A two-tier dispatch control method for microgrids considering the state of charge of hybrid energy storage

Abstract

The present invention relates to the field of microgrid energy management and dispatch control, and discloses a two-layer dispatch control method for microgrids that considers the state of charge of hybrid energy storage. The method comprises: obtaining historical data on photovoltaic and wind power generation, load power demand, hybrid energy storage power limits, and hybrid energy storage state of charge limits; using a combined prediction method of gray GM(1,N) and BP neural network in the upper-layer control to establish an upper-layer MPC model for the microgrid; estimating the initial value of the hybrid energy storage state of charge (#imgabs0#) under day-ahead dispatch and applying it to the lower-layer control; using a dynamic programming algorithm in the lower-layer control to optimize the hybrid energy storage charge and discharge power control, outputting the hybrid energy storage state of charge to the upper-layer control as a state variable; and solving the upper-layer MPC model of the microgrid based on the objective function and constraints. Compared with existing technologies, the present invention achieves the prediction of renewable energy output and load demand, and the rolling optimization dispatch of microgrid grid connection, effectively reducing grid power fluctuations.

Description

Technical Field

[0001] The present invention relates to the field of microgrid energy management and dispatching control, and in particular to a microgrid double-layer dispatching control method considering the charge state of hybrid energy storage. Background Art

[0002] The use of photovoltaic and wind power resources in microgrids improves the absorption of renewable energy and provides convenience for power grids and users. Due to the intermittent and random nature of renewable energy generation such as photovoltaic and wind power, there is a mismatch between power supply and demand after they are connected to the microgrid. Hybrid Energy Storage System (HESS) in different forms, such as batteries and supercapacitors, is used in microgrids to combine energy storage with power storage to meet load power demand and improve the stability of grid operation. Since HESS integrates energy storage devices with different characteristics, whether a reasonable scheduling strategy can be adopted to fully utilize the advantages of various energy storage components while comprehensively considering factors such as power generation and consumption balance, microgrid operation economy, and source-load forecast results to formulate a reasonable scheduling plan is the key to achieving efficient and stable operation.

[0003] Currently, the scheduling of microgrids containing HESSs only focuses on improving the optimization algorithm, without considering the real-time allocation of different energy storage devices under changes in renewable energy output and load demand. In addition, the initial state of charge (SOC) of the energy storage elements is set to a fixed value during the control process, ignoring the different requirements of the hybrid energy storage SOC initial value due to changes in renewable energy output and load demand, resulting in reduced system operation economy. Summary of the Invention

[0004] Purpose of the invention: In response to the problems existing in the prior art, the present invention provides a two-layer dispatching and control method for microgrids taking into account the state of charge of hybrid energy storage. The upper layer adopts a predictive control model, combined with the prediction results of renewable energy output and load demand at different time scales, to achieve timely tracking and dispatching of microgrid grid connection. On the basis of the upper-layer power dispatch, the lower layer is based on the initial value estimation of the hybrid energy storage SOC, and adopts a dynamic programming algorithm to optimize the hybrid energy storage power distribution of lead-acid batteries and supercapacitors, thereby suppressing the power fluctuation of the microgrid grid connection and improving the economic efficiency of hybrid energy storage charging and discharging.

[0005] Technical solution: The present invention provides a two-tier dispatching control method for a microgrid considering the state of charge of a hybrid energy storage, comprising the following steps:

[0006] S1: Obtain historical data on photovoltaic and wind power generation, load power demand, hybrid energy storage power limit, and hybrid energy storage state of charge limit;

[0007] S2: In the upper control, the grey GM (1, N) and BP neural network combined prediction method is used to establish the upper prediction control model of the microgrid, that is, the upper MPC model of the microgrid. The total power P of the hybrid energy storage system HESS in the upper MPC model of the microgrid is C (k) used to guide the hybrid energy storage charging and discharging power control in the lower-level control;

[0008] S3: estimating the initial state of charge (SOC0) of the hybrid energy storage under day-ahead scheduling, and using the initial state of charge (SOC0) of the hybrid energy storage in the lower-level control of step S4;

[0009] S4: In the lower control layer, dynamic programming algorithm is used to optimize the charge and discharge power control of the hybrid energy storage, and the charge state of the hybrid energy storage is S C (k) Output to the upper control in step S2 as a state variable;

[0010] S5: Solve the upper-level MPC model of the microgrid based on the objective function and constraint limit conditions.

[0011] Furthermore, in step 2, the combined prediction method of grey GM (1, N) and BP neural network is used to obtain the photovoltaic and wind power prediction values ​​within the forward prediction period. The prediction period T s In the prediction model, the photovoltaic and wind power forecast values ​​are received. There are N rolling optimizations in the prediction period. The state space expression of the upper-layer MPC model of the microgrid is:

[0012]

[0013] Among them, the forward prediction period of the microgrid upper MPC model is T s , the prediction time interval is Δt, there are N sampling points in the forward prediction period, according to the microgrid power balance equation, considering the SOC change of HESS, the MPC state space model is established, in which the grid-connected power P at the current moment k is selected G (k), HESS total power P C (k) SOC value of HESS C (k) is the state variable x(k), HESS power change ΔP C (k) is the control variable u(k), and the ultra-short-term power change of photovoltaic and wind power output ΔP PV (k), ΔP WT (k) and the ultra-short-term change in load demand ΔP L (k) is the disturbance variable d(k), S BC 、E SC are the capacities of the battery and supercapacitor respectively, S BC (k), S SC (k) are the SOC values ​​of the battery and supercapacitor at the current sampling moment.

[0014] To put it another way, the initial value formula of the hybrid energy storage charge state in S3 is:

[0015]

[0016] Among them, SOC k is the state of charge of HESS at the kth moment, E C，0 is the capacity value of HESS at the initial moment.

[0017] Furthermore, the dynamic planning and scheduling method for the hybrid energy storage charging and discharging state in the lower layer control of S4 is:

[0018] According to the HESS initial state of charge SOC0 calculated in S3, the sampling interval is 1 minute, the core charge state change at the sampling moment is ΔS, and the charge state transfer equation from state 1 at time k to time k+1 is:

[0019] f(k+1)=min{|f(k,l)+P eq (k)+P l |}

[0020] Among them, P eq (k) is the net load at the kth moment, P l is the charge / discharge power in the first stage.

[0021] Furthermore, in S5, different objective functions are combined into a single objective function using penalty coefficients for solution, and the objective function is:

[0022]

[0023] Among them, α and β are penalty coefficients, which are used to adjust the weights of different indicators; ΔP G Indicates the grid-connected fluctuation exceeding the limit value, k represents the current moment, ΔS C is the deviation between the SOC of HESS and the target value; N is the number of sampling points in the look-ahead period, N = Ts / Δt.

[0024] Beneficial effects:

[0025] 1. The present invention establishes upper and lower layer control strategies, realizes the rolling optimization scheduling of microgrid grid connection based on the forecast of renewable capacity output and load demand, and effectively reduces grid power fluctuation.

[0026] 2. The initial SOC scheduling value of the HESS determines the HESS's ability to supply power and absorb surplus energy on that day. The present invention introduces the initial value estimation of the hybrid energy storage state into the lower-level control to improve the consumption of new energy and enhance the microgrid's ability to cope with peak loads.

[0027] 3. The present invention uses a dynamic programming algorithm to coordinate the charging and discharging processes and power distribution of batteries and supercapacitors in the HESS, reducing the variation range of the SOC of the energy storage unit and avoiding overcharging and over-discharging of the energy storage. It makes online corrections based on real-time conditions, prioritizes charging and discharging, and adjusts the results to make power distribution more reasonable, thereby improving the utilization rate of the HESS. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a block diagram of the double-layer control strategy of the present invention;

[0029] Figure 2 This is a flow chart of the double-layer control dynamic programming of the present invention;

[0030] Figure 3 This is a graph showing the power fluctuation curve of the microgrid connected to the grid according to an embodiment of the present invention;

[0031] Figure 4 This is a charge and discharge curve diagram of a battery according to an embodiment of the present invention;

[0032] Figure 5 This is the charge and discharge curve of the supercapacitor according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention.

[0034] The present invention discloses a two-layer dispatching and control method for microgrids that considers the state of charge of hybrid energy storage. The upper layer adopts a predictive control model (MPC) and combines the prediction results of renewable energy output and load demand at different time scales to achieve timely tracking and dispatching of microgrid grid connection. Based on the upper layer power dispatch, the lower layer uses a dynamic programming algorithm to optimize the hybrid energy storage power distribution of lead-acid batteries and supercapacitors based on the initial value estimation of the hybrid energy storage SOC, suppressing the microgrid grid connection power fluctuation and improving the charging and discharging economy of the hybrid energy storage. The method comprises the following steps:

[0035] S1: Obtain historical data on photovoltaic and wind power generation, load power demand, hybrid energy storage power limit, and hybrid energy storage state of charge limit.

[0036] S2: In the upper control, the combined prediction method of grey GM (1, N) and BP neural network is used to establish the upper MPC model of the microgrid. The total power P of HESS in the upper MPC model of the microgrid is C (k) Used to guide the hybrid energy storage charging and discharging power control in the lower-level control.

[0037] The combined prediction method of grey GM (1, N) and BP neural network is used to obtain the predicted values ​​of photovoltaic and wind power within the forward prediction period. s In the prediction model, the photovoltaic and wind power forecast values ​​are received. There are N rolling optimizations in the prediction period. The state space expression of the upper-layer MPC model of the microgrid is:

[0038]

[0039] Among them, the MPC prediction look-ahead period is Ts, the prediction time interval is Δt, and there are N sampling points in the look-ahead period. According to the microgrid power balance equation, considering the SOC change of HESS, the MPC state space model is established. Among them, the grid-connected power P at the current moment k is selected. G (k), HESS total power P C (k), the SOC value Sx(k) of HESS is the state variable x(k), and the HESS power change ΔP C (k) is the control variable u(k), and the ultra-short-term power change of photovoltaic and wind power output ΔP PV (k), ΔP WT (k) and the ultra-short-term change in load demand ΔP L (k) is the disturbance variable d(k), E BC 、E SC are the capacities of the battery and supercapacitor respectively, S BC (k), S SC (k) are the SOC values ​​of the battery and supercapacitor at the current sampling moment.

[0040] Grid-connected power P G (k) is the power exchanged between the microgrid and the grid. According to the power balance equation, the expression is:

[0041] P PV +P WT -P L +P C =P G (2)

[0042] The objective function of the upper-level MPC model of the microgrid includes minimizing the grid-connected smoothing suppression and minimizing the deviation between the SOC of the hybrid energy storage and the target signal.

[0043] Objective function 1: Minimize the change of microgrid grid-connected power. The change of microgrid output power fluctuation is stable within a certain range. The optimization objectives considered by the upper control layer include minimizing the change of microgrid grid-connected power. G Expressed as:

[0044]

[0045] Among them, P PVmax 、P WTmax are the day-ahead predicted maximum values ​​of photovoltaic and wind power generation, respectively. Det is the lower limit of power fluctuation allowed by the large power grid.

[0046] Objective function 2: The deviation between the SOC of HESS and the target value is minimized.

[0047]

[0048] Where S0 is the nominal SOC value of HESS.

[0049] The constraints of the upper-layer MPC model of the microgrid include hybrid energy storage state of charge constraints and power constraints.

[0050] The hybrid energy storage state of charge constraint expression is:

[0051]

[0052] Among them, S BC (i) S sC (i) are the charge states of the battery and supercapacitor at the i-th moment; S BCmax 、S BCmin S is the upper and lower limit of the battery state of charge; sCmax 、S SCmin are the upper and lower limits of the supercapacitor voltage.

[0053] The hybrid energy storage power constraint expression is:

[0054]

[0055] Among them, P BC (i) P SC (i) are the charging and discharging power of the battery and supercapacitor at the i-th moment; P BCmax 、P BCmin P SCmax 、P SCmin They are the upper and lower limits of the charging and discharging power of batteries and supercapacitors respectively.

[0056] S3: Estimate the initial state of charge of the hybrid energy storage under day-ahead scheduling, and use the obtained initial state value in the lower-level control S4.

[0057] The initial value formula of hybrid energy storage charge state is:

[0058]

[0059] Among them, SOC k is the state of charge of HESS at the kth moment, E C，0 is the capacity value of HESS at the initial moment.

[0060] S4: In the lower control layer, dynamic programming algorithm is used to optimize the charge and discharge power control of the hybrid energy storage, and the charge state of the hybrid energy storage is S C (k) is output to the upper control layer in S2 as a state variable.

[0061] The dynamic planning and scheduling method for the hybrid energy storage charging and discharging state in the lower-level control is:

[0062] According to the HESS initial state of charge SOC0 calculated in S3, the sampling interval is 1 minute, the core charge state change at the sampling moment is ΔS, and the charge state transfer equation from state 1 at time k to time k+1 is:

[0063] f(k+1)=min{|f(k,l)+P eq (k)+P l |} (8)

[0064] Among them, P eq (k) is the net load at the kth moment, P l The charging / discharging power in the first stage, KW.

[0065] S5: Solve the upper-layer MPC model of the microgrid based on the objective function and constraint limit conditions.

[0066] In S5, different objective functions are combined into a single objective function using penalty coefficients for solution. The objective function is:

[0067]

[0068] Among them, α and β are penalty coefficients, which are used to adjust the weights of different indicators.

[0069] Take a wind-solar hybrid grid-connected power generation system as an example. Its installed PV capacity is 1 MW, its wind turbine capacity is 2.5 MW, its load is 850 kW, and its battery capacity is 2000 kWh, with a rated power of 500 kW and a rated capacity of 80%. The supercapacitor capacity is 15 kWh, with a rated power of 500 kW and a rated capacity of 98%. The upper and lower limits of the HESS remaining capacity for both energy storage components are 0.1E to 0.9E kWh, with a depth of discharge of 0.75. In the two-tier control strategy, the upper-tier MPC rolling planning time is 5 minutes, and the lower-tier real-time control interval is 1 minute. The grid power fluctuation Det is 0.2 MW, the forward prediction period Ts is 3 hours, and the penalty coefficient is 0.5.

[0070] Taking the wind-solar hybrid microgrid system as the simulation object, the proposed hybrid energy storage double-layer control strategy is applied to dispatch the microgrid grid-connected power. The microgrid grid-connected power fluctuation and HESS power change curve are shown in Figure 2. Figures 3 to 5 shown.

[0071] from Figure 3 As can be seen, grid-connected power fluctuations exceeded ±10% of installed capacity only twice during the 24-hour daily dispatch period, demonstrating that the microgrid's upper-layer MPC model achieved the desired grid-connected power fluctuation suppression effect. A comprehensive HESS power output curve shows that photovoltaic and wind power output increased from 12:00 to 16:00. The MPC controller instructed the HESS to begin pre-discharging to supply the load at 10:00, then continue charging to absorb renewable energy and reduce curtailment losses. After 16:00, the supercapacitor and battery began coordinating discharge to support grid-connected power. This continuous discharge achieved optimal power suppression and minimized power fluctuations.

[0072] from Figure 4 and Figure 5 The power curves for batteries and supercapacitors show that batteries meet the low-frequency components of the HESS power, while supercapacitors address the high-frequency output requirements. This leverages the advantages of both power and energy storage devices, reducing battery charge and discharge losses and extending their service life. The introduction of the MPC control strategy allows the HESS to pre-charge and discharge to address future fluctuations in photovoltaic and wind power, while also optimizing battery output smoothness to extend its service life.

[0073] The above embodiments are intended only to illustrate the technical concepts and features of the present invention. Their purpose is to enable those skilled in the art to understand the contents of the present invention and implement them accordingly. They are not intended to limit the scope of protection of the present invention. Any equivalent changes or modifications made in accordance with the spirit of the present invention are intended to be covered by the scope of protection of the present invention.

Claims

1. A two-tier dispatching control method for a microgrid considering the state of charge of a hybrid energy storage, characterized in that: The steps include: S1: Obtain historical data on photovoltaic and wind power generation, load power demand, hybrid energy storage power limit, and hybrid energy storage state of charge limit; S2: In the upper control, the grey GM (1, N) and BP neural network combined prediction method is used to establish the upper prediction control model of the microgrid, that is, the upper MPC model of the microgrid. The total power P of the hybrid energy storage system HESS in the upper MPC model of the microgrid is C (k) used to guide the hybrid energy storage charging and discharging power control in the lower-level control; The combined prediction method of grey GM (1, N) and BP neural network is used to obtain the predicted values ​​of photovoltaic and wind power within the forward prediction period. The prediction period T S In the prediction model, the photovoltaic and wind power forecast values ​​are received. There are N rolling optimizations in the prediction period. The state space expression of the upper-layer MPC model of the microgrid is: Among them, the forward prediction period of the microgrid upper MPC model is T S , the prediction time interval is Δt, there are N sampling points in the forward prediction period, according to the microgrid power balance equation, considering the SOC change of HESS, the MPC state space model is established, in which the grid-connected power P at the current moment k is selected G (k), HESS total power P C (k) SOC value of HESS C (k) is the state variable x(k), HESS power change ΔP C (k) is the control variable u(k), and the ultra-short-term power change of photovoltaic and wind power output ΔP PV (k), ΔP WT (k) and the ultra-short-term change in load demand ΔP L (k) is the disturbance variable d(k), E BC 、E SC are the capacities of the battery and supercapacitor respectively, S BC (k), S SC (k) are the SOC values ​​of the battery and supercapacitor at the current sampling moment; S3: estimating the initial state of charge (SOC0) of the hybrid energy storage under day-ahead scheduling, and using the initial state of charge (SOC0) of the hybrid energy storage in the lower-level control of step S4; The initial value formula of hybrid energy storage charge state is: Among them, SOC k is the state of charge of HESS at the kth moment, E C,0 is the capacity value of HESS at the initial moment; S4: In the lower control layer, dynamic programming algorithm is used to optimize the charge and discharge power control of the hybrid energy storage, and the charge state of the hybrid energy storage is S C (k) Output to the upper control in step S2 as a state variable; S5: Solve the upper-level MPC model of the microgrid based on the objective function and constraint limit conditions.

2. The microgrid dual-layer dispatching control method considering the hybrid energy storage charge state according to claim 1 is characterized in that: The dynamic planning and scheduling method for the hybrid energy storage charging and discharging state in the lower-level control of S4 is: According to the HESS initial state of charge SOC0 calculated in S3, the sampling interval is 1 minute, the core charge state change at the sampling moment is ΔS, and the charge state transfer equation from state 1 at time k to time k+1 is: f(k+1)=min{|f(k,l)+P eq (k)+P l |} Among them, P eq (k) is the net load at the kth moment, P l is the charge / discharge power in the first stage.

3. The microgrid dual-layer dispatching control method considering the hybrid energy storage charge state according to claim 1 is characterized in that: In S5, different objective functions are combined into a single objective function using penalty coefficients for solution. The objective function is: Among them, α and β are penalty coefficients, which are used to adjust the weights of different indicators; ΔP G Indicates the grid-connected fluctuation exceeding the limit value, k represents the current moment, ΔS C is the deviation between the SOC of HESS and the target value; N is the number of sampling points in the look-ahead period, N = Ts / Δt.

Images