- Email: [email protected]
Energy regulating and fluctuation stabilizing by air source heat pump and battery energy storage system in microgrid
Renewable Energy 95 (2016) 202e212
Contents lists available at ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
Energy regulating and fluctuation stabilizing by air source heat pump and battery energy storage system in microgrid Lian Yang a, b, *, Nengling Tai a, Chunju Fan a, Yuanye Meng b a b
School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China Innovation and Technology Department, GE Grid Solution, Stafford, ST17 4LX, UK
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 November 2015 Received in revised form 28 March 2016 Accepted 8 April 2016
The energy consumption statistics of buildings have shown that in China, 50%e70% of the annual energy consumption is consumed by cooling and heating systems, the majorities are air conditions and hot water supply. To cut down the investment of BESS, this paper studies the application of Air Source Heat Pump (ASHP) in a photovoltaic/Battery Energy Storage System (PV/BESS) microgrid. Energy dispatching strategy is proposed to take advantage of the thermal storage, and fluctuation stabilization strategy is proposed based on the decouple characteristics between Heat Pumps (HPs) and terminal devices. As the characteristics of fluctuation and the inputted capacity of ASHP should be considered during the stabilization, there are too many input variables for traditional fuzzy logic algorithm. According to this, this paper proposes an improved fuzzy logic algorithm, Double Fuzzy Logic (DFL) algorithm, to involve in all the necessary variables. At last, a case is simulated to verify the feasibility of the energy dispatching and fluctuation stabilization strategy. The results have verified that ASHP can decrease the capacities of PV/ BESS and stabilize the fluctuation effectively. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Microgrid energy management Fluctuation stabilize Battery energy storage system Air source heat pump Double fuzzy logic
1. Introduction Photovoltaic (PV) technology has been regarded as a solution to produce electrical energy. In recent years, PV technology has been improved greatly to achieve high energy efficiency with high durability [1e5]. While in the past the price of the PV modules was the major contribution to the cost of the systems, a downward tendency is now seen thanks to the development of silicon technologies such as Thin Film technologies, an increasing competition among manufacturers and a massive enlargement in the production capacity of PV modules [6]. However, similarly to other renewable energy sources, solar energy tends to be unsteady because it is influenced by natural and meteorological conditions [7]. Moreover, high penetration of intermittent renewable resources can bring up technical challenges including grid interconnection, power quality, reliability, protection, generation dispatch, and control [8]. The issue of how power fluctuation in PV is to be smoothed has attracted widespread interest and attention. As the levels of penetration of renewable energy rise, the technical impact
* Corresponding author. School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China. E-mail address: [email protected] (L. Yang). http://dx.doi.org/10.1016/j.renene.2016.04.019 0960-1481/© 2016 Elsevier Ltd. All rights reserved.
of PV on grid operation has led to the application of energy storage for renewables [9]. Recent papers proposed this application include a simple scheme to charge and discharge the Energy Storage System (ESS), such as storing excess power when the solar power output exceeds a threshold and discharge it back to the grid when the load demand is high [10]. In Refs. [11], a State of Charge (SOC) based smoothing control strategy was adopted to smooth out short-term power fluctuations of wind/PV hybrid system. In Refs. [12], a generalized double-shell framework for the optimal design of renewable energy systems was developed. Optimal sizing of distributed energy resources with storage systems was studied based on the yearly Joule losses, the yearly costs (capital and management), and the yearly CO2 emissions. In Refs. [13], a bidirectional converter and control system were designed for Renewable Energy Sources (RESs) and ESSs to take part in the voltage regulation and the frequency transient regulation. Ref [14] studied the economically feasibility of installing medium-scale distributed storage devices in the power system designed to lower the electricity cost for a customer-side application, assuming flexible electricity tariffs. In addition to the capability of enabling the integration of more RESs into the network, ESSs can provide many other benefits that can be summarized as: benefits related to load/generation shifting, benefits
L. Yang et al. / Renewable Energy 95 (2016) 202e212
Nomenclature C specific heat capacity of water f output power of PV g1, g2 load curves h, h' annual cost functions K,K1,K2 correction factors PASHP power consumption of ASHP PBESS_low output of BESS for stabilization Phigh high frequency component PHP capacity of HPs PHPmax extremely capacity design of HPs Phump operating power of ASHP Plow low frequency component PN rated power of ASHP PStotal stabilization command Q1 , Q2, Q3 electric quality differences between generators and load demands SBESS State of Charge of BESS SBESSmin allowed minimum state of BESS T time constant Tf time constant of filter X(s) reference value X1,X2,X3,X4 inputs of fuzzy logic modules Y(s) power consumption Greek symbols w flow of water DT temperature difference t2, t3 time period DQloss heating loss DQCOP decreased power consumption Df frequency deviation Df/Dt frequency changing rate a; b are the normalized coefficients m1i,m2j membership values Dh increased investment of HPs DPHP increased capacity of HPs
related to ancillary services and benefits related to grid system applications. In Refs. [15], the operating and maintenance cost of ESS was studied, and an ESS management was proposed based on the power forecasting module of PV. Similarly in Refs. [16,17], a cost-benefit analysis method was proposed for battery energy storage system (BESS) when BESS was applied to microgrid with PV systems. It is preferable to keep installed ESSs as small as possible, since ESSs are generally expensive. Table 1 shows an example of several ESSs' costs per output power (kW) and costs per energy(kWh) [18]. Table 1 shows that most ESSs' costs are still quite high. Thus far, various BESS-based methods of smoothing power
203
DQBESS decreased capacity of BESS Dh' decreased investment function DPASHP_low power increment of ASHP Superscript/subscripts ASHP Air Source Heat Pump ASHP_low Low frequency stabilization of ASHP BESS Battery Energy Storage System BESS_low Low frequency stabilization of BESS COP Coefficient of Performance f filter high high component HP Heat pump hump hump value loss energy losses low low component N nominal rating max maximum min minimum QASHP heating volume SBESSmax allowed maximum state of BESS Stotal total values of stabilization t time Acronyms ASHP Air Source Heat Pump BESS Battery Energy Storage System COP Coefficient of Performance DFL Double Fuzzy Logic ESS Energy Storage System HP Heat Pump LCC Life Cycle Cost MPPT Maximum Power Point Tracking PCC Point of Common Coupling PV Photovoltaic RES Renewable Energy Source SOC State of Charge
fluctuations in renewable power generation systems have been proposed, but economic cost of BESS has not been substantially reduced. As an effective way of reducing the capacity of BESS, controlling electrical appliances on the demand side was considered in Ref. [19]. In Refs. [20], the manageable loads were studied and a new formulation of shiftable loads was employed for a new load modeling method. Manageable loads are loads for which can be modified during the operating cycle without damage and degradation to the quality of the consumer. They can either be adjustable or shiftable. The ideal manageable loads should meet several requirements:
Table 1 Example of ESSs' cost. ESS
Cost per kW [1000$/kW]
Cost per kWh [1000$/kWh]
NaS battery Li-Ion battery NiMH battery High speed flywheel Electric double layer capacitor Pumped hydro
7.6 6e15
0.4 1e4
0.5e1 2 1.5e2
120e240 50~ 0.2
204
L. Yang et al. / Renewable Energy 95 (2016) 202e212
1) Power consumption of the load is large enough to compensate power fluctuation by the control of its power consumption. 2) The response speed of the load's power consumption to its reference signal is fast enough to compensate power fluctuation. 3) The loss of convenience by the control of power consumption is minor. Among electrical loads, heat pumps are commonly used in air conditioning and water heating and so on. As heat pumps can satisfy the three requirements mentioned in previous sections, they can be treated as an ideal option for control [18]. The statistics about energy consumption of buildings have shown that in China, 50%e70% of the annual energy consumption is consumed by cooling and heating systems, most of which are air conditions and hot water supply [21]. Air Source Heat Pump (ASHP) has been widely used in China and elsewhere with the advantage of great Coefficient of Performance (COP). There are several papers verified the feasibility of frequency regulation by ASHP. In Refs. [18,19,22], control characteristics of heat pumps were studied and the capacity reduction of ESS was calculated by simulation. The basic characteristic of a heat pump's power consumption was measured experimentally in a real building. In Refs. [23], a low cost state-ofart experimental stand was researched for testing ASHP under controlled evaporator ambient conditions. This innovation can be referred by ASHP to extend the water tank, for the sake of more thermal energy storage. In Refs. [24], the operation optimization of a distributed energy system was proposed to improve the overall exergy efficiency. Thermal storages were studied and used to satisfy low-quality thermal energy demands. In Refs. [25], the thermal storage tanks for space heating were selected to minimize the use of auxiliary equipment. In Refs. [26], the COP characteristic of heat pump was studied. The research shows that COP is dependent on the outside air temperature. In Refs. [27], calculation and design of the heat pumps were proposed for purposes of heating and cooling. The model involves multiple energy devices that convert a set of primary energy carriers with different energy quality levels to meet given time-varying user demands at different energy quality levels. By promoting the usage of low-temperature energy sources to satisfy low-quality thermal energy demands, the waste of highquality energy resources can be reduced, therefore improving the overall exergy efficiency. However, due to the fact that the heat conduction of buildings is complicated, and the air temperature varies with time and season, it's difficult to build an accurate model for the stabilization control. Then the fuzzy logic comes into view with the advantages on complicated system [28e35]. When something is uncertain, like measurement, it is difficult to determine its exact value by the traditional control algorithms. Then the fuzzy logic algorithm allows handling much of this uncertainty by defining fuzzy systems employing fuzzy logic. Fuzzy logic control is simple, can be easily realized, doesn't require modeling and has strong robustness, it is more appropriate to use in applications where parameters and/or model are undefined or changeable [28]. In Refs. [29,30], fuzzy logic algorithm was used to manage the flow of energy between ESSs. In the case without using fuzzy control, the stored energy of Superconducting Magnetic Energy Storage or the secondary battery reaches the maximum rated capacity or minimum rated capacity, and power control becomes impossible in some places. By modifying the power reference value using the fuzzy control, it is possible to avoid the saturation or depletion of their output [30]. In Refs. [31e35], fuzzy logic algorithm was used to manage the flow of energy between ESS and renewable energy sources. In conclusion, fuzzy logic algorithm is suitable as the control algorithm in the complicated systems. This paper studies the application of ASHP in a PV/BESS
microgrid for energy dispatching and fluctuation stabilization. An energy dispatching strategy is studied to cut down the capacity of BESS, and the stabilization strategy is proposed based on the characteristics of PV/BESS and ASHP. As the characteristics of fluctuation and the inputted capacity of ASHP should be considered during the stabilization, there are too many input variables for traditional fuzzy logic algorithm. According to this, this paper proposes an improved fuzzy logic algorithm, Double Fuzzy Logic (DFL) algorithm, to involve in all the necessary variables. At last, a case is simulated to verify the feasibility of the energy dispatching and fluctuation stabilization strategy.
2. Structure of microgrid Structure of microgrid is shown as Fig. 1. Photovoltaic provides power energy, and BESS balances the energy supply and demand. Microgrid connects to the power grid through Point of Common Coupling (PCC). Load consumptions in the microgrid consist of two parts: heating load satisfied by ASHP; common electric load such as lighting.
2.1. Photovoltaic unit Photovoltaic panel can be seen as an electrical generator, whose maximum power depends on cell temperature and solar radiation. In order to ensure the maximum efficiency whatever the conditions of sunlight and temperature are, the photovoltaic panel is associated to a power converter controlled with a Maximum Power Point Tracking (MPPT) algorithm [1,36]. Besides, the solar radiation received by inclined photovoltaic panel is the other key parameter for the power output. Hay-Davies-Klucher-Reindl anisotropic model permits to take into account the effect of clouds on terrestrial irradiance through a clearness index and a diffuse fraction [37]. It was shown that the weather and the MPPT strategy had aggregated the output fluctuation of PV unit. Stochastic of clearness leads to the stochastic fluctuation of output power. The fluctuation should be stabilized by auxiliary devices such as energy storage devices.
2.2. Battery energy storage system RESs output the fluctuated power into microgrid, and the fluctuation requires to be stabilized by frequency regulating devices such as BESS. Assuming that State of Charge (SOC) of BESS is SBESS, SBESS should be maintained in a certain range, or else life damage will be caused to BESS. We can get:
Fig. 1. Structure of microgrid.
L. Yang et al. / Renewable Energy 95 (2016) 202e212
SBESSmin SBESS ðtÞ SBESSmax
(1)
where SBESSmin is allowed minimum state; SBESSmax is the allowed maximum state. Capacity of BESS can be decomposed into two parts: one for fluctuation stabilization, the other for peak-valley regulation. Although BESS delivers better response characteristics and low power losses, it expends larger investment. Besides, life loss will be aggravated by the stabilizing operation.
2.3. Air Source Heat Pump Among electrical appliances, heat pumps, which are used for air conditioning and water heating and so on, are attractive options for compensation of power fluctuation [18]. ASHP is composed of Heat Pump (HP), water tank (thermal storage), terminal units, etc., shown in Fig. 2. 1) HP. HP is a closed loop composed of compressor, condenser, throttling gear and evaporator. Compressor circulates the refrigerant through the loop for heat transmission. 2) Water tank. ASHP has a tank to store hot water. As ASHP is taken into account of microgrid management, the capacity of water tank is larger than the usual ones, for the propose of storing more heating energy. 3) Terminal units. Heat stored in the water tank is dispatched by the terminal units. One portion of the heat is used for air conditioning, while the other satisfies the hot water demand. The heating volume of ASHP can be expressed as follows:
QASHP ¼ wC DT
(2)
where QASHP is the heating volume; w is the flow of water; C is the specific heat capacity of water; DT is the temperature difference. Power consumption corresponding to the heating volume is expressed by COP, shown as follows:
COP ¼
QASHP PASHP
(3)
where PASHP is the power consumption of ASHP. COP of ASHP is affected by HP's performance and ambient temperature. COP ascends nonlinearly with the rising of ambient temperature [23]. It means that during the whole daytime, ASHP will consume less electric power for the same quality of heat during high-temperature period. BESS regulates the peak-valley load by charging or discharging the electric energy, while ASHP can regulate the thermal load by
205
charging or discharging the thermal energy. BESS has better response characteristics, while ASHP has greater economic efficiency. Thus, ASHP can operate as an assistance of BESS by devoting into energy dispatching and fluctuation stabilization. 3. Microgrid energy management strategy Microgrid system operates at a low voltage distribution, and has several distributed energy resources. Microgrid system also has the ability to operate connected to the power grid (on grid) or disconnected to the power grid (off grid/islanded) [38]. When operating as on grid, microgrid system can trade electric energy with the grid. In this paper, electric energy can only flow into microgrid owing to the actual difficulties from power utilities and microgrid operators. In order to reduce its influence and dependence on the power grid, microgrid purchases electric energy only during the valley price period. Power is traded based on load prediction. Microgrid predicts the solar condition and load of coming days, and then purchases electric energy to fill the vacancy. The energy flow in microgrid is shown as Fig. 3. Operating as a thermal energy storage device, water tank can regulate microgrid energy together with BESS. It means that water tank can bridge the gap between HP and thermal demand, which releases HP for controllable energy consumption. 3.1. Energy dispatching during the whole day time The energy states of BESS and water tank drop to the minimum values in the morning, measured during a whole day time. In daytime, solar radiation and ambient temperature rise up, leading to the increase of PV's output and HPs' COP. Accompanied by the rising of energy supply, BESS starts to charge energy, as well as ASHP. More electric energy will be consumed if more HPs are put into operation. HPs supply more heat than the real-time demand and then the redundant heat is stored by the water tank. At night, PV ceases, ambient temperature drops, meanwhile COP of HPs falls. A large portion of HPs are paused for less energy consumption. The water tank supplies the thermal demand, while BESS sustains ASHP and other electric loads. Fig. 4 shows the regulation principle of microgrid. In Fig. 4, f(t) is the output power of PV; g1(t) is the former load curve that ASHP operated according to the real-time demand; g2(t) is the regulated load curve that ASHP participated in the energy dispatching; Q1, Q2 and Q3 are electric quality differences between generators and load demands. If microgrid operates with g1(t), the electric quality which should be regulated by BESS is Q1 þ Q2; or else microgrid operates with g2(t), then electric quality BESS should regulate is Q1 þ Q3. Associated by ASHP, BESS can cut down its own capacity of Q2Q3. Besides, HPs with greater COP can cut down the total power consumption during the whole day time. 3.2. Optimal sizing of BESS and ASHP The rated capacity of HPs plays an important role in g2(t). Adding more HPs is an effective way to increase the rated capacity. g2(t)
Fig. 2. Structure of ASHP
Fig. 3. Energy flow in the microgrid.
206
L. Yang et al. / Renewable Energy 95 (2016) 202e212
DQBESS ¼
Z
Z ½ f ðtÞ g1 ðtÞdt
t2
½ f ðtÞ g2 ðtÞdt
(9)
t3
where, t2 is the period of f(t) being greater than g1(t); t3 is the period of f(t) being greater than g2(t). The decreased investment corresponding to DQBESS is shown as follows:
Dh0 ðQBESS Þ ¼ h0 ðQBESS þ DQBESS Þ h0 ðQBESS Þ
Fig. 4. Energy dispatching in microgrid.
should get closer to f(t), for the propose of cutting down the capacity of BESS. Besides, g2(t) should be planned based on such premise that user demands are satisfied.
Z24
Z24
g2 ðtÞdt DQloss þ DQCOP
g1 ðtÞdt ¼ 0
(4)
0
where, DQloss is heating loss; DQCOP decreased power consumption caused by different COP. The extremely capacity design of HPs, PHPmax, is that: (1), the HPs work during the period of g2(t) < f(t), and then all the HPs cease during the period of g2(t)>f(t); or (2), all the energy of f(t)(having subtracted the energy for other electric load) are consumed by ASHP during the period of g2(t) < f(t). However, this will increase the investment of HPs. Hence the balance between the increasing investment of HPs and the decreasing investment of BESS should be discussed. The increased investment Dh of HPs with the increased capacity DPHP is shown as follows:
Dh ¼ hðPHP þ DPHP Þ hðPHP Þ
(5)
where, Functionh() is the annual cost of ASHP, it was calculated by converting the Life Cycle Cost (LCC) to each year; PHP is the capacity of HPs. The decreased power consumptionDQCOP with DPHP is shown as follows:
DQCOP ¼ DW$
1 1 COP2 COP1
(6)
where, DW is the transferred heat from night time to day time (convert to each year); COP1 is the average COP during the day time; COP2 is the average COP during the night time. Water tank should be expended for more energy storage. The expanded capacity is assumed as DQwater:
DQwater ¼
K$COP1 $ cwater $DT
Z ½g2 ðtÞ g1 ðtÞdt
where Function h0 ðÞ is the annual cost, it was also calculated by converting LCC to each year. Comparing the increased investment of ASHP and the decreased investment of BESS, we can get the optimization function shown as equation (11):
minDhðASHPÞ Dh0 ðQBESS Þ
(11)
According to Equation (11), the capacity of ASHP and BESS can be optimized for lower investment. 3.3. Fluctuation stabilized by BESS and ASHP Stochastic of clearness leads to the output fluctuation of PV, energy consumption of microgrid loads will also fluctuate due to the users' stochastic behaviors. BESS had been working on energy stabilization in many microgrids. With the controllable merit provided by water tank, HPs can regulate the power consumption as the response to fluctuation and the assistance to BESS. In the HP unit, there is a valve that controls the refrigerant flow. The control of the valve is carried out via hydraulic system. The HP unit has two external terminals that correspond to “Load Up” and “Load Down”. The signal given to these terminals controls the valve and changes the amount of refrigerant. As a result, the power consumption of HP is changed. Power consumption characteristic of HP is modeled using a first order transfer function, shown as Equation (12).
YðsÞ ¼
1 XðsÞ 1 þ Ts
(12)
Where, Y(s) is the power consumption, X(s) is the reference value and T is the time constant of the model. The daily peak-valley characteristics of microgrid load should be regulated by ASHP and BESS together, moreover the frequency fluctuation also requires to be stabilized. During the stabilization period in microgrid, the assistance of ASHP can cut down the charge/discharge current of BESS, which means that the loss of BESS is lowed. 4. Double Fuzzy Logic model
(7)
t1
where, K is the margin coefficient, cwater is heating ratio of water; DT is the temperature difference between unheated water and heated water; t1 is the period of g2(t) being greater thang1(t). The increased investment of ASHP corresponding to the increased capacity of DPHP is shown as follows:
DhðASHPÞ ¼ hðPHP þ DPHP ; DQCOP ; DQwater Þ hðPHP Þ
(10)
(8)
where, Functionh() is the annual cost of ASHP, it was calculated by converting the Life Cycle Cost (LCC) to each year. The decreased capacity DQBESS of BESS is shown as follows:
This section aims at DFL model for frequency fluctuation. Fluctuation caused by the randomness of PV and load should be stabilized. Assuming thatPStotal is the stabilization command: positive PStotal needs BESS and ASHP to supply energy; negative PStotal needs BESS and ASHP to absorb energy. BESS is the non-inertial device with fast response, while ASHP is the inertial device with less responding rate. Thus the high frequency fluctuation can only be compensated by BESS, while the low frequency fluctuation can be stabilized by BESS and ASHP together. High-pass filter is employed to pick up high frequency componentPhigh fromPStotal, andPhigh is responded by BESS. The rest fluctuation stabilizing command, assumed asPlow, is responded by both ASHP and BESS, shown as equations (13) and (14):
L. Yang et al. / Renewable Energy 95 (2016) 202e212
Phigh ðsÞ ¼
sTf P ðsÞ 1 þ sTf Stotal
Plow ðsÞ ¼ PStotal ðsÞ Phigh ðsÞ ¼
(13)
1 P ðsÞ 1 þ sTf Stotal
(14)
whereTf is the time constant of filter. When analyzing the response dispatch between BESS and ASHP, operating characteristics of each device should be considered for the charge-discharge constraints under different states. The stabilization command should be dispatched based on the characteristics of BESS and ASHP, hence the control objectives of DFL are shown as follows: (1) The required capacity and charge/discharge current of BESS should be limited as possible, so as to cutting down the loss of BESS and the investment of microgrid. (2) If SOC of BESS reaches the upper limit, energy absorption should be reduced during the charging state, and energy release should be encouraged during the discharging state. Otherwise, SOC of BESS reaches the lower limit, energy absorption should be encouraged during the charging state, and energy release should be reduced during the discharging state. (3) If HPs run under heavy load, unloading is encouraged; or else if HPs run under fractional load, load increasing is encouraged. There are four inputs of DFL, one for BESS, one for ASHP, the others for the stabilization dispatch. Energy states of each storage device are picked up as two inputs. State of Charge of BESS at time t, SOC(t), is assigned as one input of DFL. Load coefficient of ASHP, Phump(t)/PN, is assigned as another input, where Phump(t) is the operating power at time t; and PN is the rated power of ASHP. The other two inputs should be assigned to the variables that reflect the energy fluctuation in microgrid. The fluctuation Plow, related to the frequency deviation Df in microgrid, is picked up as the total stabilization command of BESS and ASHP. The frequency changing rate Df/Dt is picked up as one of the dispatch coefficient. Then we can get the expressions of the four inputs:
X1 ðtÞ ¼ SOCðtÞ
(15)
X2 ðtÞ ¼ Plow ðtÞ
(16)
X3 ðtÞ ¼
Phump ðtÞ PN
(17)
X4 ðtÞ ¼
Df Dt
(18)
Fig. 5 shows the DFL model for stabilization dispatch. Inputs X1~X4 are processed by Fuzzy Logic submodules Fuzzy1 and Fuzzy2 respectively, and output the correction factors K1 and K2 to weighting function. After weighted by coefficient a (0 < a < 1), we can get the ultimate value K of DFL.
a¼
aCBESS CBESS þ bCASHP
(19)
Where, CBESS is the inputted capacity of BESS for stabilization; CASHP is the inputted capacity of ASHP; a and b are the normalized coefficients, they are set considering capacities transforming and economic weights. Submodule Fuzzy1: Fuzzy 1 inputs [X1,X2] and outputs K1. If the value of Plow(t) is appropriate and BESS has enough margin for stabilization (charging or discharging), Fuzzy Logic rule will turn down the correction factor K1. Or else, BESS has not enough margin, Fuzzy Logic rule will turn up the correction factor K1, and ASHP will undertake more stabilization assignment. Submodule Fuzzy2: Fuzzy 2 inputs [X3,X4] and outputs K2. It adjusts the stabilization command based on the operating state of ASHP and the changing rate of fluctuation. Firstly, heavy load state of ASHP is discussed. If the value of X4 is negative, Fuzzy Logic rule will turn up the output K2 to remit the pressure of ASHP; otherwise X4 is positive, K2 will be turned down. Secondly, under load of ASHP is discussed. If the value of X4 is negative, K2 will be turned down; otherwise, K2 will be turned up. The primary elements of a fuzzy logic model are (1) fuzzy sets, (2) membership function (3) IF-THEN rules [39]. A fuzzy set is labeled by a linguistic term, and the linguistic term is a word such as “NB”, “NM”, etc. The fuzzy sets and IF-THEN rules construct a rule base of fuzzy logic model. Membership function is the effective method for comprehensive evaluation based on multi factors. Membership function is a curve that defines how each point in the input space is mapped to a membership value between 0 and 1. It provides a measure for the degree of an element to a fuzzy set. It fully defines fuzzy set and it can take any form, but there are some common examples that appear in real applications. There are different shapes of membership functions: triangular, trapezoidal, piecewise-linear, Gaussian and so on. In most literature, triangular membership functions are used due to having flexible formula to calculate [40]. Thus the triangular shape is employed in this paper. Membership functions of Fuzzy 1 and Fuzzy 2 are shown as Fig. 6 (a) and (b) respectively. The rule of Fuzzy Logic will affect the control result directly. Fig. 6 has shown the inputs, outputs and membership functions of two Fuzzy Logic submodules. The fuzzy rules of them are shown as Tables 2 and 3. Deduction results of fuzzy rules are fuzzy sets, which should be cleared by weighting method. Correction factor K1 is calculated by Equation (20):
PP K1 ¼
i
j
m1i ðX1 ðtÞÞm2j ðX2 ðtÞÞDkij
PP i
Fig. 5. Double Fuzzy Logic model.
207
j
m1i ðX1 ðtÞÞm2j ðX2 ðtÞÞ
(20)
Where, m1i ðX1 ðtÞÞ and m2j ðX2 ðtÞÞ are membership values of the inputs X1(t) and X2(t) respectively; the set of i is {NL, NM, ZO, PM, PL}; set of j is {NB, NM, NS, ZO, PS, PM, PB}; Dkij is the output corresponding to m1i ðX1 ðtÞÞ and m2j ðX2 ðtÞÞ. The calculation of correction factor K2 is similar to that of K1, thus is omitted for simplification. Stabilization command of BESS and ASHP should be coordinated in the DFL, hence we can get:
208
L. Yang et al. / Renewable Energy 95 (2016) 202e212
Fig. 6. Membership functions of DFL.
Table 2 Rule base of Fuzzy 1. X1
NB NM ZO PM PB
DPASHP
NB NM NS ZO PS PM PB
NB
NM
NS
ZO
PS
PM
PB
PB PM ZO NM NB
PM PS ZO NS NM
PS PS ZO NS NS
ZO ZO ZO ZO ZO
NS NS ZO PS PS
NM NS ZO PS PM
NB NM ZO PM PB
NB
NM
NS
ZO
PS
PM
PB
NB NM NS ZO NS ZO PS
NB NS ZO PS ZO PS PM
NB ZO PS PM PS PM PB
PB PB PB PB PB PB PB
PB PM PS PM PS ZO NB
PM PS ZO PS ZO NS NB
PS ZO NS ZO NS NM NB
X4
Plow ðtÞ ¼ PBESS
low ðtÞ
þ DPASHP
low ðtÞ
(21)
Where, PBESS_low(t) is the output power of BESS for stabilization, positive value is assumed under charging model; DPASHP_low(t) is the power increment of ASHP, shown as Equation (22):
DPASHP
low ðtÞ
¼ DPASHP
¼ PASHP ðtÞ PASHP ðt 1Þ
(22)
Stabilization command betweenPBESS_low(t) and DPASHP_low(t) is dispatched by the DFL, shown as follows:
PBESS
low ðtÞ
1Þ þ Plow ðtÞ low ðt 1Þ $KðtÞ
low ðt
DPASHP
X2
Table 3 Rule base of Fuzzy 2. X3
low ðtÞ
¼ Plow ðtÞ DPASHP
low ðt
1Þ $½1 KðtÞ
(23) (24)
According to the equations (23) and (24), BESS and ASHP stabilize the fluctuation based on their own operating status. The DFL model set forth above has involved the SOC of BESS with Fuzzy 1, and considered the output of ASHP with Fuzzy 2. Then K1 and K2 are normalized by weighting coefficient. At last, DFL outputs the factor K for stabilization.
5. Case analyses This paper analyzed the actual microgrid case in Meizhou Island in China. The island is supplied through a 15 km submarine cable, which has caused great losses. Local photovoltaic generators are required to replace the diesel generators and decrease the electric transmission losses. The capacities of PV and BESS were designed for the energy balance during only one day period, because the microgrid can be sustained through the submarine cable, the continuous rainy weather principle can be ignored. Since the advantages of ASHP to other heating devices were verified in many literature, this section only analysis different operation strategies of ASHP. The simplified dynamic model of the microgrid is shown as Fig. 7. Parameters are shown as Table 4. Simulation results are shown as Figs. 8e15. Fig. 8 shows the statistical load profiles of the microgrid. The statistics were counted by the average data during the winter period. And they were counted according to the local communities, containing the resident buildings, schools and hotels, provided by Zhongnan Construction Group. Power consumption of the whole day was
L. Yang et al. / Renewable Energy 95 (2016) 202e212
209
Fig. 7. Simplified model of the microgrid.
Table 4 Parameters of the microgrid model. Hu t Kpt Kit Vu t Tcon Tf Kb
5.19 3 0.6 1 0.02 5 380
Tb R1 k1 Trh Fhp Trh
0.04 0.04 0.03 30 0.3 7
12 MWh, while the peak and valley values were 0.75 MW and 0.25 MW. It can be seen that during the whole day time, heating load occupied about 60% of total load. Besides, heating load occupied much more during evening time. At that time demands were supplied by BESS. Then capacity of PV was designed as 12 MWh/ 2.5 MWp, for the strategy verification. Fluctuation characteristics of PV can be seen from Fig. 9, where the random data of illumination were referred to the local meteorology. Fig. 10 shows a typical load curve during the whole day time. This load curve expresses the peak-valley characteristic and random characteristic. Load demands from 17:00 to 24:00 were quite great. The heating demand was satisfied by ASHP in real-time. Only BESS was applied for regulation and stabilization. SOC of BESS is shown as Fig. 11. It can
Fig. 9. Output characteristics of PV.
Fig. 10. Typical load curve of the microgrid.
Fig. 8. Statistical load curves of the microgrid.
be seen that the variance of SOC during the whole day time was about 7 MWh. The variance value of SOC can be used to design the installed capacity of BESS, based on the planned upper limit and lower limit (in terms of percentage). Then, ASHP was applied to the microgrid. ASHP regulated the cooling load consumption by ice tankers, and the regulated load is shown as Fig. 12. ASHP supplied more quantity of heat to the water tank during the day time, when the value of COP was higher. A part of energy demand was
210
L. Yang et al. / Renewable Energy 95 (2016) 202e212
Fig. 11. SOC of BESS in the microgrid without ASHP.
Fig. 14. Stabilization dispatch between BESS and ASHP.
Fig. 12. Regulated load curve of the microgrid.
Fig. 15. SOC of BESS in the microgrid with ASHP.
transferred from night time to day time. Besides, ASHP supplied fluctuation stabilization together with BESS. Fig. 13 shows the response characteristics of the two devices. The response speed of BESS was much faster than that of ASHP, owing to that time constant of BESS was much smaller. Thus the high-frequency fluctuation can only be stabilized by BESS. The stabilization was dispatched under the DFL strategy expressed in Section 4, and the results are shown as Fig. 14. It can be seen that stabilization of BESS varied faster and more severe than that of ASHP. Since ASHP undertook partial fluctuation and energy storage, BESS gave a more stable and less variance performance on SOC, shown in Fig. 15. During the first design (capacity of PV was 12 MWh/2.5 MWp), values of SOC at 0:00 and 24:00 are 3 MWh and 5 MWh respectively. These results were due to the fact that ASHP exhausted less power under greater COP value. During the second design, the capacity of PV was designed as 10 MWh/2.1 MWp, so that energy balance of microgrid could be maintained. Then the variance of SOC decreased to 3 MWh, and 4 MWh of batteries could be cut down or dropped out as reserve. Correspondingly, the expended capacity of
Fig. 13. Bode diagram of BESS and ASHP.
water tank was about 42 m3. 6. Conclusions This paper has proposed a microgrid energy management strategy to reduce the cost of photovoltaic and Battery Energy Storage System. Air Source Heat Pump had been introduced to a PV/ BESS microgrid for energy dispatching and fluctuation stabilization. In the greater power grid, the peak-valley load can be regulated by the time-of-use price or pumped-storage power station to achieve a lower unit regulation price. In the microgrid however, renewable energy has the fluctuation characteristics; load is smaller and has the obviously peak-valley characteristics. As a result, only Battery Energy Storage System had been widely applied. If the residential load occupies a large proportion, heating load will directly increase the capacity of Battery Energy Storage Systems. The application of Air Source Heat Pump decreases the capacities of photovoltaic and batteries. Besides, assisted by Air Source Heat Pump for stabilizing the power fluctuation, losses of battery energy storage system are decreased. Consequently, the application of ice storage air condition in microgrid has greater operation reliability and economic efficiency. Different from photovoltaic, the output characteristics of wind generators are determined by the local meteorological conditions. If the wind energy is abundant during day time and lacking during night period, which is similar to the load consumption characteristics, then the wind farm is suitable for the strategy proposed in this paper. Otherwise, once the output of wind generators is not similar to the load consumption characteristics, the wind generators are not suitable for this strategy. This paper considers a microgrid connected with a greater grid, which is different from the stand-alone microgrid. Capacities of photovoltaic and Battery Energy Storage System should be designed based on the continuous rainy weather condition. Application of Air Source Heat Pump in the stand-alone microgrid
L. Yang et al. / Renewable Energy 95 (2016) 202e212
should be expended based on the strategy proposed in this paper. In addition, as shown in statistics, cooling and heating loads occupy 50%e70% of the total power consumption in buildings. There are other manageable loads that can also be used in microgrid energy management. Further studies have been started on Ice Storage Air Condition (ISAC), which is one of the potential solutions for microgrid energy management. Acknowledgement This research was supported by Science and technology project of Education Ministry, China, No.113023A. The authors would like to thank Hengxu HA at GE Grid Solution; Jinshi Chen at Zhongnan Construction Group and Chengen Wu at Fujian electric power research institute, for their assistance in getting the required data to carry out this research. References [1] A. Bouabdallah, J.C. Olivier, S. Bourguet, M. Machmoum, E. Schaeffer, Safe sizing methodology applied to a standalone photovoltaic system, Renew. Energy 80 (2015) 266e274. [2] R. Bakhshi, J. Sadeh, H.R. Mosaddegh, Optimal economic designing of gridconnected photovoltaic systems with multiple inverters using linear and nonlinear module models based on Genetic algorithm, Renew. Energy 72 (2014) 386e394. [3] A.M. Dizqah, A. Maheri, K. Busawon, An accurate method for the PV model identification based on a genetic algorithm and the interior-point method, Renew. Energy 72 (2014) 212e222. [4] M. Mohsenzadeh, R. Hosseini, A photovoltaic/thermal system with a combination of a booster diffuse reflector and vacuum tube for generation of electricity and hot water production, Renew. Energy 78 (2015) 245e252. [5] P.M. Almeida, P.G. Barbosa, J.G. Oliveira, Digital proportional multi-resonant current controller for improving grid-connected photovoltaic systems, Renew. Energy 76 (2015) 662e669. [6] G. Graditi, D. Colonnese, N. Femia, Efficiency and reliability comparison of DCDC converters for single phase grid connected photovoltaic inverters, in: Power Electronics Electrical Drives Automation and Motion (SPEEDAM), 2010 International Symposium on. IEEE, 2010, pp. 140e147. [7] X.Y. Wang, V.D. Mahinda, S.S. Choi, Determination of battery storage capacity in energy buffer for wind farm, IEEE Trans. Energy Convers. 23 (3) (2008) 868e878. [8] S. Teleke, M.E. Baran, A. Huang, S. Bhattacharya, Control strategies for battery energy storage for wind farm dispatching, IEEE Trans. Energy Convers. 24 (3) (2009) 725e732. [9] T. Sercan, E.B. Mesut, B. Subhashish, A.Q. Huang, Rule-based control of battery energy storage for dispatching intermittent renewable sources, IEEE Trans. Sustain. energy 1 (3) (2010) 117e124. [10] M. Mehos, D. Kabel, P. Smithers, Planting the seed, IEEE Power Energy Mag. 7 (3) (2009) 55e62. [11] L. Xiangjun, H. Dong, L. Xiaokang, Battery energy storage station (BESS)-Based smoothing control of photovoltaic (PV) and wind power generation fluctuations, IEEE Trans. Sustain. Energy 4 (2) (2013) 464e473. [12] M.L. Di Silvestre, G. Graditi, E.R. Sanseverino, A generalized framework for optimal sizing of distributed energy resources in micro-grids using an indicator-based swarm approach. Industrial Informatics, IEEE Trans. 10 (1) (2014) 152e162. [13] Graditi G, Ippolito M G, Telaretti E, et al. An innovative conversion device to the grid interface of combined RES-based generators and electric storage systems. Ind. Electron. IEEE Trans.. 62(4): 2540e2550. [14] G. Graditi, M.G. Ippolito, E. Telaretti, et al., Technical and economical assessment of distributed electrochemical storages for load shifting applications: an Italian case study, Renew. Sustain. Energy Rev. 57 (2016) 515e523. [15] C. Chen, S. Duan, T. Cai, et al., Smart energy management system for optimal microgrid economic operation, Renew. Power Gener. IET 5 (3) (2011) 258e267. [16] A. Nottrott, J. Kleissl, B. Washom, Energy dispatch schedule optimization and cost benefit analysis for grid-connected, photovoltaic-battery storage systems, Renew. Energy 55 (2013) 230e240. [17] Y. Yang, N. Yang, H. Li, Cost-benefit study of dispersed battery storage to increase penetration of photovoltaic systems on distribution feeders, in: IEEE PES General Meeting Conference & Exposition, 2014, pp. 1e5. [18] S. Kawachi, J. Baba, H. Hagiwara, Energy capacity reduction of energy storage system in microgrid by use of heat pump: characteristic study by use of actual machine, in: IEEE 14th Power Electronics and Motion Control Conf, vol. 11, 2010, pp. 52e58. [19] S. Kawachi, H. Hagiwara, J. Baba, Modeling and simulation of heat pump air conditioning unit intending energy capacity reduction of energy storage system in microgrid, in: IEEE Power Electronics and Applications Conf, 2011,
211
pp. 1e9. [20] G. Graditi, M.L. Di Silvestre, R. Gallea, et al., Heuristic-based shiftable loads optimal management in smart micro-grids, Ind. Inf. IEEE Trans. 11 (1) (2015) 271e280. [21] GB 50189-2005 Design Standard for Energy Efficiency of Public Buildings [S], 2005 (in Chinese). [22] S. Kawachi, H. Hagiwara, J. Baba, Feasibility study on short period power fluctuation compensation using heat pump air conditioning system, J. Int. Counc. Electr. Eng. 2 (2) (2012) 139e145. [23] V.S. Cr aciun, C. Bojesen, M.B. Blarke, A flexible and low cost experimental stand for air source heat pump for Smart Buildings, in: Optimization of Electrical and Electronic Equipment (OPTIM), 13th Int. Conf. IEEE, 2012, pp. 1040e1045. [24] M. Di Somma, B. Yan, N. Bianco, et al., Operation optimization of a distributed energy system considering energy costs and exergy efficiency, Energy Convers. Manag. 103 (2015) 739e751. [25] W. Gu, Z. Wu, R. Bo, Modeling, planning and optimal energy management of combined cooling, heating and power microgrid: a review, Int. J. Electr. Power Energy Syst. 54 (2014) 26e37. [26] S. Obara, S. Watanabe, B. Rengarajan, Operation method study based on the energy balance of an independent microgrid using solar-powered water electrolyzer and an electric heat pump, Energy 36 (8) (2011) 5200e5213. [27] J. Perko, V. Dugec, D. Topic, Calculation and design of the heat pumps. Energetics (IYCE), in: Proceedings of the 3rd Int. Youth Conf. IEEE, 2011, pp. 1e7. [28] S. Faquir, A. Yahyaouy, H. Tairi, J. Sabor, A type-1 fuzzy logic algorithm to manage the flow of energy in a stand-alone PV/wind/battery hybrid system, Intell. Syst. Comput. Vis. ISCV (2015) 1e6. [29] M. Mohammedi, M. Becherif, A. Aboubou, O. Kraa, Fuzzy logic and passivity based control applied to hybrid DC power source using fuel cell and battery, in: Systems and Control (ICSC), 4th Int. Conf, 2015, pp. 510e515. [30] T. Ise, M. Kita, A. Taguchi, A hybrid energy storage with a SMES and secondary battery, IEEE Appl. Supercond. 15 (2) (2005) 1915e1918. [31] D.O. Elamine, E.H. Nfaoui, B. Jaouad, Multi-agent system based on fuzzy control and prediction using NN for smart microgrid energy management, Intell. Syst. Comput. Vis. ISCV (2015) 1e6. [32] S. Zhang, Y. Mishra, M. Shahidehpour, Fuzzy-logic based frequency controller for wind farms augmented with energy storage systems, IEEE Power Syst. 99 (2015) 1e9. [33] P.I. Gonzalez, G.A.J. Calderon, M. Calderon Godoy, A. Ramiro Gonzalez, Experimental automation platform of stand-alone hybrid renewable energy systems: fuzzy logic application and exergy analysis, in: 6th Int. Renewable Energy Congress (IREC), 2015, pp. 1e6. [34] N.L. Diaz, W. Dan, T. Dragicevic, J.C. Vasquez, J.M. Guerrero, Fuzzy droop control loops adjustment for stored energy balance in distributed energy storage system, in: 9th Int. Conf. Power Electronics and ECCE Asia (ICPE-ECCE Asia), 2015, pp. 728e735. [35] S. Zhang, Y. Mishra, M. Shahidehpour, Fuzzy-logic based frequency controller for wind farms augmented with energy storage systems, IEEE Power Syst. 99 (2015) 1e9. [36] A. Bouabdallah, S. Bourguet, J.C. Olivier, Optimal sizing of a stand-alone photovoltaic system, in: IEEE Int. Conf. Renewable Energy Research and Applications (ICRERA), 2013, pp. 543e548. [37] A. Bouabdallah, S. Bourguet, J.C. Olivier, Photovoltaic energy for the fixed and tracking system based on the modeling of solar radiation, IEEE Conf. Ind. Electron. Soc. (2013) 1821e1826. [38] B.S. Hartono, Y. Budiyanto, R. Setiabudy, Review of microgrid technology, IEEE Int. Conf. Qual. Res. QiR (2013) 127e132. [39] A.M. Aziz, A new perspective on the choice of fuzzy membership functions in multitarget tracking systems, in: Aerospace Conference, 2015 IEEE. IEEE, 2015, pp. 1e8. [40] E.B. Kumar, M. Sundaresan, Edge detection using trapezoidal membership function based on fuzzy's mamdani inference system, in: 2014 International Conference on. IEEE, Computing for Sustainable Global Development (INDIACom), 2014, pp. 515e518.
Lian Yang He was born in Anhui, China, in May 1988. He received the B$Sc. degree in electrical engineering from Hefei University of Technology, Hefei, China, in 2010. From 2010, he is studying for the Ph.D. degree in electrical engineering from Shanghai Jiao Tong University, Shanghai, China. Besides, he is now positioned as BOND researcher in Innovation and Technology Department of GE Grid Solution (former Alstom Grid), UK. His research interests conclude the control strategy of active distribution systems and smart grid.
Nengling Tai He received the B.Sc., M.Sc., and Ph.D. degrees in electrical engineering from Huazhong University of Science and Technology(HUST), Wuhan, China, in 1994, 1997, and 2000, respectively. Currently, he is the associate present and professor in the Department of Power Electrical Engineering, Shanghai Jiao Tong University, Shanghai, China.
Chunju Fan She was born in Jiangsu, China, in May 1967. She received the B.S. degree in electrical engineering from Hefei University of Technology, Hefei, China, in 1990, the
212
L. Yang et al. / Renewable Energy 95 (2016) 202e212
M.S. degree in electrical engineering from Tianjin University, Tianjin, China, in 1993, and the Ph.D. degree in electrical engineering from Shanghai Jiao Tong University, Shanghai, in 2005. Currently, she is an Associate Professor at Shanghai Jiao Tong University, where she has been since 1993.
Yuanye Meng He received his M.Eng degree (with honors) in Electrical Engineering from University of Southampton, UK. He is now positioned as a researcher in Innovation and Technology department of GE Grid Solution (former Alstom Grid), UK. His current research interest is advanced microgrid energy management.
Contents lists available at ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
Energy regulating and fluctuation stabilizing by air source heat pump and battery energy storage system in microgrid Lian Yang a, b, *, Nengling Tai a, Chunju Fan a, Yuanye Meng b a b
School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China Innovation and Technology Department, GE Grid Solution, Stafford, ST17 4LX, UK
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 November 2015 Received in revised form 28 March 2016 Accepted 8 April 2016
The energy consumption statistics of buildings have shown that in China, 50%e70% of the annual energy consumption is consumed by cooling and heating systems, the majorities are air conditions and hot water supply. To cut down the investment of BESS, this paper studies the application of Air Source Heat Pump (ASHP) in a photovoltaic/Battery Energy Storage System (PV/BESS) microgrid. Energy dispatching strategy is proposed to take advantage of the thermal storage, and fluctuation stabilization strategy is proposed based on the decouple characteristics between Heat Pumps (HPs) and terminal devices. As the characteristics of fluctuation and the inputted capacity of ASHP should be considered during the stabilization, there are too many input variables for traditional fuzzy logic algorithm. According to this, this paper proposes an improved fuzzy logic algorithm, Double Fuzzy Logic (DFL) algorithm, to involve in all the necessary variables. At last, a case is simulated to verify the feasibility of the energy dispatching and fluctuation stabilization strategy. The results have verified that ASHP can decrease the capacities of PV/ BESS and stabilize the fluctuation effectively. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Microgrid energy management Fluctuation stabilize Battery energy storage system Air source heat pump Double fuzzy logic
1. Introduction Photovoltaic (PV) technology has been regarded as a solution to produce electrical energy. In recent years, PV technology has been improved greatly to achieve high energy efficiency with high durability [1e5]. While in the past the price of the PV modules was the major contribution to the cost of the systems, a downward tendency is now seen thanks to the development of silicon technologies such as Thin Film technologies, an increasing competition among manufacturers and a massive enlargement in the production capacity of PV modules [6]. However, similarly to other renewable energy sources, solar energy tends to be unsteady because it is influenced by natural and meteorological conditions [7]. Moreover, high penetration of intermittent renewable resources can bring up technical challenges including grid interconnection, power quality, reliability, protection, generation dispatch, and control [8]. The issue of how power fluctuation in PV is to be smoothed has attracted widespread interest and attention. As the levels of penetration of renewable energy rise, the technical impact
* Corresponding author. School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China. E-mail address: [email protected] (L. Yang). http://dx.doi.org/10.1016/j.renene.2016.04.019 0960-1481/© 2016 Elsevier Ltd. All rights reserved.
of PV on grid operation has led to the application of energy storage for renewables [9]. Recent papers proposed this application include a simple scheme to charge and discharge the Energy Storage System (ESS), such as storing excess power when the solar power output exceeds a threshold and discharge it back to the grid when the load demand is high [10]. In Refs. [11], a State of Charge (SOC) based smoothing control strategy was adopted to smooth out short-term power fluctuations of wind/PV hybrid system. In Refs. [12], a generalized double-shell framework for the optimal design of renewable energy systems was developed. Optimal sizing of distributed energy resources with storage systems was studied based on the yearly Joule losses, the yearly costs (capital and management), and the yearly CO2 emissions. In Refs. [13], a bidirectional converter and control system were designed for Renewable Energy Sources (RESs) and ESSs to take part in the voltage regulation and the frequency transient regulation. Ref [14] studied the economically feasibility of installing medium-scale distributed storage devices in the power system designed to lower the electricity cost for a customer-side application, assuming flexible electricity tariffs. In addition to the capability of enabling the integration of more RESs into the network, ESSs can provide many other benefits that can be summarized as: benefits related to load/generation shifting, benefits
L. Yang et al. / Renewable Energy 95 (2016) 202e212
Nomenclature C specific heat capacity of water f output power of PV g1, g2 load curves h, h' annual cost functions K,K1,K2 correction factors PASHP power consumption of ASHP PBESS_low output of BESS for stabilization Phigh high frequency component PHP capacity of HPs PHPmax extremely capacity design of HPs Phump operating power of ASHP Plow low frequency component PN rated power of ASHP PStotal stabilization command Q1 , Q2, Q3 electric quality differences between generators and load demands SBESS State of Charge of BESS SBESSmin allowed minimum state of BESS T time constant Tf time constant of filter X(s) reference value X1,X2,X3,X4 inputs of fuzzy logic modules Y(s) power consumption Greek symbols w flow of water DT temperature difference t2, t3 time period DQloss heating loss DQCOP decreased power consumption Df frequency deviation Df/Dt frequency changing rate a; b are the normalized coefficients m1i,m2j membership values Dh increased investment of HPs DPHP increased capacity of HPs
related to ancillary services and benefits related to grid system applications. In Refs. [15], the operating and maintenance cost of ESS was studied, and an ESS management was proposed based on the power forecasting module of PV. Similarly in Refs. [16,17], a cost-benefit analysis method was proposed for battery energy storage system (BESS) when BESS was applied to microgrid with PV systems. It is preferable to keep installed ESSs as small as possible, since ESSs are generally expensive. Table 1 shows an example of several ESSs' costs per output power (kW) and costs per energy(kWh) [18]. Table 1 shows that most ESSs' costs are still quite high. Thus far, various BESS-based methods of smoothing power
203
DQBESS decreased capacity of BESS Dh' decreased investment function DPASHP_low power increment of ASHP Superscript/subscripts ASHP Air Source Heat Pump ASHP_low Low frequency stabilization of ASHP BESS Battery Energy Storage System BESS_low Low frequency stabilization of BESS COP Coefficient of Performance f filter high high component HP Heat pump hump hump value loss energy losses low low component N nominal rating max maximum min minimum QASHP heating volume SBESSmax allowed maximum state of BESS Stotal total values of stabilization t time Acronyms ASHP Air Source Heat Pump BESS Battery Energy Storage System COP Coefficient of Performance DFL Double Fuzzy Logic ESS Energy Storage System HP Heat Pump LCC Life Cycle Cost MPPT Maximum Power Point Tracking PCC Point of Common Coupling PV Photovoltaic RES Renewable Energy Source SOC State of Charge
fluctuations in renewable power generation systems have been proposed, but economic cost of BESS has not been substantially reduced. As an effective way of reducing the capacity of BESS, controlling electrical appliances on the demand side was considered in Ref. [19]. In Refs. [20], the manageable loads were studied and a new formulation of shiftable loads was employed for a new load modeling method. Manageable loads are loads for which can be modified during the operating cycle without damage and degradation to the quality of the consumer. They can either be adjustable or shiftable. The ideal manageable loads should meet several requirements:
Table 1 Example of ESSs' cost. ESS
Cost per kW [1000$/kW]
Cost per kWh [1000$/kWh]
NaS battery Li-Ion battery NiMH battery High speed flywheel Electric double layer capacitor Pumped hydro
7.6 6e15
0.4 1e4
0.5e1 2 1.5e2
120e240 50~ 0.2
204
L. Yang et al. / Renewable Energy 95 (2016) 202e212
1) Power consumption of the load is large enough to compensate power fluctuation by the control of its power consumption. 2) The response speed of the load's power consumption to its reference signal is fast enough to compensate power fluctuation. 3) The loss of convenience by the control of power consumption is minor. Among electrical loads, heat pumps are commonly used in air conditioning and water heating and so on. As heat pumps can satisfy the three requirements mentioned in previous sections, they can be treated as an ideal option for control [18]. The statistics about energy consumption of buildings have shown that in China, 50%e70% of the annual energy consumption is consumed by cooling and heating systems, most of which are air conditions and hot water supply [21]. Air Source Heat Pump (ASHP) has been widely used in China and elsewhere with the advantage of great Coefficient of Performance (COP). There are several papers verified the feasibility of frequency regulation by ASHP. In Refs. [18,19,22], control characteristics of heat pumps were studied and the capacity reduction of ESS was calculated by simulation. The basic characteristic of a heat pump's power consumption was measured experimentally in a real building. In Refs. [23], a low cost state-ofart experimental stand was researched for testing ASHP under controlled evaporator ambient conditions. This innovation can be referred by ASHP to extend the water tank, for the sake of more thermal energy storage. In Refs. [24], the operation optimization of a distributed energy system was proposed to improve the overall exergy efficiency. Thermal storages were studied and used to satisfy low-quality thermal energy demands. In Refs. [25], the thermal storage tanks for space heating were selected to minimize the use of auxiliary equipment. In Refs. [26], the COP characteristic of heat pump was studied. The research shows that COP is dependent on the outside air temperature. In Refs. [27], calculation and design of the heat pumps were proposed for purposes of heating and cooling. The model involves multiple energy devices that convert a set of primary energy carriers with different energy quality levels to meet given time-varying user demands at different energy quality levels. By promoting the usage of low-temperature energy sources to satisfy low-quality thermal energy demands, the waste of highquality energy resources can be reduced, therefore improving the overall exergy efficiency. However, due to the fact that the heat conduction of buildings is complicated, and the air temperature varies with time and season, it's difficult to build an accurate model for the stabilization control. Then the fuzzy logic comes into view with the advantages on complicated system [28e35]. When something is uncertain, like measurement, it is difficult to determine its exact value by the traditional control algorithms. Then the fuzzy logic algorithm allows handling much of this uncertainty by defining fuzzy systems employing fuzzy logic. Fuzzy logic control is simple, can be easily realized, doesn't require modeling and has strong robustness, it is more appropriate to use in applications where parameters and/or model are undefined or changeable [28]. In Refs. [29,30], fuzzy logic algorithm was used to manage the flow of energy between ESSs. In the case without using fuzzy control, the stored energy of Superconducting Magnetic Energy Storage or the secondary battery reaches the maximum rated capacity or minimum rated capacity, and power control becomes impossible in some places. By modifying the power reference value using the fuzzy control, it is possible to avoid the saturation or depletion of their output [30]. In Refs. [31e35], fuzzy logic algorithm was used to manage the flow of energy between ESS and renewable energy sources. In conclusion, fuzzy logic algorithm is suitable as the control algorithm in the complicated systems. This paper studies the application of ASHP in a PV/BESS
microgrid for energy dispatching and fluctuation stabilization. An energy dispatching strategy is studied to cut down the capacity of BESS, and the stabilization strategy is proposed based on the characteristics of PV/BESS and ASHP. As the characteristics of fluctuation and the inputted capacity of ASHP should be considered during the stabilization, there are too many input variables for traditional fuzzy logic algorithm. According to this, this paper proposes an improved fuzzy logic algorithm, Double Fuzzy Logic (DFL) algorithm, to involve in all the necessary variables. At last, a case is simulated to verify the feasibility of the energy dispatching and fluctuation stabilization strategy.
2. Structure of microgrid Structure of microgrid is shown as Fig. 1. Photovoltaic provides power energy, and BESS balances the energy supply and demand. Microgrid connects to the power grid through Point of Common Coupling (PCC). Load consumptions in the microgrid consist of two parts: heating load satisfied by ASHP; common electric load such as lighting.
2.1. Photovoltaic unit Photovoltaic panel can be seen as an electrical generator, whose maximum power depends on cell temperature and solar radiation. In order to ensure the maximum efficiency whatever the conditions of sunlight and temperature are, the photovoltaic panel is associated to a power converter controlled with a Maximum Power Point Tracking (MPPT) algorithm [1,36]. Besides, the solar radiation received by inclined photovoltaic panel is the other key parameter for the power output. Hay-Davies-Klucher-Reindl anisotropic model permits to take into account the effect of clouds on terrestrial irradiance through a clearness index and a diffuse fraction [37]. It was shown that the weather and the MPPT strategy had aggregated the output fluctuation of PV unit. Stochastic of clearness leads to the stochastic fluctuation of output power. The fluctuation should be stabilized by auxiliary devices such as energy storage devices.
2.2. Battery energy storage system RESs output the fluctuated power into microgrid, and the fluctuation requires to be stabilized by frequency regulating devices such as BESS. Assuming that State of Charge (SOC) of BESS is SBESS, SBESS should be maintained in a certain range, or else life damage will be caused to BESS. We can get:
Fig. 1. Structure of microgrid.
L. Yang et al. / Renewable Energy 95 (2016) 202e212
SBESSmin SBESS ðtÞ SBESSmax
(1)
where SBESSmin is allowed minimum state; SBESSmax is the allowed maximum state. Capacity of BESS can be decomposed into two parts: one for fluctuation stabilization, the other for peak-valley regulation. Although BESS delivers better response characteristics and low power losses, it expends larger investment. Besides, life loss will be aggravated by the stabilizing operation.
2.3. Air Source Heat Pump Among electrical appliances, heat pumps, which are used for air conditioning and water heating and so on, are attractive options for compensation of power fluctuation [18]. ASHP is composed of Heat Pump (HP), water tank (thermal storage), terminal units, etc., shown in Fig. 2. 1) HP. HP is a closed loop composed of compressor, condenser, throttling gear and evaporator. Compressor circulates the refrigerant through the loop for heat transmission. 2) Water tank. ASHP has a tank to store hot water. As ASHP is taken into account of microgrid management, the capacity of water tank is larger than the usual ones, for the propose of storing more heating energy. 3) Terminal units. Heat stored in the water tank is dispatched by the terminal units. One portion of the heat is used for air conditioning, while the other satisfies the hot water demand. The heating volume of ASHP can be expressed as follows:
QASHP ¼ wC DT
(2)
where QASHP is the heating volume; w is the flow of water; C is the specific heat capacity of water; DT is the temperature difference. Power consumption corresponding to the heating volume is expressed by COP, shown as follows:
COP ¼
QASHP PASHP
(3)
where PASHP is the power consumption of ASHP. COP of ASHP is affected by HP's performance and ambient temperature. COP ascends nonlinearly with the rising of ambient temperature [23]. It means that during the whole daytime, ASHP will consume less electric power for the same quality of heat during high-temperature period. BESS regulates the peak-valley load by charging or discharging the electric energy, while ASHP can regulate the thermal load by
205
charging or discharging the thermal energy. BESS has better response characteristics, while ASHP has greater economic efficiency. Thus, ASHP can operate as an assistance of BESS by devoting into energy dispatching and fluctuation stabilization. 3. Microgrid energy management strategy Microgrid system operates at a low voltage distribution, and has several distributed energy resources. Microgrid system also has the ability to operate connected to the power grid (on grid) or disconnected to the power grid (off grid/islanded) [38]. When operating as on grid, microgrid system can trade electric energy with the grid. In this paper, electric energy can only flow into microgrid owing to the actual difficulties from power utilities and microgrid operators. In order to reduce its influence and dependence on the power grid, microgrid purchases electric energy only during the valley price period. Power is traded based on load prediction. Microgrid predicts the solar condition and load of coming days, and then purchases electric energy to fill the vacancy. The energy flow in microgrid is shown as Fig. 3. Operating as a thermal energy storage device, water tank can regulate microgrid energy together with BESS. It means that water tank can bridge the gap between HP and thermal demand, which releases HP for controllable energy consumption. 3.1. Energy dispatching during the whole day time The energy states of BESS and water tank drop to the minimum values in the morning, measured during a whole day time. In daytime, solar radiation and ambient temperature rise up, leading to the increase of PV's output and HPs' COP. Accompanied by the rising of energy supply, BESS starts to charge energy, as well as ASHP. More electric energy will be consumed if more HPs are put into operation. HPs supply more heat than the real-time demand and then the redundant heat is stored by the water tank. At night, PV ceases, ambient temperature drops, meanwhile COP of HPs falls. A large portion of HPs are paused for less energy consumption. The water tank supplies the thermal demand, while BESS sustains ASHP and other electric loads. Fig. 4 shows the regulation principle of microgrid. In Fig. 4, f(t) is the output power of PV; g1(t) is the former load curve that ASHP operated according to the real-time demand; g2(t) is the regulated load curve that ASHP participated in the energy dispatching; Q1, Q2 and Q3 are electric quality differences between generators and load demands. If microgrid operates with g1(t), the electric quality which should be regulated by BESS is Q1 þ Q2; or else microgrid operates with g2(t), then electric quality BESS should regulate is Q1 þ Q3. Associated by ASHP, BESS can cut down its own capacity of Q2Q3. Besides, HPs with greater COP can cut down the total power consumption during the whole day time. 3.2. Optimal sizing of BESS and ASHP The rated capacity of HPs plays an important role in g2(t). Adding more HPs is an effective way to increase the rated capacity. g2(t)
Fig. 2. Structure of ASHP
Fig. 3. Energy flow in the microgrid.
206
L. Yang et al. / Renewable Energy 95 (2016) 202e212
DQBESS ¼
Z
Z ½ f ðtÞ g1 ðtÞdt
t2
½ f ðtÞ g2 ðtÞdt
(9)
t3
where, t2 is the period of f(t) being greater than g1(t); t3 is the period of f(t) being greater than g2(t). The decreased investment corresponding to DQBESS is shown as follows:
Dh0 ðQBESS Þ ¼ h0 ðQBESS þ DQBESS Þ h0 ðQBESS Þ
Fig. 4. Energy dispatching in microgrid.
should get closer to f(t), for the propose of cutting down the capacity of BESS. Besides, g2(t) should be planned based on such premise that user demands are satisfied.
Z24
Z24
g2 ðtÞdt DQloss þ DQCOP
g1 ðtÞdt ¼ 0
(4)
0
where, DQloss is heating loss; DQCOP decreased power consumption caused by different COP. The extremely capacity design of HPs, PHPmax, is that: (1), the HPs work during the period of g2(t) < f(t), and then all the HPs cease during the period of g2(t)>f(t); or (2), all the energy of f(t)(having subtracted the energy for other electric load) are consumed by ASHP during the period of g2(t) < f(t). However, this will increase the investment of HPs. Hence the balance between the increasing investment of HPs and the decreasing investment of BESS should be discussed. The increased investment Dh of HPs with the increased capacity DPHP is shown as follows:
Dh ¼ hðPHP þ DPHP Þ hðPHP Þ
(5)
where, Functionh() is the annual cost of ASHP, it was calculated by converting the Life Cycle Cost (LCC) to each year; PHP is the capacity of HPs. The decreased power consumptionDQCOP with DPHP is shown as follows:
DQCOP ¼ DW$
1 1 COP2 COP1
(6)
where, DW is the transferred heat from night time to day time (convert to each year); COP1 is the average COP during the day time; COP2 is the average COP during the night time. Water tank should be expended for more energy storage. The expanded capacity is assumed as DQwater:
DQwater ¼
K$COP1 $ cwater $DT
Z ½g2 ðtÞ g1 ðtÞdt
where Function h0 ðÞ is the annual cost, it was also calculated by converting LCC to each year. Comparing the increased investment of ASHP and the decreased investment of BESS, we can get the optimization function shown as equation (11):
minDhðASHPÞ Dh0 ðQBESS Þ
(11)
According to Equation (11), the capacity of ASHP and BESS can be optimized for lower investment. 3.3. Fluctuation stabilized by BESS and ASHP Stochastic of clearness leads to the output fluctuation of PV, energy consumption of microgrid loads will also fluctuate due to the users' stochastic behaviors. BESS had been working on energy stabilization in many microgrids. With the controllable merit provided by water tank, HPs can regulate the power consumption as the response to fluctuation and the assistance to BESS. In the HP unit, there is a valve that controls the refrigerant flow. The control of the valve is carried out via hydraulic system. The HP unit has two external terminals that correspond to “Load Up” and “Load Down”. The signal given to these terminals controls the valve and changes the amount of refrigerant. As a result, the power consumption of HP is changed. Power consumption characteristic of HP is modeled using a first order transfer function, shown as Equation (12).
YðsÞ ¼
1 XðsÞ 1 þ Ts
(12)
Where, Y(s) is the power consumption, X(s) is the reference value and T is the time constant of the model. The daily peak-valley characteristics of microgrid load should be regulated by ASHP and BESS together, moreover the frequency fluctuation also requires to be stabilized. During the stabilization period in microgrid, the assistance of ASHP can cut down the charge/discharge current of BESS, which means that the loss of BESS is lowed. 4. Double Fuzzy Logic model
(7)
t1
where, K is the margin coefficient, cwater is heating ratio of water; DT is the temperature difference between unheated water and heated water; t1 is the period of g2(t) being greater thang1(t). The increased investment of ASHP corresponding to the increased capacity of DPHP is shown as follows:
DhðASHPÞ ¼ hðPHP þ DPHP ; DQCOP ; DQwater Þ hðPHP Þ
(10)
(8)
where, Functionh() is the annual cost of ASHP, it was calculated by converting the Life Cycle Cost (LCC) to each year. The decreased capacity DQBESS of BESS is shown as follows:
This section aims at DFL model for frequency fluctuation. Fluctuation caused by the randomness of PV and load should be stabilized. Assuming thatPStotal is the stabilization command: positive PStotal needs BESS and ASHP to supply energy; negative PStotal needs BESS and ASHP to absorb energy. BESS is the non-inertial device with fast response, while ASHP is the inertial device with less responding rate. Thus the high frequency fluctuation can only be compensated by BESS, while the low frequency fluctuation can be stabilized by BESS and ASHP together. High-pass filter is employed to pick up high frequency componentPhigh fromPStotal, andPhigh is responded by BESS. The rest fluctuation stabilizing command, assumed asPlow, is responded by both ASHP and BESS, shown as equations (13) and (14):
L. Yang et al. / Renewable Energy 95 (2016) 202e212
Phigh ðsÞ ¼
sTf P ðsÞ 1 þ sTf Stotal
Plow ðsÞ ¼ PStotal ðsÞ Phigh ðsÞ ¼
(13)
1 P ðsÞ 1 þ sTf Stotal
(14)
whereTf is the time constant of filter. When analyzing the response dispatch between BESS and ASHP, operating characteristics of each device should be considered for the charge-discharge constraints under different states. The stabilization command should be dispatched based on the characteristics of BESS and ASHP, hence the control objectives of DFL are shown as follows: (1) The required capacity and charge/discharge current of BESS should be limited as possible, so as to cutting down the loss of BESS and the investment of microgrid. (2) If SOC of BESS reaches the upper limit, energy absorption should be reduced during the charging state, and energy release should be encouraged during the discharging state. Otherwise, SOC of BESS reaches the lower limit, energy absorption should be encouraged during the charging state, and energy release should be reduced during the discharging state. (3) If HPs run under heavy load, unloading is encouraged; or else if HPs run under fractional load, load increasing is encouraged. There are four inputs of DFL, one for BESS, one for ASHP, the others for the stabilization dispatch. Energy states of each storage device are picked up as two inputs. State of Charge of BESS at time t, SOC(t), is assigned as one input of DFL. Load coefficient of ASHP, Phump(t)/PN, is assigned as another input, where Phump(t) is the operating power at time t; and PN is the rated power of ASHP. The other two inputs should be assigned to the variables that reflect the energy fluctuation in microgrid. The fluctuation Plow, related to the frequency deviation Df in microgrid, is picked up as the total stabilization command of BESS and ASHP. The frequency changing rate Df/Dt is picked up as one of the dispatch coefficient. Then we can get the expressions of the four inputs:
X1 ðtÞ ¼ SOCðtÞ
(15)
X2 ðtÞ ¼ Plow ðtÞ
(16)
X3 ðtÞ ¼
Phump ðtÞ PN
(17)
X4 ðtÞ ¼
Df Dt
(18)
Fig. 5 shows the DFL model for stabilization dispatch. Inputs X1~X4 are processed by Fuzzy Logic submodules Fuzzy1 and Fuzzy2 respectively, and output the correction factors K1 and K2 to weighting function. After weighted by coefficient a (0 < a < 1), we can get the ultimate value K of DFL.
a¼
aCBESS CBESS þ bCASHP
(19)
Where, CBESS is the inputted capacity of BESS for stabilization; CASHP is the inputted capacity of ASHP; a and b are the normalized coefficients, they are set considering capacities transforming and economic weights. Submodule Fuzzy1: Fuzzy 1 inputs [X1,X2] and outputs K1. If the value of Plow(t) is appropriate and BESS has enough margin for stabilization (charging or discharging), Fuzzy Logic rule will turn down the correction factor K1. Or else, BESS has not enough margin, Fuzzy Logic rule will turn up the correction factor K1, and ASHP will undertake more stabilization assignment. Submodule Fuzzy2: Fuzzy 2 inputs [X3,X4] and outputs K2. It adjusts the stabilization command based on the operating state of ASHP and the changing rate of fluctuation. Firstly, heavy load state of ASHP is discussed. If the value of X4 is negative, Fuzzy Logic rule will turn up the output K2 to remit the pressure of ASHP; otherwise X4 is positive, K2 will be turned down. Secondly, under load of ASHP is discussed. If the value of X4 is negative, K2 will be turned down; otherwise, K2 will be turned up. The primary elements of a fuzzy logic model are (1) fuzzy sets, (2) membership function (3) IF-THEN rules [39]. A fuzzy set is labeled by a linguistic term, and the linguistic term is a word such as “NB”, “NM”, etc. The fuzzy sets and IF-THEN rules construct a rule base of fuzzy logic model. Membership function is the effective method for comprehensive evaluation based on multi factors. Membership function is a curve that defines how each point in the input space is mapped to a membership value between 0 and 1. It provides a measure for the degree of an element to a fuzzy set. It fully defines fuzzy set and it can take any form, but there are some common examples that appear in real applications. There are different shapes of membership functions: triangular, trapezoidal, piecewise-linear, Gaussian and so on. In most literature, triangular membership functions are used due to having flexible formula to calculate [40]. Thus the triangular shape is employed in this paper. Membership functions of Fuzzy 1 and Fuzzy 2 are shown as Fig. 6 (a) and (b) respectively. The rule of Fuzzy Logic will affect the control result directly. Fig. 6 has shown the inputs, outputs and membership functions of two Fuzzy Logic submodules. The fuzzy rules of them are shown as Tables 2 and 3. Deduction results of fuzzy rules are fuzzy sets, which should be cleared by weighting method. Correction factor K1 is calculated by Equation (20):
PP K1 ¼
i
j
m1i ðX1 ðtÞÞm2j ðX2 ðtÞÞDkij
PP i
Fig. 5. Double Fuzzy Logic model.
207
j
m1i ðX1 ðtÞÞm2j ðX2 ðtÞÞ
(20)
Where, m1i ðX1 ðtÞÞ and m2j ðX2 ðtÞÞ are membership values of the inputs X1(t) and X2(t) respectively; the set of i is {NL, NM, ZO, PM, PL}; set of j is {NB, NM, NS, ZO, PS, PM, PB}; Dkij is the output corresponding to m1i ðX1 ðtÞÞ and m2j ðX2 ðtÞÞ. The calculation of correction factor K2 is similar to that of K1, thus is omitted for simplification. Stabilization command of BESS and ASHP should be coordinated in the DFL, hence we can get:
208
L. Yang et al. / Renewable Energy 95 (2016) 202e212
Fig. 6. Membership functions of DFL.
Table 2 Rule base of Fuzzy 1. X1
NB NM ZO PM PB
DPASHP
NB NM NS ZO PS PM PB
NB
NM
NS
ZO
PS
PM
PB
PB PM ZO NM NB
PM PS ZO NS NM
PS PS ZO NS NS
ZO ZO ZO ZO ZO
NS NS ZO PS PS
NM NS ZO PS PM
NB NM ZO PM PB
NB
NM
NS
ZO
PS
PM
PB
NB NM NS ZO NS ZO PS
NB NS ZO PS ZO PS PM
NB ZO PS PM PS PM PB
PB PB PB PB PB PB PB
PB PM PS PM PS ZO NB
PM PS ZO PS ZO NS NB
PS ZO NS ZO NS NM NB
X4
Plow ðtÞ ¼ PBESS
low ðtÞ
þ DPASHP
low ðtÞ
(21)
Where, PBESS_low(t) is the output power of BESS for stabilization, positive value is assumed under charging model; DPASHP_low(t) is the power increment of ASHP, shown as Equation (22):
DPASHP
low ðtÞ
¼ DPASHP
¼ PASHP ðtÞ PASHP ðt 1Þ
(22)
Stabilization command betweenPBESS_low(t) and DPASHP_low(t) is dispatched by the DFL, shown as follows:
PBESS
low ðtÞ
1Þ þ Plow ðtÞ low ðt 1Þ $KðtÞ
low ðt
DPASHP
X2
Table 3 Rule base of Fuzzy 2. X3
low ðtÞ
¼ Plow ðtÞ DPASHP
low ðt
1Þ $½1 KðtÞ
(23) (24)
According to the equations (23) and (24), BESS and ASHP stabilize the fluctuation based on their own operating status. The DFL model set forth above has involved the SOC of BESS with Fuzzy 1, and considered the output of ASHP with Fuzzy 2. Then K1 and K2 are normalized by weighting coefficient. At last, DFL outputs the factor K for stabilization.
5. Case analyses This paper analyzed the actual microgrid case in Meizhou Island in China. The island is supplied through a 15 km submarine cable, which has caused great losses. Local photovoltaic generators are required to replace the diesel generators and decrease the electric transmission losses. The capacities of PV and BESS were designed for the energy balance during only one day period, because the microgrid can be sustained through the submarine cable, the continuous rainy weather principle can be ignored. Since the advantages of ASHP to other heating devices were verified in many literature, this section only analysis different operation strategies of ASHP. The simplified dynamic model of the microgrid is shown as Fig. 7. Parameters are shown as Table 4. Simulation results are shown as Figs. 8e15. Fig. 8 shows the statistical load profiles of the microgrid. The statistics were counted by the average data during the winter period. And they were counted according to the local communities, containing the resident buildings, schools and hotels, provided by Zhongnan Construction Group. Power consumption of the whole day was
L. Yang et al. / Renewable Energy 95 (2016) 202e212
209
Fig. 7. Simplified model of the microgrid.
Table 4 Parameters of the microgrid model. Hu t Kpt Kit Vu t Tcon Tf Kb
5.19 3 0.6 1 0.02 5 380
Tb R1 k1 Trh Fhp Trh
0.04 0.04 0.03 30 0.3 7
12 MWh, while the peak and valley values were 0.75 MW and 0.25 MW. It can be seen that during the whole day time, heating load occupied about 60% of total load. Besides, heating load occupied much more during evening time. At that time demands were supplied by BESS. Then capacity of PV was designed as 12 MWh/ 2.5 MWp, for the strategy verification. Fluctuation characteristics of PV can be seen from Fig. 9, where the random data of illumination were referred to the local meteorology. Fig. 10 shows a typical load curve during the whole day time. This load curve expresses the peak-valley characteristic and random characteristic. Load demands from 17:00 to 24:00 were quite great. The heating demand was satisfied by ASHP in real-time. Only BESS was applied for regulation and stabilization. SOC of BESS is shown as Fig. 11. It can
Fig. 9. Output characteristics of PV.
Fig. 10. Typical load curve of the microgrid.
Fig. 8. Statistical load curves of the microgrid.
be seen that the variance of SOC during the whole day time was about 7 MWh. The variance value of SOC can be used to design the installed capacity of BESS, based on the planned upper limit and lower limit (in terms of percentage). Then, ASHP was applied to the microgrid. ASHP regulated the cooling load consumption by ice tankers, and the regulated load is shown as Fig. 12. ASHP supplied more quantity of heat to the water tank during the day time, when the value of COP was higher. A part of energy demand was
210
L. Yang et al. / Renewable Energy 95 (2016) 202e212
Fig. 11. SOC of BESS in the microgrid without ASHP.
Fig. 14. Stabilization dispatch between BESS and ASHP.
Fig. 12. Regulated load curve of the microgrid.
Fig. 15. SOC of BESS in the microgrid with ASHP.
transferred from night time to day time. Besides, ASHP supplied fluctuation stabilization together with BESS. Fig. 13 shows the response characteristics of the two devices. The response speed of BESS was much faster than that of ASHP, owing to that time constant of BESS was much smaller. Thus the high-frequency fluctuation can only be stabilized by BESS. The stabilization was dispatched under the DFL strategy expressed in Section 4, and the results are shown as Fig. 14. It can be seen that stabilization of BESS varied faster and more severe than that of ASHP. Since ASHP undertook partial fluctuation and energy storage, BESS gave a more stable and less variance performance on SOC, shown in Fig. 15. During the first design (capacity of PV was 12 MWh/2.5 MWp), values of SOC at 0:00 and 24:00 are 3 MWh and 5 MWh respectively. These results were due to the fact that ASHP exhausted less power under greater COP value. During the second design, the capacity of PV was designed as 10 MWh/2.1 MWp, so that energy balance of microgrid could be maintained. Then the variance of SOC decreased to 3 MWh, and 4 MWh of batteries could be cut down or dropped out as reserve. Correspondingly, the expended capacity of
Fig. 13. Bode diagram of BESS and ASHP.
water tank was about 42 m3. 6. Conclusions This paper has proposed a microgrid energy management strategy to reduce the cost of photovoltaic and Battery Energy Storage System. Air Source Heat Pump had been introduced to a PV/ BESS microgrid for energy dispatching and fluctuation stabilization. In the greater power grid, the peak-valley load can be regulated by the time-of-use price or pumped-storage power station to achieve a lower unit regulation price. In the microgrid however, renewable energy has the fluctuation characteristics; load is smaller and has the obviously peak-valley characteristics. As a result, only Battery Energy Storage System had been widely applied. If the residential load occupies a large proportion, heating load will directly increase the capacity of Battery Energy Storage Systems. The application of Air Source Heat Pump decreases the capacities of photovoltaic and batteries. Besides, assisted by Air Source Heat Pump for stabilizing the power fluctuation, losses of battery energy storage system are decreased. Consequently, the application of ice storage air condition in microgrid has greater operation reliability and economic efficiency. Different from photovoltaic, the output characteristics of wind generators are determined by the local meteorological conditions. If the wind energy is abundant during day time and lacking during night period, which is similar to the load consumption characteristics, then the wind farm is suitable for the strategy proposed in this paper. Otherwise, once the output of wind generators is not similar to the load consumption characteristics, the wind generators are not suitable for this strategy. This paper considers a microgrid connected with a greater grid, which is different from the stand-alone microgrid. Capacities of photovoltaic and Battery Energy Storage System should be designed based on the continuous rainy weather condition. Application of Air Source Heat Pump in the stand-alone microgrid
L. Yang et al. / Renewable Energy 95 (2016) 202e212
should be expended based on the strategy proposed in this paper. In addition, as shown in statistics, cooling and heating loads occupy 50%e70% of the total power consumption in buildings. There are other manageable loads that can also be used in microgrid energy management. Further studies have been started on Ice Storage Air Condition (ISAC), which is one of the potential solutions for microgrid energy management. Acknowledgement This research was supported by Science and technology project of Education Ministry, China, No.113023A. The authors would like to thank Hengxu HA at GE Grid Solution; Jinshi Chen at Zhongnan Construction Group and Chengen Wu at Fujian electric power research institute, for their assistance in getting the required data to carry out this research. References [1] A. Bouabdallah, J.C. Olivier, S. Bourguet, M. Machmoum, E. Schaeffer, Safe sizing methodology applied to a standalone photovoltaic system, Renew. Energy 80 (2015) 266e274. [2] R. Bakhshi, J. Sadeh, H.R. Mosaddegh, Optimal economic designing of gridconnected photovoltaic systems with multiple inverters using linear and nonlinear module models based on Genetic algorithm, Renew. Energy 72 (2014) 386e394. [3] A.M. Dizqah, A. Maheri, K. Busawon, An accurate method for the PV model identification based on a genetic algorithm and the interior-point method, Renew. Energy 72 (2014) 212e222. [4] M. Mohsenzadeh, R. Hosseini, A photovoltaic/thermal system with a combination of a booster diffuse reflector and vacuum tube for generation of electricity and hot water production, Renew. Energy 78 (2015) 245e252. [5] P.M. Almeida, P.G. Barbosa, J.G. Oliveira, Digital proportional multi-resonant current controller for improving grid-connected photovoltaic systems, Renew. Energy 76 (2015) 662e669. [6] G. Graditi, D. Colonnese, N. Femia, Efficiency and reliability comparison of DCDC converters for single phase grid connected photovoltaic inverters, in: Power Electronics Electrical Drives Automation and Motion (SPEEDAM), 2010 International Symposium on. IEEE, 2010, pp. 140e147. [7] X.Y. Wang, V.D. Mahinda, S.S. Choi, Determination of battery storage capacity in energy buffer for wind farm, IEEE Trans. Energy Convers. 23 (3) (2008) 868e878. [8] S. Teleke, M.E. Baran, A. Huang, S. Bhattacharya, Control strategies for battery energy storage for wind farm dispatching, IEEE Trans. Energy Convers. 24 (3) (2009) 725e732. [9] T. Sercan, E.B. Mesut, B. Subhashish, A.Q. Huang, Rule-based control of battery energy storage for dispatching intermittent renewable sources, IEEE Trans. Sustain. energy 1 (3) (2010) 117e124. [10] M. Mehos, D. Kabel, P. Smithers, Planting the seed, IEEE Power Energy Mag. 7 (3) (2009) 55e62. [11] L. Xiangjun, H. Dong, L. Xiaokang, Battery energy storage station (BESS)-Based smoothing control of photovoltaic (PV) and wind power generation fluctuations, IEEE Trans. Sustain. Energy 4 (2) (2013) 464e473. [12] M.L. Di Silvestre, G. Graditi, E.R. Sanseverino, A generalized framework for optimal sizing of distributed energy resources in micro-grids using an indicator-based swarm approach. Industrial Informatics, IEEE Trans. 10 (1) (2014) 152e162. [13] Graditi G, Ippolito M G, Telaretti E, et al. An innovative conversion device to the grid interface of combined RES-based generators and electric storage systems. Ind. Electron. IEEE Trans.. 62(4): 2540e2550. [14] G. Graditi, M.G. Ippolito, E. Telaretti, et al., Technical and economical assessment of distributed electrochemical storages for load shifting applications: an Italian case study, Renew. Sustain. Energy Rev. 57 (2016) 515e523. [15] C. Chen, S. Duan, T. Cai, et al., Smart energy management system for optimal microgrid economic operation, Renew. Power Gener. IET 5 (3) (2011) 258e267. [16] A. Nottrott, J. Kleissl, B. Washom, Energy dispatch schedule optimization and cost benefit analysis for grid-connected, photovoltaic-battery storage systems, Renew. Energy 55 (2013) 230e240. [17] Y. Yang, N. Yang, H. Li, Cost-benefit study of dispersed battery storage to increase penetration of photovoltaic systems on distribution feeders, in: IEEE PES General Meeting Conference & Exposition, 2014, pp. 1e5. [18] S. Kawachi, J. Baba, H. Hagiwara, Energy capacity reduction of energy storage system in microgrid by use of heat pump: characteristic study by use of actual machine, in: IEEE 14th Power Electronics and Motion Control Conf, vol. 11, 2010, pp. 52e58. [19] S. Kawachi, H. Hagiwara, J. Baba, Modeling and simulation of heat pump air conditioning unit intending energy capacity reduction of energy storage system in microgrid, in: IEEE Power Electronics and Applications Conf, 2011,
211
pp. 1e9. [20] G. Graditi, M.L. Di Silvestre, R. Gallea, et al., Heuristic-based shiftable loads optimal management in smart micro-grids, Ind. Inf. IEEE Trans. 11 (1) (2015) 271e280. [21] GB 50189-2005 Design Standard for Energy Efficiency of Public Buildings [S], 2005 (in Chinese). [22] S. Kawachi, H. Hagiwara, J. Baba, Feasibility study on short period power fluctuation compensation using heat pump air conditioning system, J. Int. Counc. Electr. Eng. 2 (2) (2012) 139e145. [23] V.S. Cr aciun, C. Bojesen, M.B. Blarke, A flexible and low cost experimental stand for air source heat pump for Smart Buildings, in: Optimization of Electrical and Electronic Equipment (OPTIM), 13th Int. Conf. IEEE, 2012, pp. 1040e1045. [24] M. Di Somma, B. Yan, N. Bianco, et al., Operation optimization of a distributed energy system considering energy costs and exergy efficiency, Energy Convers. Manag. 103 (2015) 739e751. [25] W. Gu, Z. Wu, R. Bo, Modeling, planning and optimal energy management of combined cooling, heating and power microgrid: a review, Int. J. Electr. Power Energy Syst. 54 (2014) 26e37. [26] S. Obara, S. Watanabe, B. Rengarajan, Operation method study based on the energy balance of an independent microgrid using solar-powered water electrolyzer and an electric heat pump, Energy 36 (8) (2011) 5200e5213. [27] J. Perko, V. Dugec, D. Topic, Calculation and design of the heat pumps. Energetics (IYCE), in: Proceedings of the 3rd Int. Youth Conf. IEEE, 2011, pp. 1e7. [28] S. Faquir, A. Yahyaouy, H. Tairi, J. Sabor, A type-1 fuzzy logic algorithm to manage the flow of energy in a stand-alone PV/wind/battery hybrid system, Intell. Syst. Comput. Vis. ISCV (2015) 1e6. [29] M. Mohammedi, M. Becherif, A. Aboubou, O. Kraa, Fuzzy logic and passivity based control applied to hybrid DC power source using fuel cell and battery, in: Systems and Control (ICSC), 4th Int. Conf, 2015, pp. 510e515. [30] T. Ise, M. Kita, A. Taguchi, A hybrid energy storage with a SMES and secondary battery, IEEE Appl. Supercond. 15 (2) (2005) 1915e1918. [31] D.O. Elamine, E.H. Nfaoui, B. Jaouad, Multi-agent system based on fuzzy control and prediction using NN for smart microgrid energy management, Intell. Syst. Comput. Vis. ISCV (2015) 1e6. [32] S. Zhang, Y. Mishra, M. Shahidehpour, Fuzzy-logic based frequency controller for wind farms augmented with energy storage systems, IEEE Power Syst. 99 (2015) 1e9. [33] P.I. Gonzalez, G.A.J. Calderon, M. Calderon Godoy, A. Ramiro Gonzalez, Experimental automation platform of stand-alone hybrid renewable energy systems: fuzzy logic application and exergy analysis, in: 6th Int. Renewable Energy Congress (IREC), 2015, pp. 1e6. [34] N.L. Diaz, W. Dan, T. Dragicevic, J.C. Vasquez, J.M. Guerrero, Fuzzy droop control loops adjustment for stored energy balance in distributed energy storage system, in: 9th Int. Conf. Power Electronics and ECCE Asia (ICPE-ECCE Asia), 2015, pp. 728e735. [35] S. Zhang, Y. Mishra, M. Shahidehpour, Fuzzy-logic based frequency controller for wind farms augmented with energy storage systems, IEEE Power Syst. 99 (2015) 1e9. [36] A. Bouabdallah, S. Bourguet, J.C. Olivier, Optimal sizing of a stand-alone photovoltaic system, in: IEEE Int. Conf. Renewable Energy Research and Applications (ICRERA), 2013, pp. 543e548. [37] A. Bouabdallah, S. Bourguet, J.C. Olivier, Photovoltaic energy for the fixed and tracking system based on the modeling of solar radiation, IEEE Conf. Ind. Electron. Soc. (2013) 1821e1826. [38] B.S. Hartono, Y. Budiyanto, R. Setiabudy, Review of microgrid technology, IEEE Int. Conf. Qual. Res. QiR (2013) 127e132. [39] A.M. Aziz, A new perspective on the choice of fuzzy membership functions in multitarget tracking systems, in: Aerospace Conference, 2015 IEEE. IEEE, 2015, pp. 1e8. [40] E.B. Kumar, M. Sundaresan, Edge detection using trapezoidal membership function based on fuzzy's mamdani inference system, in: 2014 International Conference on. IEEE, Computing for Sustainable Global Development (INDIACom), 2014, pp. 515e518.
Lian Yang He was born in Anhui, China, in May 1988. He received the B$Sc. degree in electrical engineering from Hefei University of Technology, Hefei, China, in 2010. From 2010, he is studying for the Ph.D. degree in electrical engineering from Shanghai Jiao Tong University, Shanghai, China. Besides, he is now positioned as BOND researcher in Innovation and Technology Department of GE Grid Solution (former Alstom Grid), UK. His research interests conclude the control strategy of active distribution systems and smart grid.
Nengling Tai He received the B.Sc., M.Sc., and Ph.D. degrees in electrical engineering from Huazhong University of Science and Technology(HUST), Wuhan, China, in 1994, 1997, and 2000, respectively. Currently, he is the associate present and professor in the Department of Power Electrical Engineering, Shanghai Jiao Tong University, Shanghai, China.
Chunju Fan She was born in Jiangsu, China, in May 1967. She received the B.S. degree in electrical engineering from Hefei University of Technology, Hefei, China, in 1990, the
212
L. Yang et al. / Renewable Energy 95 (2016) 202e212
M.S. degree in electrical engineering from Tianjin University, Tianjin, China, in 1993, and the Ph.D. degree in electrical engineering from Shanghai Jiao Tong University, Shanghai, in 2005. Currently, she is an Associate Professor at Shanghai Jiao Tong University, where she has been since 1993.
Yuanye Meng He received his M.Eng degree (with honors) in Electrical Engineering from University of Southampton, UK. He is now positioned as a researcher in Innovation and Technology department of GE Grid Solution (former Alstom Grid), UK. His current research interest is advanced microgrid energy management.