A Hybrid Energy Storage Dispatch Method and System for Microgrids

Abstract

This invention belongs to the field of microgrids, specifically relating to a hybrid energy storage scheduling method for microgrids, comprising the following steps: acquiring historical load curves, renewable energy forecast data, and energy storage unit parameters; using historical load curves and renewable energy forecast data to determine the charging and discharging strategy of the energy storage unit; initially setting the SOC safety range, continuously optimizing multi-objective weighting coefficients, and obtaining the filtering frequency band threshold; periodically performing MPC scheduling optimization, issuing power commands in real time, forming a closed-loop control of the SOC safety range, multi-objective weighting coefficients, and filtering frequency band thresholds; under the premise of hybrid energy storage in the microgrid, extending the service life of the energy storage unit and maintaining the optimal charging and discharging frequency during the scheduling process, continuously outputting the optimal power command sequence to assist scheduling through rolling optimization; and, with the assistance of SOH estimation, scheduling control of each sub-device of the energy storage unit is performed, effectively solving the problem of overuse of a single energy storage sub-device.

Description

Technical Field

[0001] This invention relates to the field of microgrids, and more specifically to a hybrid energy storage dispatching method and system for microgrids. Background Technology

[0002] A microgrid is a small-scale power system composed of distributed power sources, loads, energy storage, power distribution, and control systems. It is an autonomous power system capable of self-control, protection, and management, possessing complete power generation, distribution, and consumption functions, and effectively optimizing energy within the grid. Microgrids are a concept relative to traditional large power grids. Traditional large power grids primarily operate on a unidirectional basis, with electricity traveling from power plants through transmission lines to users. Microgrids, however, operate on a bidirectional basis, enabling local power generation and balancing, maximizing energy efficiency. With the global energy crisis, increased energy consumption, and the rise of new energy technologies, new energy power generation is becoming increasingly widespread, gradually forming a new energy and electricity market. However, the energy density of new energy sources is generally low, and high-power generation requires careful site selection, thus they are intermittent power sources. The emergence of microgrid technology provides an important technological direction for the efficient utilization of these new energy sources.

[0003] Hybrid energy storage systems in microgrids typically include different types of energy storage technologies, such as batteries, supercapacitors, and flywheels. Each type of energy storage unit has different characteristics. For example, batteries have high energy density but limited charge-discharge cycles, while supercapacitors have high power density and long cycle life. Therefore, scheduling methods need to consider how to rationally allocate loads so that each type of energy storage unit can play its advantages and avoid overuse of a particular type, especially batteries that are prone to aging. Summary of the Invention

[0004] (I) Purpose of the Invention

[0005] To address the technical problems existing in the background art, this invention proposes a microgrid hybrid energy storage scheduling method, which features extended service life of energy storage units and maintenance of optimal charging and discharging frequency.

[0006] (II) Technical Solution

[0007] To address the aforementioned technical problems, this invention provides a microgrid hybrid energy storage dispatch method, comprising the following steps:

[0008] Obtain historical load curves, renewable energy forecast data, and energy storage unit parameters. Use the historical load curves and renewable energy forecast data to determine the charging and discharging strategies for the energy storage units.

[0009] Initially set the SOC safe range, continuously optimize the multi-objective weight coefficients, and obtain the filter frequency band threshold;

[0010] Periodically perform MPC scheduling optimization, issue power commands in real time, and form a closed-loop control of SOC safety range, multi-objective weight coefficients and filter frequency band thresholds;

[0011] By estimating the state of energy loss (SOH) online using a Kalman filter, the aging degree of energy storage units can be tracked in real time, thereby optimizing energy storage scheduling strategies.

[0012] Preferably, the historical load curve involves electricity demand, renewable energy forecast data involves power generation, and energy storage parameters involve capacity and lifetime. The net load calculation formula based on the interaction of these three factors is: P net (t)=P load (t)-P re (t);

[0013] Among them, P net (t): Net load at time t (unit: kW);

[0014] P load (t): Historical load curve power at time t (unit: kW);

[0015] P re (t): Predicted renewable energy power at time t (unit: kW);

[0016] Net load represents the actual power that needs to be supplemented by energy storage units or the power grid;

[0017] When P net When (t)>0, the energy storage unit needs to discharge;

[0018] When P net When (t) < 0, the energy storage unit can be charged.

[0019] Preferably, the energy storage charging and discharging power constraint formula is:

[0020]

[0021] P storage (t): Actual charging and discharging power of energy storage at time t (unit: kW);

[0022] P rated Rated power of the energy storage unit (unit: kW, from energy storage parameters);

[0023] Energy storage capacity is limited by the rated value to avoid overcharging or over-discharging. Rated power P rated These are key parameters for energy storage units.

[0024] Preferably, the initial range of the SOC safe interval is (20% to 80%).

[0025] Optimize weighting coefficients for multi-objective optimization problems, including factors such as grid cost, energy storage unit attenuation cost, and tracking error cost;

[0026] Filtering frequency thresholds are used for signal processing to filter out high-frequency noise or low-frequency interference.

[0027] Preferably, MPC scheduling optimization is performed periodically, with each cycle lasting at least 15 minutes. Power commands for future periods are generated based on the latest data and prediction models. The MPC objective function is:

[0028]

[0029] N: Length of the predicted time domain (e.g., N=4 corresponds to 1 hour);

[0030] C grid (k): The grid electricity price (yuan / kWh) for the kth time period;

[0031] w1, w2, w3: Weighting coefficients (must satisfy w1+w2+w3=1);

[0032] β: Battery degradation coefficient (related to SOHR);

[0033] P ref (k): Reference power command (such as grid dispatch requirements).

[0034] The preferred SOC dynamic model is:

[0035]

[0036] SOC(k): The state of charge of the battery in the k-th time period (range: [0,1]);

[0037] P(k): The charging and discharging power in the kth time period (P>0 for discharging, P<0 for charging);

[0038] η: Charge / discharge efficiency (ηdis discharge, ηch charge);

[0039] Δt: Control period (15 minutes);

[0040] C nom Battery nominal capacity (kWh);

[0041] SOH C :Health status (capacity decay rate, range: [0,1]).

[0042] Preferably, the constraint formula for the SOC safe interval is: SOC min ≤SOC(k)≤SOC max ,

[0043] SOC min SOC max Dynamic safety interval [0.2, 0.8];

[0044] Through optimization calculation using the MPC objective function, the SOC is always in the range of [0.2, 0.8], the power command does not exceed the rated value of the converter, and the optimal power command sequence for the next 15 minutes is output.

[0045] Preferably, the Kalman filter is a recursive algorithm suitable for state estimation in dynamic systems. It processes noise and updates the estimated value in real time. For energy storage units, the state of energy (SOH) is usually related to capacity decay or internal resistance increase. Changes in capacity and internal resistance can be used as state variables, while real-time data such as voltage, current and temperature are used as observation inputs.

[0046] Preferably, the SOH estimation is integrated into the MPC scheduling, and their charging and discharging strategies are adjusted according to the SOH of each energy storage unit, which includes batteries, supercapacitors and flywheels.

[0047] The present invention also provides a microgrid hybrid energy storage dispatch system, comprising:

[0048] Data acquisition module: acquires historical load curves, renewable energy forecast data, and energy storage unit parameters; and uses historical load curves and renewable energy forecast data to determine the charging and discharging strategy of the energy storage unit.

[0049] Initialization configuration module: Initializes the SOC safe range, continuously optimizes multi-objective weight coefficients, and obtains filter frequency band thresholds;

[0050] Rolling optimization module: Periodically performs MPC scheduling optimization, issues power commands in real time, and forms a closed-loop control of SOC safety range, multi-objective weight coefficients and filter frequency band thresholds;

[0051] Health monitoring module: It estimates the state of energy (SOH) online through a Kalman filter and tracks the aging of energy storage units in real time, thereby optimizing energy storage scheduling strategies.

[0052] The above-mentioned technical solution of the present invention has the following beneficial technical effects: under the premise of microgrid hybrid energy storage, the service life of energy storage units is extended and the optimal charging and discharging frequency is maintained during the scheduling process. The optimal power command sequence is continuously output to assist the scheduling through rolling optimization. With the assistance of SOH estimation, the scheduling and control of each sub-device of the energy storage unit is carried out, which effectively solves the problem of overuse of a single energy storage sub-device. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of the method of the present invention;

[0054] Figure 2 This is a schematic diagram of the process of the present invention;

[0055] Figure 3 This is a schematic diagram of the rolling optimization flowchart structure of the present invention. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. It should be understood that these descriptions are merely exemplary and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0057] like Figure 1-3 As shown, the present invention proposes a microgrid hybrid energy storage dispatch method, which includes the following steps:

[0058] Step S100: Obtain historical load curves, renewable energy forecast data, and energy storage unit parameters; use historical load curves and renewable energy forecast data to determine the charging and discharging strategy of the energy storage unit.

[0059] Step S200: Initially set the SOC safe range, continuously optimize the multi-objective weight coefficients, and obtain the filter frequency band threshold;

[0060] Step S300: Periodically perform MPC scheduling optimization, issue power commands in real time, and form a closed-loop control of SOC safe range, multi-objective weight coefficient and filter frequency band threshold;

[0061] Step S400: Estimate SOH online using a Kalman filter to track the aging of energy storage units in real time, thereby optimizing the energy storage scheduling strategy.

[0062] Historical load curves involve electricity demand, renewable energy forecast data involves power generation, and energy storage parameters involve capacity and lifetime. The net load calculation formula, which relates these three factors, is as follows: P net (t)=P load (t)-P re (t);

[0063] Among them, P net (t): Net load at time t (unit: kW);

[0064] P load (t): Historical load curve power at time t (unit: kW);

[0065] P re (t): Predicted renewable energy power at time t (unit: kW);

[0066] Net load represents the actual power that needs to be supplemented by energy storage units or the power grid;

[0067] When P net When (t)>0, the energy storage unit needs to discharge;

[0068] When P net When (t) < 0, the energy storage unit can be charged.

[0069] The energy storage charging and discharging power constraint formula is:

[0070]

[0071] P storage (t): Actual charging and discharging power of energy storage at time t (unit: kW);

[0072] P rated Rated power of the energy storage unit (unit: kW, from energy storage parameters);

[0073] Energy storage capacity is limited by the rated value to avoid overcharging or over-discharging. Rated power P rated These are key parameters for energy storage units.

[0074] The purpose of initially setting the SOC safety range, optimizing the weighting coefficients, and the filtering frequency band thresholds is to improve system performance, security, and reliability.

[0075] The initial range of the SOC safety zone is (20% to 80%). This is to prevent the energy storage unit from being overcharged or over-discharged, and to avoid the energy storage unit from operating in an excessively high or low charge state, so as to prevent damage to the energy storage unit or the occurrence of safety hazards.

[0076] On the other hand, it can slow down the aging rate of energy storage units and extend their lifespan.

[0077] Optimizing weighting coefficients is used for multi-objective optimization problems. For example, in the case of factors such as grid cost, energy storage unit attenuation cost, and tracking error cost, it is necessary to adjust the weighting coefficients of each objective to achieve the best overall performance. It can be understood that the sum of the weighting coefficients is one.

[0078] Finally, the filter band threshold is used for signal processing to filter out high-frequency noise or low-frequency interference, ensuring signal accuracy, retaining useful signals and removing noise in data processing.

[0079] For example, in energy storage systems:

[0080] SOC range limits the depth of charge and discharge to protect battery life;

[0081] When optimizing the weighting coefficients, priority is given to smoothing grid power fluctuations (high-frequency filtering), while also considering the SOC change rate (weighting allocation).

[0082] The filtering threshold removes measurement noise to ensure the accuracy of SOC estimation.

[0083] By setting these parameters appropriately, the system can achieve an optimal balance between safety, efficiency, and stability.

[0084] Periodically perform MPC scheduling optimization and issue power commands in real time to form a closed-loop control of SOC safety range, multi-objective weight coefficients and filter frequency band thresholds. In scenarios where MPC (Model Predictive Control) scheduling optimization is performed every 15 minutes and power commands are issued in real time, the design of SOC safety range, optimized weight coefficients and filter frequency band thresholds can significantly improve the system's control accuracy, safety and economy.

[0085] Control cycle: MPC optimization is triggered every 15 minutes to generate power commands for future periods based on the latest data and prediction models.

[0086] Real-time delivery: Optimized power commands (such as charging and discharging power, grid interaction power, etc.) are delivered to the execution equipment (such as energy storage units) in real time.

[0087] Closed-loop feedback: The system status (such as SOC and power) is collected in real time by sensors, filtered, and then input into the MPC model to form closed-loop control.

[0088] The SOC dynamic model is:

[0089]

[0090] SOC(k): The state of charge of the battery in the k-th time period (range: [0,1]);

[0091] P(k): The charging and discharging power in the kth time period (P>0 for discharging, P<0 for charging);

[0092] η: Charge / discharge efficiency (ηdis discharge, ηch charge);

[0093] Δt: Control period (15 minutes);

[0094] C nom Battery nominal capacity (kWh);

[0095] SOH C :Health status (capacity decay rate, range: [0,1]).

[0096] The objective function of MPC is:

[0097]

[0098] N: Length of the predicted time domain (e.g., N=4 corresponds to 1 hour);

[0099] C grid (k): The grid electricity price (yuan / kWh) for the kth time period;

[0100] w1, w2, w3: Weighting coefficients (must satisfy w1+w2+w3=1);

[0101] β: Battery degradation coefficient (related to SOHR);

[0102] P ref (k): Reference power command (such as grid dispatch requirements).

[0103] The constraint formula for the SOC safe interval is: SOC min ≤SOC(k)≤SOC max ,

[0104] SOC min SOC max Dynamic safety interval (e.g., [0.2, 0.8]);

[0105] Actual value based on SOH C Adjustments (such as SOH) C When SOC is <0.9, max =0.75);

[0106] It should be noted that: through the optimization calculation of the MPC objective function, the SOC is always in the range of [0.2, 0.8], the power command does not exceed the rated value of the converter, and the optimal power command sequence for the next 15 minutes is output.

[0107] The first power command (i.e. the command at the current moment) is sent to the execution device in real time, and the remaining command sequence is used as a reference to dynamically correct the feedback in the next cycle, thus realizing closed-loop feedback.

[0108] Understandably, the actual SOC change can be monitored during the optimization calculation of the MPC objective function. If the deviation from the predicted value exceeds a threshold (e.g., ±5%), the MPC objective function can be re-optimized to ensure that the SOC is always within the dynamic safety range.

[0109] In one embodiment, peak-valley arbitrage is achieved through 15-minute MPC scheduling, while extending the lifespan of the energy storage unit.

[0110] The parameters are set as follows:

[0111] SOC range: 25% to 75% (25% capacity is reserved to account for prediction errors);

[0112] Weighting coefficients: grid cost: energy storage unit attenuation cost: tracking error is 0.7:0.2:0.1;

[0113] Filtering threshold: The power command is low-pass filtered at 10Hz to prevent frequent switching of the converter;

[0114] Inputting the parameters into the MPC objective function yields the following results: daily revenue increased by 12%, and energy storage unit cycle life was extended by 15%.

[0115] In an optional embodiment, the priorities of different optimization objectives in the MPC objective function are defined to achieve dynamic trade-offs among multiple objectives.

[0116] The weighting factors include: economic efficiency (such as maximizing electricity price difference revenue), energy storage unit lifespan (such as reducing charge-discharge cycle depth), and grid demand (such as smoothing load fluctuations);

[0117] In the objective function, weights are assigned to different objectives (e.g., economic efficiency weight 0.6, energy storage unit lifetime weight 0.3, grid demand weight 0.1).

[0118] Dynamically adjust weights: Update weights in real time based on external conditions (such as peak electricity prices, emergency frequency regulation needs).

[0119] As another example, during periods of low electricity market prices (such as early morning), economic factors are given priority for charging; when frequency regulation demand is urgent, grid factors are given priority for responding to power commands.

[0120]

[0121] The SOH (State of Health) of energy storage units is estimated online using a Kalman filter, and the aging degree of the energy storage units is tracked in real time, thereby optimizing energy storage management strategies (such as dynamically adjusting the SOC safe range and limiting charge and discharge power). The following is the specific implementation logic and its collaborative application with MPC control:

[0122] Among them, the Kalman filter is a recursive algorithm suitable for state estimation in dynamic systems. It can handle noise and update the estimated value in real time. For energy storage units, the state of energy (SOH) is usually related to capacity decay or internal resistance increase. Changes in capacity and internal resistance can be used as state variables, while real-time data such as voltage, current and temperature are used as observation inputs.

[0123] SOH characterization indicators:

[0124] Capacity degradation rate: SOH_Capacity = Current maximum capacity / Initial rated capacity × 100% (If below 80%, it is considered the end of the life).

[0125] Internal resistance growth rate: SOH_Resistance = current internal resistance / initial internal resistance × 100% (if it exceeds 150%, it is considered a failure).

[0126] By estimating the State of Harm (SOH) index, charging and discharging strategies can be dynamically adjusted to avoid accelerated aging due to overstress; the remaining battery life can be predicted, and maintenance plans can be optimized.

[0127] Energy storage unit aging models need to describe the change of SOH over time, and empirical models are usually used:

[0128] Capacity decay model (empirical formula): SOH_Capacity(k+1)=SOH_Capacity(k)-α·I·Δt / Q_initial

[0129] Where α is the aging coefficient (which needs to be calibrated online);

[0130] I represents the charging and discharging current;

[0131] Q_initial is the initial capacity.

[0132] The SOH estimation results are dynamically fed back to the MPC scheduling to achieve aging-adaptive energy storage unit management.

[0133] In one embodiment, SOH estimation is integrated into MPC scheduling. The charging and discharging strategies of each energy storage unit are adjusted based on the SOH of each unit. The energy storage units include batteries, supercapacitors, and flywheels. In grid frequency regulation, supercapacitors handle high-frequency fluctuations, batteries handle medium- and long-term demands, and flywheels provide instantaneous support. When it is necessary to limit the charging and discharging power of aging batteries, supercapacitors are used preferentially for high-frequency operations. The response threshold of the flywheel is adjusted according to its wear condition. Real-time SOH estimation can dynamically adjust the task allocation of each device and extend the overall system life.

[0134] The core formula for energy storage scheduling is:

[0135]

[0136] P sc Supercapacitor power distribution (kW)

[0137] P bat t: Battery power distribution (kW)

[0138] P grid Power exchange between the grid and the grid (kW)

[0139] Δt: Duration of power fluctuation

[0140] Application scenario: Combining real-time estimation of SOH and real-time power allocation decision, the charging and discharging tasks of energy storage units are allocated, and when ifΔt < 10s, supercapacitors are used for instantaneous discharge;

[0141] When if10s≤Δt≤1h, use the battery for continuous discharge.

[0142] The present invention also provides a microgrid hybrid energy storage dispatch system, comprising:

[0143] Data acquisition module: acquires historical load curves, renewable energy forecast data, and energy storage unit parameters; and uses historical load curves and renewable energy forecast data to determine the charging and discharging strategy of the energy storage unit.

[0144] Initialization configuration module: Initializes the SOC safe range, continuously optimizes multi-objective weight coefficients, and obtains filter frequency band thresholds;

[0145] Rolling optimization module: Periodically performs MPC scheduling optimization, issues power commands in real time, and forms a closed-loop control of SOC safety range, multi-objective weight coefficients and filter frequency band thresholds;

[0146] Health monitoring module: It estimates the state of energy (SOH) online through a Kalman filter and tracks the aging of energy storage units in real time, thereby optimizing energy storage scheduling strategies.

[0147] It should be understood that the specific embodiments described above are merely illustrative or explanatory of the principles of the invention and do not constitute a limitation thereof. Therefore, any modifications, equivalent substitutions, improvements, etc., made without departing from the spirit and scope of the invention should be included within the protection scope of the invention. Furthermore, the appended claims are intended to cover all variations and modifications falling within the scope and boundaries of the appended claims, or equivalent forms of such scope and boundaries.

Claims

1. A microgrid hybrid energy storage dispatch method, characterized in that, Includes the following steps: Obtain historical load curves, renewable energy forecast data, and energy storage unit parameters. Use the historical load curves and renewable energy forecast data to determine the charging and discharging strategies for the energy storage units. Initially set the SOC safe range, continuously optimize the multi-objective weight coefficients, and obtain the filter frequency band threshold; Periodically perform MPC scheduling optimization, issue power commands in real time, and form a closed-loop control of SOC safety range, multi-objective weight coefficients and filter frequency band thresholds; The SOH is estimated online by using a Kalman filter to track the aging of energy storage units in real time, thereby optimizing the energy storage scheduling strategy. MPC scheduling optimization is performed periodically, with each cycle lasting at least 15 minutes. Power commands for future time periods are generated based on the latest data and predictive models. The MPC objective function is: ； N: Length of the prediction time domain; Electricity price in the k-th time period, unit: yuan / kWh; w1,w2,w 3： The weighting coefficients satisfy w1+w2+w3=1; Battery degradation coefficient; Reference power command; The SOC dynamic model is: ； The state of charge of the battery in the k-th time period. Range: [0,1]; The charging and discharging power in the k-th time period. Discharge, Charge; Charge and discharge efficiency; Control cycle, It lasts for 15 minutes; Battery nominal capacity, unit: kWh; : Healthy status, capacity decay rate range: [0,1].

2. The microgrid hybrid energy storage dispatch method according to claim 1, characterized in that, Historical load curves involve electricity demand, renewable energy forecast data involves power generation, and energy storage parameters involve capacity and lifetime. The net load calculation formula is based on the interaction of these three factors: ； in, Net load at time t, unit: kW; Historical load curve power at time t, unit: kW; : Forecasted renewable energy power at time t, in kW; Net load represents the actual power that needs to be supplemented by energy storage units or the power grid; when At this time, the energy storage unit needs to discharge; when The energy storage unit can be charged at this time.

3. The microgrid hybrid energy storage dispatch method according to claim 2, characterized in that, The energy storage charging and discharging power constraint formula is: ； Actual charge / discharge power of energy storage at time t, unit: kW; Rated power of the energy storage unit, unit: kW; Energy storage capacity is limited by the rated value to avoid overcharging or over-discharging. Rated power These are key parameters for energy storage units.

4. The microgrid hybrid energy storage dispatch method according to claim 1, characterized in that, The initial range of the SOC safe zone is 20% to 80%; Optimize weighting coefficients for multi-objective optimization problems, including grid cost, energy storage unit attenuation cost, and tracking error cost factors; Filtering frequency thresholds are used for signal processing to filter out high-frequency noise or low-frequency interference.

5. The microgrid hybrid energy storage dispatch method according to claim 1, characterized in that, The constraint formula for the SOC safe zone is: ； , Dynamic safety interval [0.2, 0.8]; Through optimization calculation using the MPC objective function, the SOC is always in the range of [0.2, 0.8], the power command does not exceed the rated value of the converter, and the optimal power command sequence for the next 15 minutes is output.

6. The microgrid hybrid energy storage dispatch method according to claim 1, characterized in that, The Kalman filter is a recursive algorithm suitable for state estimation in dynamic systems. It processes noise and updates the estimated value in real time. For energy storage units, the state of energy (SOH) is usually related to capacity decay or internal resistance increase. Changes in capacity and internal resistance are used as state variables, while real-time data of voltage, current and temperature are used as observation inputs.

7. A microgrid hybrid energy storage dispatch method according to claim 1, characterized in that, The State of Health (SOH) estimation is integrated into the MPC scheduling, and the charging and discharging strategies of each energy storage unit are adjusted according to the SOH of each unit, which includes one of the following: battery, supercapacitor, and flywheel.

8. A microgrid hybrid energy storage dispatch system, employing a microgrid hybrid energy storage dispatch method as described in any one of claims 1 to 7, characterized in that, include: Data acquisition module: acquires historical load curves, renewable energy forecast data, and energy storage unit parameters; and uses historical load curves and renewable energy forecast data to determine the charging and discharging strategy of the energy storage unit. Initialization configuration module: Initializes the SOC safe range, continuously optimizes multi-objective weight coefficients, and obtains filter frequency band thresholds; Rolling optimization module: Periodically performs MPC scheduling optimization, issues power commands in real time, and forms a closed-loop control of SOC safety range, multi-objective weight coefficients and filter frequency band thresholds; Health monitoring module: It estimates the state of energy (SOH) online through a Kalman filter and tracks the aging of energy storage units in real time, thereby optimizing energy storage scheduling strategies.

Images