mail: [email protected]). Flinders University, Adelaide, SA 5042, Australia (e-mail: [email protected]). Flinders University, Adelaide, SA 5042, Australia (e-mail: [email protected]). ∗∗∗ Flinders University, Adelaide, SA 5042, Australia (e-mail: [email protected]). [email protected]). [email protected]). Abstract: Underwater vehicles are increasingly being used for the inspection of marine strucAbstract: Underwater vehicles are increasingly being used for the inspection of marine strucAbstract: Underwater vehicles are being used the of structures, bathymetry and underwater intervention applications. manoeuvring performance Abstract: Underwater vehicles are increasingly increasingly being used for for The the inspection inspection of marine marine structures, bathymetry and underwater intervention applications. The manoeuvring performance Abstract: Underwater vehicles are increasingly usedoffor thethruster inspection of marine tures, bathymetry and underwater intervention applications. The manoeuvring performance of these vehicles isand critically dependant on thebeing quality the model which strucmust tures, bathymetry underwater intervention applications. The manoeuvring performance of these vehicles is critically dependant on the quality of the thruster model which must tures, bathymetry and underwater intervention applications. Thethruster manoeuvring performance of these vehicles critically dependant on the quality of the model which must faithfully representis the relationship between thruster inputs and corresponding outputs. of these vehicles is critically dependant on the quality of the thruster model which must corresponding outputs. faithfully represent the relationship between the thruster and of these vehicles critically dependant on quality ofinputs the thruster model which must faithfully represent relationship between inputs and outputs. Thruster manufacturers typically provide onlythe thethruster essential electrical, mechanical and thrust faithfully representis the the relationship between the thruster inputs and corresponding corresponding outputs. Thruster manufacturers typically provide only the essential electrical, mechanical and thrust faithfully represent the relationship between the thruster inputs and corresponding outputs. Thruster manufacturers typically provide only the essential electrical, mechanical and thrust performance characteristics, for this reason the identification of the thruster’s parameters is Thruster manufacturers typically provide only essential electrical, mechanical and thrust performance characteristics, for this this reason thethe identification of the the thruster’s thruster’s parameters is Thruster provide only the essential electrical, mechanical andvehicle. thrust performance characteristics, for reason the identification of parameters is especially manufacturers important for typically constructing a reliable control system for thruster’s the underwater performance characteristics, for this reason the identification of the parameters is especially important for constructing a reliable control system for the underwater vehicle. performance characteristics, for the is identification of the parameters is especially important constructing aa reliable control system for the underwater vehicle. Physical testing usingfor some formthis of reason test bench normally used for thruster’s identifying the thruster especially important for constructing reliable control system for the underwater vehicle. the thruster Physical testing using some form of test bench is normally used for identifying especially important for constructing a reliable control system for the underwater vehicle. Physical testing using some form of test bench is normally used for identifying the thruster parameters. These test benches usually consist of a water tank and the thruster which is clamped Physical testing using some form of test bench normally usedthe forthruster identifying the thruster parameters. These test benches usually consist of ais tank and which is clamped Physical testing using some form of test bench iswater normally used forthruster identifying the thruster parameters. These test usually consist of water tank and the which is to some form of instrumented reaction frame for model parameters. The cost of parameters. These test benches benches usually consist of aameasuring water tankthe and the thruster which is clamped clamped to some form of instrumented reaction frame for measuring the model parameters. The cost of of parameters. test benchesto usually consist of ameasuring water and the is clamped to some formThese of instrumented instrumented reaction framefactors for measuring the model parameters. The cost a test bench varies according two major whichtank arethe themodel sizethruster and thewhich instrumentation to some form of reaction frame for parameters. The cost of a test bench varies according to two major factors which are the size and the instrumentation to some formvaries ofdepends instrumented reaction frame for measuring the model parameters. of a test bench according two factors which the size the which strongly on theto physical quantities that we are are trying toand measure. InThe this cost article a test bench varies according to two major major factors which are the sizeto and the instrumentation instrumentation which strongly depends on the physical quantities that we are trying measure. In this article a test bench varies according to two major factors which are the size and the instrumentation which strongly depends on the physical quantities that we are trying to measure. In this article we will make a state of the art survey of existing test benches in order to provide a technical which strongly depends on the physical quantities that we are trying to measure. In athis article to provide technical we will make a state of the art survey of existing test benches in order which strongly the quantities that are trying to measure. In aathis article we will make a state of the art survey of existing test benches in to provide technical basis that will allow to build a low-cost test bench. A thruster will be using the we will make a depends state us of on the artphysical survey of existing test we benches in order order to modelled provide technical basis that will allow us to build a low-cost test bench. A thruster will be modelled using the we will make a allow stateand of the art survey of existing test benches in order topaper. provide a using technical basis that willbench allow us to build a low-cost low-cost test bench. A thruster will be modelled using the resulting test thebuild outcomes will be presented at thruster the end of this basis that will us to a test bench. A will be modelled the resulting test bench and the outcomes will be presented at the end of this paper. basis thattest willbench allowand us to a low-cost test bench. A will bepaper. modelled using the resulting the outcomes will presented at the this resulting test bench and thebuild outcomes will be be presented at thruster the end end of of this paper. Copyright ©test 2019. The Authors. Published by Elsevier Ltd. All rights reserved. resulting bench and the outcomes will be presented at the end of this paper. Keywords: System Identification, Test Bench, Thruster Model. Keywords: Keywords: System System Identification, Identification, Test Test Bench, Bench, Thruster Thruster Model. Model. Keywords: System Identification, Test Bench, Thruster Model. Keywords: System Identification, Test Bench, Thruster Model. 1. INTRODUCTION the thruster and the measurement of the different physical 1. INTRODUCTION the thruster and the measurement of the different physical 1. INTRODUCTION the thruster and measurement of the different physical quantities needed. size of the test will depend 1. INTRODUCTION the thruster and the theThe measurement of thebench different physical quantities needed. The size of the test bench will depend 1. INTRODUCTION the thruster and the measurement of the different physical A thruster is an fundamental element in an autonomous quantities needed. The size of the test bench will depend on the size of the thruster being identified, its realization is quantities needed. The size of the test bench will depend A thruster is an fundamental element in an autonomous on the size of the thruster being identified, its realization is A thruster is fundamental element in an autonomous quantities needed. The size of the test bench will depend underwater oran remotely operated vehicle (AUV/ROV), it on the size of the thruster being identified, its realization is multidisciplinary because it requires skills in electrical and A thruster is an fundamental element in an autonomous on the size of the thruster being identified, its realization is underwater oran remotely operated vehicle (AUV/ROV), it multidisciplinary because it requires skills in electrical and A thruster fundamental element an autonomous underwater remotely vehicle (AUV/ROV), it on the size ofengineering. the thruster identified, its realization is consists of aisor motor and operated a propeller to in generate a thrust because it requires skills electrical and mechanical Abeing literature studyin made by Guibunderwater ormotor remotely operated vehicle (AUV/ROV), it multidisciplinary multidisciplinary becauseA itliterature requires skills in electrical and consists of a and a propeller to generate a thrust mechanical engineering. study made by Guibunderwater or remotely operated vehicle (AUV/ROV), it consists aa motor and aa(Aras propeller generate aa thrust multidisciplinary because itliterature requires skills in electrical and that can of move the vehicle et al. to (2013a)). Given that mechanical engineering. A study made by Guibert (2005), shows that several laboratories have developed consists of motor and propeller to generate thrust mechanical engineering. A literature study made by Guibthat can move the vehicle et al. to (2013a)). Given that ert (2005), shows that several laboratories have consists a ismotor andwithin a(Aras propeller generate a thrust that can move the vehicle (Aras et (2013a)). Given that mechanical engineering. A literature study madedeveloped by Guibthe thruster located the innermost control loop (2005), that laboratories have developed various testshows benches inseveral order to model thrusters. that can of move the vehicle (Aras et al. al. (2013a)). Given that ert ert (2005), shows thatin several laboratories have developed the thruster is located within the innermost control loop various test benches order to model thrusters. that can move the vehicle (Aras et al. (2013a)). Given that the thruster is located within the innermost control loop ert (2005), shows that several laboratories have developed of the vehicle control system, then it is essential in to order various test benches in order to model thrusters. the thruster located withinthen the itinnermost loop various test benches in order to model thrusters. of the the vehicleis control system, is essentialcontrol in to to order The organization of this article will be thrusters. as follows: a state the thruster iscontrol located withinthen the control loop The of vehicle system, it is in various test benches in order to will model to the achieve good manoeuvring control that the thruster organization of this article be as follows: a of vehicle control system, then itinnermost is essential essential in thruster to order order to achieve good manoeuvring control that the The organization of this article will be as follows: aa state state of the art review of the literature on the test benches The organization of this article will be as follows: state of the vehicle control system, then it is essential in to order to achieve good manoeuvring control that the thruster model accurately represents thecontrol thruster characteristics. of the art review of the literature on the test benches to achieve good manoeuvring that the thruster model accurately represents the thruster characteristics. The organization of this article will be as follows: a state of the art review of the literature on the test benches already realized will be introduced in Section 2 while of the art review of the literature on the test benches to achieve good manoeuvring control that the thruster model accurately represents the thruster characteristics. The subject of thruster modelling and control are still already realized will be introduced in Section 2 while modelsubject accurately represents the thruster characteristics. The of thruster modelling and control are still of the art review of the literature on the test benches already realized will be introduced in Section 2 while mentioning the advantages and disadvantages of each of already realized will be introduced in Sectionof 2each while model accurately represents thestudies thruster characteristics. The subject of thruster modelling and control are still the object of research in the mentioning the advantages advantages and disadvantages disadvantages of The subject ofvarious thruster modelling and published control are still the object of various research studies published in the already realized will be introduced in Section 2 while mentioning the and of each of them. In Section 3 we will make a comparison between the mentioning the advantages and disadvantages of each of The subject ofvarious thruster and control still the object of research studies published in the literature (Muljowidodo etmodelling al. (2009); Yang et al. are (2015)). them. In Section 3 we will make a comparison between the the object of various research studies published in the literature (Muljowidodo et al. (2009); Yang et al. (2015)). mentioning the advantages and aadisadvantages of and eachthe of In 33 we comparison between test benches based onwill themake measurement sensors them. In Section Section we will make comparison between the the of various studies published in the them. literature (Muljowidodo et Yang et (2015)). The object composition of anresearch underwater thruster isal. generally test benches based on the measurement sensors and the literature (Muljowidodo et al. al. (2009); (2009); Yang et is al.generally (2015)). The composition of an underwater thruster them. In Section 3 we will make a comparison between test benches based on the measurement sensors and the mechanical reaction frame. The description of the realized test benchesreaction based on the The measurement sensors and the literature (Muljowidodo et al.and (2009); Yang et (2015)). mechanical The of underwater is basedcomposition on an electrical motor the thruster thrust force depends frame. description of the realized The composition of an an underwater thruster isal.generally generally based on an electrical motor and the thrust force depends test benches based on the The measurement sensors and the mechanical reaction frame. description the realized bench will be discussed in Section 44of by detailing mechanical reaction frame. The description of the realized The composition of an underwater thruster is generally based on an electrical motor and the thrust force depends on the model of the motor, the propeller and other test bench will be discussed in Section by detailing based onmodel an electrical motor andthe the propeller thrust force depends on the of the motor, and other mechanical reaction frame. The description of the realized test bench will be discussed in Section 4 by detailing the mechanical part (tank, mechanical reaction frame and test bench will be discussed in Section 4 by detailing based an make electrical andmathematical the propeller thrust force depends on the model of the motor, the and other factors that obtaining the model more the mechanical part (tank, mechanical reaction frame and on theon model ofobtaining themotor motor, the propeller and other factors make the mathematical more test bench be (tank, discussed in Section 4 byacquisition detailing the mechanical part reaction frame and thruster) andwill electronic partmechanical (sensors and data the mechanical part (tank, mechanical reaction frame and on the that model ofobtaining the motor, the propeller model and other factors that make the mathematical model more complex. thruster) and electronic part (sensors and data acquisition factors that make obtaining the mathematical model more complex. the mechanical part (tank, mechanical reaction frame thruster) and electronic part (sensors and data acquisition system). In Section 5 we present the results obtained and thruster) and electronic part (sensors and data acquisition factors that make obtaining the mathematical model more complex. In Section 55 we present the results obtained and complex. thruster) and electronic (sensors and data acquisition To obtain a model of a thruster without going through system). system). In present the obtained and we make aa Section comparison with aa commercial thruster. In system). In Section 5 we wepart present the results results obtained and complex. To obtain a model of a thruster without going through we make comparison with commercial thruster. In To obtain a model of a thruster without going through system). In 5conclusions we with present the results obtained and a complex mathematical development, it isgoing necessary to we make aa Section comparison aa commercial thruster. In the final section the and perspectives will be To obtain a model of a thruster without through we make comparison with commercial thruster. In a complex mathematical development, it is necessary to the final section the conclusions and perspectives will be obtain amathematical modelto ofconstruct a thruster through aTo development, it is necessary to we make aand comparison with a commercial thruster. In identification thewithout model from measured final the and will presented discussed. ausecomplex complex mathematical development, it from isgoing necessary to the the final section section the conclusions conclusions and perspectives perspectives will be be use identification to construct the model measured presented and discussed. a complex mathematical development, it is necessary to use identification to construct the model from measured the final section the conclusions and perspectives will be input-output data that is obtained from a test bench. A presented and discussed. use identification construct the model input-output datato that is obtained from afrom test measured bench. A presented and discussed. use construct the model measured input-output that is from test bench. and discussed. test identification bench for data the to identification of underwater thrusters is presented 2. STATE OF ART ON THE TEST BENCHES input-output data that is obtained obtained from aafrom testthrusters bench. A A test bench for the identification of underwater is 2. STATE OF ART ON THE TEST BENCHES input-output data that is obtained from a test testthrusters bench. A test bench for the identification of underwater is 2. STATE OF ART generally constituted by (Guibert (2005)): tank filled test bench for the identification of underwater thrusters is 2. STATE OF ART ON ON THE THE TEST TEST BENCHES BENCHES generally constituted by (Guibert (2005)): a test tank filled test bench for the identification of underwater thrusters is 2. STATE OF ART ON THE TEST BENCHES generally constituted by (Guibert (2005)): a test tank filled with water, the system of propulsion (motor + propeller) The first underwater thruster test bench realized by Dygenerally constituted by (Guibert (2005)): a test tank filled with water, the system of propulsion (motor + propeller) The first underwater thruster test bench realized by Dygenerally constituted by (Guibert (2005)): a test tank filled with water, the system of propulsion (motor + propeller) The first underwater thruster test bench realized by and instrumentation which will make it possible to control namical System and Control Laboratory (DSCL) is Dydewithinstrumentation water, the system of propulsion (motor + propeller) The firstSystem underwater thruster test bench realized byis Dyand which will make it possible to control namical and Control Laboratory (DSCL) dewith water, the system of propulsion (motor + propeller) The first underwater thruster test bench realized by and deand instrumentation instrumentation which which will will make make it it possible possible to to control control namical namical System System and and Control Control Laboratory Laboratory (DSCL) (DSCL) is is Dydeand instrumentation which makePublished it possible to control System and Control Laboratory (DSCL) is de2405-8963 Copyright © 2019. Thewill Authors. by Elsevier Ltd. All namical rights reserved. Peer review under responsibility of International Federation of Automatic Control. 10.1016/j.ifacol.2019.12.316
Abdelmalek Laidani et al. / IFAC PapersOnLine 52-21 (2019) 254–259
scribed in Whitcom and Yoerger (1999), it consists of a cylinder tank with a depth of 3.5 m and a diameter of 5 m, a support structure composed of a fixed horizontal beam that overhangs the tank and a vertical beam with pivot connection that keeps the thruster in the tank. A force sensor is located in the other non-submerged end of the pivoting beam at the same distance as the thruster, therefore the sensor directly returns thrust measurements generated by the thruster. The following Fig. 1 shows the test bench:
255
Fig. 3. Test bench of DOE: testing SQUID AUV
Fig. 4. Test bench of CAUVR Fig. 1. DSCL test bench This test bench was made to test a thruster of 1 kWatts with a voltage of 120 Volts. This thruster is based on a brushless electrical motor and a propeller of 246 mm diameter. In the instrumentation part, a force sensor ranging from -960 to +960 Newton was used as well as an encoder with a resolution of 4096 Pulses/Revolution. This first test bench of DSCL could measure the force generated by the thruster in one direction only, to measure the force in the other direction it was necessary to reverse the disposition of the thruster. A second test bench was made to correct the deficiencies of the first (Bachmayer et al. (1999)). This test bench is composed of a vertical beam fixed by four brackets. A 6-axis force sensor is inserted between the thruster and the vertical beam (Fig. 2). This test stand is designed to support and test a wide variety of thrusters. At Florida Atlantic University’s Department of Ocean Engineering (DOE) (Guibert (2005)), tests on the SQUID Autonomous Uunderwater Vehicule (AUV) were conducted in a 26,500 liters test tank with a section of 1,7 m2 using an elastic element equipped with a strain gauge to measure the thrust produced (Fig. 3). Fig. 4 shows the test bench of the Centre of AUV Research (CAUVR) as used in Healey et al. (1995) and Cody (2003) to measure the thrust of a thruster to obtain its model.
Fig. 2. Second test bench of DSCL
Fig. 5. Test bench of DSL
Fig. 6. The THL 404 Thruster mounted on the LDTN’s Elastic Structure A simple triangular shaped structure with a force sensor placed in the arm is exposed to traction/compression from the thruster. In Cooke (1989) the Deep Submergence Laboratory (DSL) test bench is presented, it consists of a submerged structure in a tank of approximately 3758 liters. The thruster is mounted and compensated so that only the thrust can be exerted on the force sensor (see Fig. 5). To measure the thrust of the THL 404 thruster Deniellou et al. (1998), the Laboratoire de D´eveloppement des Technologies Nouvelles of ENSTA-Bretagne (LDTN) opted for the production of an elastic structure equipped with a displacement sensor instead of a strain gauge. The Fig. 6 shows the principle of thrust measurement.
256
Abdelmalek Laidani et al. / IFAC PapersOnLine 52-21 (2019) 254–259
Fig. 7. The IREENA’s test bench
Fig. 9. The CUMS’s test bench
Fig. 8. The BlueRobotics’s test bench for testing T100 Thruster The Institut de Recherche en Energie Electrique of Nantes Atlantique (IREENA) in collaboration with the Institut de Cr´eativit´e Industrielle (ICI) have made the realization of a test bench to allow the validation of models and control laws of thruster in the context of Guibert (2005) thesis. The Fig. 7 shows an isometric view of the realized structure. This test bench consists of a tank of 3 x 3 x 1.5 m for Length, Width and Height (L,W,H) respectively, a thruster based on a brushless IP67 motor which can be submerged in water, a fixed structure holding the thruster and a force sensor integrated in the mounting structure, as well as other sensors for measuring water velocity and hydrodynamic torque. BlueRobotics, a company specializing in the design and manufacture of components for underwater robotics, has developed a test bench to identify its T100 thruster (Thruster based brushless motor M100). The structure is L-shaped with one degree of freedom allowing rotation as showing in Fig. 8 (BlueRobotics (2019)). A weight placed on the horizontal part just above the force sensor makes it possible to measure the thrust in both directions of rotation without making any changes to the positioning of the thruster. At the Center for Unmanned System Studies (CUMS), a thruster has been designed to be implemented on their underwater robot. The thruster parameters were identified by the test bench shown in Fig. 9 (Muljowidodo et al. (2009)). In Juca et al. (2012) a test bench was presented and used for the identification of a thruster model mounted in the Remotely Operated Vehicle (LAURS ROV), the same bench was also used for identifying other thruster models from Tecnadyne Inc, see Fig. 10. 3. COMPARISON OF TESTS BENCHES Now we will summarize and compare all the characteristics of the various test benches described above. This comparison will provide us with a technical basis for the design selection and realization of our test bench. The points on which we are going to base these comparisons are given in Table 1 and Table 2.
Fig. 10. The LAURS ROV’s test bench We can notice from Table 1 that most of the test benches are equipped with strain gauge sensors for measuring the thrust, this is related to their simplicity of implementation and their cost which is less expensive then a 6-axis force sensor (DSL2 and IREENA) or displacement sensor (LDTN). Also, most test benches immerse the force sensor in the water in order to make a direct measurement of the thrust force, this choice requires some means of sealing the sensor which may inadvertently affect the sensor performance. Some test benches are equipped with additional sensors (e.g: velocimeter) for measuring the velocity of the fluid (DSCL2,DSL and IREENA), this makes it possible to identify other parameters possibly influencing the thrust measurements and thus improve the accuracy of the thruster model, but this will significantly increase the cost of the test bench. The structure comparison in Table 2 shows that there are two approaches, the first is the use of a fixed structure Table 1. Comparison of test bench measurement sensors Test Bench DSCL1 DSCL2 DoE CAUVR DSL LDTN IREENA BlueRobotics CUMS LAURS
Thrust Measurement Strain gauge 6-Axis force sensor Strain gauge Strain gauge Strain gauge Displacementsensor 6-Axis force sensor Strain gauge
Immersed Sensor No Yes
Velocimeter
No Yes Yes Yes
No No Yes No
Yes
Yes
No
No
Strain gauge Strain gauge
No No
No No
No Yes
Abdelmalek Laidani et al. / IFAC PapersOnLine 52-21 (2019) 254–259
257
that does not allow any degree of freedom, however the advantage of this solution is the measurement of the thrust force in both directions using a 6-axis load cell without changing the thruster configuration. The second is to use a structure that allows movement in one degree of freedom, the force sensor is removed out of the water so as to avoid the necessity of sealing the sensor, however the measurement of thrust is only in one direction. The cost of a test bench depends on the following parameters:
Fig. 11. Test bench used for identification
• The type and variety of sensors employed (strain gauge, 6 axis, displacement, tachometer and velocimeter). • The type of material used in the structure part. • The size of the thruster under test. • The dimensions of the test tank.
The objective of this comparison is to have a general overview of the technical choices and criteria that must be taken into consideration for the realization of a test bench destined for the identification or control of an underwater thruster. 4. DESCRIPTION OF THE TEST BENCH REALIZED Our goal is to build a low-cost test bench that will allow us to identify and control an underwater thruster that will also be low-cost, this thruster is designed and realized by us and will be integrated in our future underwater vehicle. The desired requirements for the test bench to be designed and realised are as follows: • The test bench will be used for the identification of small thrusters (<1kWatt), the tank must be of rectangular form to allow for the inclusion of a side observation window, • The test structure must be adapted to the tank, with low drag so as not to impact on the water flow, simple to construct, and able to measure the thrust in both directions, • A low-cost thruster with a ducted propeller and a sealed brushless motor, that is easy to build and to control, will be mounted to the test structure, • The use of a force and current sensor that can be interfaced to an acquisition card should be used for digitizing measurements. According to the specifications mentioned above, we were inspired by the BlueRobotics test bench, that is simple Table 2. Structure comparison of test benches Test Bench DSCL1 DSCL2 DoE CAUVR DSL LDTN IREENA BlueRobotics CUMS LAURS
Structure Type 1DoF Fixed Fixed 1DoF Fixed Fixed Fixed 1DoF
Thrust Direction 1 2 2 1 1 2 2 2
Thruster Power 1.000 kW 1.000 kW 1.000 kW 0.333 kW 0.360 kW THL 404 THL 404 T100/T200
Overall Cost >100 $ >1000 $ <100 $ <100 $ >100 $ >100 $ >1000 $ <100 $
1DoF 1DoF
1 1
0.400 kW 1.000 kW
<100 $ <100 $
Fig. 12. The thruster assembly and easy to build, destined for testing small thrusters, the supporting structure is (L) shaped form, and the entire structure is easily transportable. In the following paragraphs we will present each constituent element: 4.1 Test Tank The dimensions of the tank are: 800x325x415 mm for L, W and H, respectively. An opening on the side allows to observe the thruster during its immersion in water as shown in Fig. 11. The tank is designed for small thruster tests, therefore, its dimensions are sufficient to center the thruster so as not to impose a boundary wall effect (force feedback) on the measurement of the thrust. 4.2 Support Structure It is composed of a support of L-shaped form, with a vertical section of 320 mm immersed on which the thruster will be fixed and another horizontal section of 220 mm that will transfer the force onto the force sensor located outside the tank. This support allows the rotational movement around an axis of 8 mm and fixed by bearings. 4.3 Propulsion The thruster that we want to identify was designed by us. It is based on an integrated brushless motor, reference A2212. The fairing of the thruster and its 3 bladed propeller are made by a 3D printer. Table 3 shows the dimensions of the thruster and Fig. 12 shows its assembly. Table 3. Thruster Dimensions Diameter Length Blades Material
Propeller 66 mm 20 mm 3 ABS
Fairing/Duct 70 mm 98 mm – ABS
4.4 Electronic Components A 5 kg strain gauge load cell is used to measure the thrust force generated. The principle of measurement with
Abdelmalek Laidani et al. / IFAC PapersOnLine 52-21 (2019) 254–259
258
Fig. 14. Thrust force generated by the thruster.
Fig. 13. Schematic of the developed test bench. a strain gauge is based on a Wheatstone bridge with variable resistors that are sensitive to deformation of the load cell. The variation of the voltage at the bridge is very small, so it is necessary to use an amplifier. A circuit which integrates an amplifier and a converter with a resolution of 24 bits (HX711) was used to make the sampling and digitization of the thrust force, then transmit it in numerical format to an acquisition card. The current absorbed by the thruster is measured by an ACS712 Hall effect current sensor, it produces an analog output that is measured by the Analog to Digital Converters (ADC) of the acquisition card. An Electronic Speed Controller (ESC) of 30 A is used to control the thruster. Fig. 13 shows the schematic diagram that explains the acquisition system. Table 4 shows the components used in the realization of the test bench and their approximate prices. The total price of the bench does not exceed 100$, which makes it low-cost and the results that will be presented in the next section will show its efficiency. Table 4. Itemised list and costs of constituent parts used to build the test bench Items Power Supply Acquisition Card Force Sensor 24 bits ADC Current Sensor Support Structure Other accessories Water Tank Overall Cost
Type ATX Arduino 5 kg Strain gauge HX711 ACS712 30A Stainless Steel Wires Recovery Article
Approx. Price 20 $ 10 $ 05 $ 03 $ 05 $ 10 $ 10 $ 00 $ 70 $
5. RESULTS AND COMPARISONS 5.1 Results To characterise the thruster we need to obtain the measurements concerning the thrust force generated and the corresponding current consumed by the the thruster,
Fig. 15. Current consumed by thruster for forward/reverse directions. across the speed range of the thruster from 0 to 100% which corresponds to a Pulse Width Modulation (PWM) from 1000 to 2000 µs. Fig. 14 shows the thrust force curve generated in the forward and reverse directions. We notice that there are 3 operating zones: the dead zone in which the thruster doesn’t produce force and is of the order of 50 us, then a linear zone which is the operating zone between 50 and 1800 us, but beyond this zone the thruster enters into a saturation zone where the thrust force remains constant despite the increase of the PWM signal. The thrust force generated in the forward and reverse direction is approximately 12.25 and 10.78 Newton respectively which is equivalent to 1.25 and 1.10 Kgf. The current consumed by the thruster is shown in Fig. 15, we note that the maximum current reaches 21 A which is equivalent to a power of about 250 Watts under a supply voltage of 12 Volts. In Fig. 16 the curve shows the efficiency of the thruster, where efficiency can be determined from the thrust force generated with respect to the power consumed. We note that the efficiency in the forward direction is 0.28 Newton/Watt which is greater than in the reverse direction which is equal to 0.15 Newton/Watt. Fig. 17 shows that the relationship between current and thrust force has a quadratic form and this implies that the Torque-Thrust relationship is also quadratic because the torque is proportional to the current.
Abdelmalek Laidani et al. / IFAC PapersOnLine 52-21 (2019) 254–259
259
constructed by us using a 3D printing technique. The curve presented in Fig. 14 can be transformed into a polynomial of N degrees by neglecting the higher order dynamics of the thruster. This polynomial will be integrated in the mathematical model of an underwater vehicle in order to get a simulation behavior closer to the reality so that the control laws can be validated before they are implemented on the vehicle. The comparison in Table 5 shows that the thruster that we have realized and tested using the test bench described above, does not perform as well as the T100 but its performance is nevertheless satisfactory given its cost. It must be taken into account that the respective tests are made using different test benches, and that the T100 values used in the comparison are values obtained from the manufacturer’s data sheet (BlueRobotics (2019)).
Fig. 16. Thruster efficiency.
REFERENCES
Fig. 17. Current vs. Thrust. 5.2 Comparison In order to validate the thruster that we have built and analyzed with the test bench, we will make a comparison with another thruster that is supposed to have similar performances. This thruster is the T100 from BlueRobotics (BlueRobotics (2019)), The Table 5 shows the main features of the comparison. Table 5. Comparison with T100 Thruster Electrical/Physical /Performance Max Forward Thrust Max Reverse Thrust Min Thrust Rotational Speed Operating Voltage Max Current Max Power Length Diameter Weight Propeller Diameter Cost(without ESC) Cost(with ESC)
T100
Our Thruster
2.36 kgf 1.85 kgf 0.01 kgf 300-4200 rev/min 6-16 Volts 12.5 Amps 136 Watts 102 mm 100 mm 295 gr 76 mm 119 $ 144 $
1.25 kgf (12.25 N) 1.10 kgf (10.78 N) 0.05 kgf (0.049 N) 700-15000 rev/min 12 Volts 21.0 Amps 250 Watts 98 mm 70 mm 152 gr 66 mm 45 $ 65 $
6. CONCLUSION We have presented in this article a low-cost test bench destined for the identification of an underwater thruster, this bench was realized by taking into account the advantages and disadvantages of existing test benches. The thruster that we have tested is a thruster designed and
Aras, M., Abdullah, S., A.Rahman, and Aziz, M. (2013a). Thruster modeling for underwater vehicle using system identification method. International Journal of Advanced Robotic Systems, 10, 252–264. Bachmayer, R., Whitcomb, L., and Grosenbaugh, M. (1999). Development, comparison, and preliminary experimental validation of nonlinear dynamic thruster models. IEEE Journal of Oceanic Engineering, 24, 481– 494. BlueRobotics (2019). BlueRobotics test bench. https:// www.bluerobotics.com. Cody, S. (2003). An Experimental Study of the Response of Small Tunnel Thrusters to Triangular and Square Wave Inputs. Master’s thesis, Naval Postgraduate School, Monterey, Canada. Cooke, J. (1989). Incorporating Thruster Dynamics in the Control of an Underwater Vehicle. Master’s thesis, MIT, USA. Deniellou, L., Gallou, Y., Gourmelen, P., and Seube, N. (1998). Force control of underwater thrusters with application to auv motion control. IEEE Oceanic Engineering Society, 2. Guibert, C. (2005). Modelisation et commande en pous´ee de moteurs ` a courant alternatifs en propulsion navale. Ph.D. thesis, Universit´e de Nantes. Healey, A., Rock, S., Cody, S., Miles, D., and Brown, J. (1995). Toward an improved understanding of thruster dynamics for underwater vehicles. IEEE Journal of Oceanic Engineering, 20, 354–361. Juca, J., Saito, M., CezarAdamowski, J., Takase, F., and Maruyama, N. (2012). Modelling and identification of yaw motion of an open-frame underwater robot. Journal of Intelligent and Robotic Systems, 66, 37–56. Muljowidodo, K., Adi, N., Prayogo, N., and Budiyono, A. (2009). Design and testing of underwater thruster for shrimp rov-itb. Indian Journal of Marine Sciences, 38, 232–237. Whitcom, L. and Yoerger, D. (1999). Development, comparison, and preliminary experimental validation of nonlinear dynamic thruster models. IEEE Journal of Oceanic Engineering, 24, 481–494. Yang, R., Clement, B., Mansour, A., Li, M., and Wu, N. (2015). Modeling of a complex-shaped underwater vehicle for robust control scheme. Journal of Intelligent and Robotic Systems.