A method and system for coordinated operation control of a multi-energy complementary energy storage system

Abstract

This invention discloses a collaborative operation control method and system for a multi-energy complementary energy storage system, relating to the field of energy storage system optimization and scheduling technology. The method includes: constructing a nonlinear coupled model of the multi-energy complementary energy storage system; acquiring new energy and load forecast data and calculating a new energy absorption capacity index to mitigate grid-connected power fluctuations and reduce system curtailment rate; configuring corresponding parameters of a hierarchical model predictive control architecture based on the index; and generating collaborative charging and discharging control commands for multiple types of energy storage through hierarchical rolling time-domain solution. This invention solves the technical problems of existing multi-energy storage regulation methods lacking a unified quantitative evaluation basis, having a single scheduling level without continuous self-correction capability, and being difficult to adapt to system operation deviations caused by random fluctuations in new energy and load. It achieves the technical effects of dynamic collaborative adjustment of multiple energy storage devices, stable grid-connected output, and reduced new energy waste.

Description

Technical Field

[0001] This invention relates to the field of energy storage system optimization and scheduling technology, and in particular to a method and system for coordinated operation control of multi-energy complementary energy storage systems. Background Technology

[0002] Multi-energy complementary energy storage systems are the core carriers for high-proportion renewable energy consumption, and their coordinated control effect directly determines grid connection stability and curtailment management level. Existing technologies mostly employ simplified models of single energy storage systems or single-layer scheduling architectures for regulation. Conventional linear modeling struggles to characterize the nonlinear coupling characteristics of the conversion losses between electricity, heat, and hydrogen, and scheduling strategies lack quantitative indicators to reflect the system's absorption capacity. Without quantifying renewable energy absorption capacity as a basis for scheduling, single-layer optimization struggles to balance global planning and real-time adjustment, lacks hierarchical rolling solutions and closed-loop correction mechanisms, and cannot accurately output coordinated charging and discharging commands for multiple energy storage systems under random source-load fluctuations, making it difficult to effectively suppress grid-connected power fluctuations and curtailment issues. Summary of the Invention

[0003] This application provides a collaborative operation control method and system for a multi-energy complementary energy storage system, which solves the technical problems of existing multi-energy storage regulation methods lacking a unified quantitative evaluation basis, having a single scheduling level and no continuous self-correction capability, and being difficult to adapt to system operation deviations caused by random fluctuations in new energy sources and loads.

[0004] The first aspect of this application provides a method for coordinated operation control of a multi-energy complementary energy storage system. The method includes: constructing a nonlinear coupled model of the multi-energy complementary energy storage system to characterize the dynamic nonlinear characteristics of the multi-energy flow conversion and storage processes; acquiring new energy power generation forecast data and load forecast data, and calculating a new energy absorption capacity index; with the joint optimization objective of smoothing grid-connected power fluctuations and reducing system curtailment rate, configuring corresponding parameters of a hierarchical model predictive control architecture based on the new energy absorption capacity index, and constructing a hierarchical model predictive control architecture; and performing a rolling time-domain solution on the nonlinear coupled model through the hierarchical model predictive control architecture to generate charging and discharging coordinated control commands for various types of energy storage.

[0005] A second aspect of this application provides a collaborative operation control system for a multi-energy complementary energy storage system. The system includes: a coupling model construction module for constructing a nonlinear coupling model of the multi-energy complementary energy storage system to characterize the dynamic nonlinear characteristics of the multi-energy flow conversion and storage processes; a prediction data acquisition module for acquiring new energy power generation prediction data and load prediction data, and calculating a new energy absorption capacity index; a control architecture construction module for configuring corresponding parameters of a hierarchical model predictive control architecture based on the new energy absorption capacity index, with the joint optimization objectives of smoothing grid-connected power fluctuations and reducing system curtailment rate; and a control command generation module for performing rolling time-domain solutions on the nonlinear coupling model through the hierarchical model predictive control architecture to generate charging and discharging collaborative control commands for various types of energy storage.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application constructs a multi-energy coupled simulation model of electricity, heat, and hydrogen, combines source-load prediction data to quantify the system's renewable energy acceptance capacity, sets multi-level optimization constraints in layers and iteratively solves energy storage control commands, and continuously corrects the model and scheduling strategy based on on-site operation measurement data. This dynamically adapts to wind and solar power output and random load fluctuations, achieves precise control of multiple types of energy storage, effectively stabilizes grid-connected power, improves the local consumption level of renewable energy, and achieves the technical effect of dynamic coordinated adjustment of multiple energy storage devices, stable grid-connected power output, and reduced renewable energy waste. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 This is a flowchart illustrating a collaborative operation control method for a multi-energy complementary energy storage system provided in an embodiment of this application.

[0009] Figure 2 This is a schematic diagram of the structure of a multi-energy complementary energy storage system collaborative operation control system provided in an embodiment of this application.

[0010] Figure labeling: Coupled model construction module 1, prediction data acquisition module 2, control architecture construction module 3, control command generation module 4. Detailed Implementation

[0011] This application provides a collaborative operation control method and system for a multi-energy complementary energy storage system, which solves the technical problems of existing multi-energy storage regulation methods lacking a unified quantitative evaluation basis, having a single scheduling level and no continuous self-correction capability, and being difficult to adapt to system operation deviations caused by random fluctuations in new energy sources and loads.

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0013] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0014] Example 1, as Figure 1 As shown, a method for coordinated operation control of a multi-energy complementary energy storage system is provided, wherein the method includes: A nonlinear coupling model of a multi-energy complementary energy storage system is constructed to characterize the dynamic nonlinear characteristics of the multi-energy flow conversion and storage process.

[0015] In this embodiment, the multi-energy complementary energy storage system is a comprehensive energy system that integrates various energy storage devices such as electrochemical energy storage, thermal energy storage, and hydrogen energy storage, and coordinates the management of multiple energy flows including electrical energy, thermal energy, and hydrogen energy to enhance the absorption capacity of new energy sources. The nonlinear coupling model is deployed within the control and computing unit and describes the energy conversion, loss, and interrelationships of various types of energy storage through mathematical equations, reflecting the nonlinear variation of efficiency with operating conditions.

[0016] Specifically, the construction of the nonlinear coupling model includes: taking electrochemical energy storage, thermal energy storage and hydrogen energy storage as basic energy storage units, and establishing nonlinear mathematical descriptions of each energy storage unit in the energy conversion, storage and loss process.

[0017] Furthermore, by coupling and relating the energy conversion efficiency functions and power balance constraints among the electrical, thermal, and hydrogen multi-energy flows, the nonlinear mathematical descriptions of the aforementioned units are constructed into a multi-energy-dimensional nonlinear coupled model. This model accurately reflects the nonlinear dynamic characteristics of energy-power relationships during the conversion and storage processes of the multi-energy flows, providing a precise system state description basis for subsequent coordinated control.

[0018] By acquiring new energy power generation forecast data and load forecast data, the new energy absorption capacity index is calculated.

[0019] In this embodiment of the application, the new energy absorption capacity index is calculated by the ratio of surplus new energy power to adjustable energy storage capacity, and is used as a numerical indicator to quantitatively evaluate the system's new energy absorption capacity and judge the risk of energy curtailment.

[0020] Optionally, new energy output forecast data and load demand forecast data at multiple time scales can be collected, and outlier removal, missing value completion, and time series alignment can be performed on the original data in sequence to obtain a standardized forecast dataset.

[0021] Based on this forecast dataset, the power surplus between the predicted output of new energy sources and the upper limit of grid-connected transmission power is calculated, as well as the total adjustable capacity that each energy storage unit can provide under the current energy storage conditions. The ratio of the power surplus to the total adjustable capacity is used as the new energy absorption capacity index. This index is used to characterize the system's absorption margin for new energy power generation in the current and future periods, providing a quantitative basis for the parameter configuration of the hierarchical control architecture.

[0022] With the joint optimization goal of smoothing grid-connected power fluctuations and reducing system curtailment rate, a hierarchical model predictive control architecture is constructed by configuring the corresponding parameters of the hierarchical model predictive control architecture based on the new energy absorption capacity index.

[0023] In this embodiment, the hierarchical model predictive control architecture is deployed in the energy control equipment and is divided into three levels: global planning, collaborative scheduling, and real-time control. It is a hierarchical computational control system that generates energy storage control commands by rolling optimization at different cycles.

[0024] In one embodiment of this application, the construction of the hierarchical model predictive control architecture specifically includes: building a three-layer control architecture comprising a strategic layer, a coordination layer, and an execution layer. Based on the renewable energy absorption capacity index, the optimization time domain, constraints, and decision weights for each layer are set to achieve the joint optimization objective of smoothing grid-connected power fluctuations and reducing system curtailment rate.

[0025] Through the above configuration, each level can work together at different time scales to achieve hierarchical and coordinated control over the operation status of the multi-energy complementary energy storage system.

[0026] The nonlinear coupled model is solved in the rolling time domain by the hierarchical model predictive control architecture to generate charging and discharging coordinated control commands for various types of energy storage.

[0027] Specifically, the system updates real-time operating status and forecast data within each control cycle, solves the optimization problem at each level in the order of strategic layer, coordination layer, and execution layer, and passes the optimization result to the next layer as constraint boundary after solving at each level.

[0028] Through rolling iterations, the charging and discharging power and operating sequence of each type of energy storage within the current control cycle are output as charging and discharging coordinated control commands. These commands are used to dynamically adapt to random fluctuations in new energy output and load, enabling coordinated operation control of multiple types of energy storage.

[0029] Furthermore, the method provided in this application embodiment includes: Nonlinear mathematical descriptions of the energy conversion, storage, and loss processes corresponding to electrochemical energy storage, thermal energy storage, and hydrogen energy storage are established respectively. By coupling and associating the nonlinear mathematical descriptions with the energy conversion efficiency function and power balance constraint relationship between the electric, thermal, and hydrogen multi-energy flows, a multi-energy-dimensional nonlinear coupling model is constructed.

[0030] Specifically, firstly, multiple sets of operating data for the electrochemical energy storage device within its rated power range were collected. Ten gradients were evenly selected for the test conditions, ranging from 10%, 20%, to 100% of the rated charge / discharge power. Simultaneously, the corresponding ambient temperature, charge / discharge efficiency, internal resistance voltage drop, and self-discharge rate parameters were recorded. A second-order polynomial fitting was used to obtain a nonlinear function of charge / discharge efficiency as a function of charge / discharge power and ambient temperature. The independent variables of the charge / discharge efficiency function are the current charge / discharge power and ambient temperature, and the dependent variable is the actual energy conversion efficiency under the corresponding operating condition. The coefficient of determination for the fitting was no less than 0.95.

[0031] Discrete state equations are constructed with a step length of 1 minute. The state variable is the current state of charge of the stored energy, and the input variable is the current charging and discharging power. The loss term includes internal resistance Joule loss and self-discharge loss per unit time. The internal resistance Joule loss is calculated as the product of the square of the current charging and discharging current and the nominal internal resistance of the electrochemical energy storage unit. The self-discharge loss is calculated as the product of the rated self-discharge rate per unit time and the current energy storage capacity. The value of the state of charge at the next moment is obtained by superimposing the initial value of the state of charge with the difference between the effective charging and discharging power and the loss. This fully characterizes the dynamic process of energy conversion, storage and loss of electrochemical energy storage.

[0032] Next, multiple sets of operational data were collected from the thermal energy storage device within its rated charge / discharge power range. Ten test gradients were selected, ranging from 10% to 100% of the rated charge / discharge power, and the temperature of the thermal storage medium, ambient temperature, heat exchange efficiency, and heat dissipation parameters were recorded simultaneously. A nonlinear function of heat exchange efficiency as a function of charge / discharge power was obtained using quadratic polynomial fitting, while a linear fitting was used to calculate the relationship between heat dissipation per unit time and the temperature difference between the thermal storage medium and the environment. The coefficients of determination for both fittings were not less than 0.95.

[0033] A thermal equilibrium equation of state is constructed with a step length of 1 minute. The state variable is the heat storage capacity of the thermal storage device at the current moment, the input variable is the charging and discharging power at the current moment, and the loss term is the natural heat dissipation per unit time. The heat storage value at the next moment is obtained by superimposing the initial value of the heat storage capacity with the difference between the effective charging and discharging power and the heat dissipation loss. This represents the dynamic process of thermal energy conversion, storage and heat dissipation loss in thermal energy storage.

[0034] Tests were conducted on the hydrogen energy storage system, which includes an electrolysis hydrogen production unit, a fuel cell power generation unit, and a hydrogen storage unit. For both the electrolysis hydrogen production unit and the fuel cell power generation unit, 10 test operating points were selected at 10% to 100% of the rated power. The input power and output energy were recorded under each operating condition. The nonlinear functions of the electrolysis hydrogen production efficiency and the fuel cell power generation efficiency as a function of the operating power were obtained by fitting a third-order polynomial. The coefficient of determination for the fitting was not less than 0.95.

[0035] The hydrogen storage unit uses leakage rate and compression energy consumption per unit time as loss parameters. The compression energy consumption is calculated as the product of the compression power consumption coefficient corresponding to the unit hydrogen production and the current hydrogen production rate. The state equation for hydrogen storage capacity is constructed with a step size of 1 minute. The state variable is the hydrogen mass storage in the hydrogen storage tank at the current moment. The input variable is the hydrogen production rate by electrolysis, and the output variable is the hydrogen consumption rate of the fuel cell. The loss term includes the hydrogen leakage and compression energy consumption per unit time. The hydrogen storage capacity at the next moment is obtained by superimposing the difference between the hydrogen production, hydrogen consumption and loss of the initial value of hydrogen storage capacity. This represents the dynamic process of energy conversion, storage and loss of hydrogen energy storage.

[0036] For electric heating devices that convert electricity to heat, electrolytic hydrogen production devices that convert electricity to hydrogen, and fuel cell devices that convert hydrogen to electricity, 10 steady-state operating points were selected at 10% power intervals within the full rated power range. After each operating point was run stably for 10 minutes, the average values ​​of the input electrical power and the output thermal power or hydrogen energy were recorded. With the input power as the independent variable and the energy conversion efficiency as the dependent variable, the efficiency function of each energy conversion device was obtained by fitting a third-order polynomial. The coefficient of determination of the fitting of all efficiency functions was not less than 0.95, which served as the basis for quantitative calculation of energy transfer between multiple energy flows.

[0037] Next, power balance equations are set for three types of energy nodes: electrical nodes, thermal nodes, and hydrogen nodes. It is agreed that the power of energy storage charging, heating, and hydrogen production is positive, while the power of discharging, releasing heat, and generating electricity is negative. The hydrogen mass flow rate of the hydrogen node is uniformly converted to power dimensions using the lower calorific value of hydrogen, maintaining the same power unit as the electrical and thermal nodes to ensure consistent calculation dimensions for the three types of energy nodes. For electrical nodes, the instantaneous algebraic sum of the power generated by new energy sources, the power of electrochemical energy storage charging and discharging, the power generated by fuel cells, the power consumed by electric heating, the power transmitted to the grid, and the power consumed by the load is zero. For thermal nodes, the instantaneous algebraic sum of the power of thermal energy storage charging and discharging, the power output of electric heating, and the power of the heat load is zero. For hydrogen nodes, the instantaneous algebraic sum of the converted power of hydrogen production by electrolysis, the converted power of hydrogen consumption by fuel cells, and the converted power of changes in hydrogen storage is zero.

[0038] Finally, the discrete state equations of the three types of energy storage and the energy conversion efficiency functions are substituted into the power balance equations of the corresponding nodes. The unified state vectors are the state of charge of electrochemical energy storage, the heat storage of thermal energy storage, and the hydrogen storage of hydrogen energy storage. The unified input vectors are the power of new energy generation, the grid-connected power, the electrical load power, and the thermal load power. The Euler discretization method is used to integrate them into a multi-energy-dimensional nonlinear coupled state-space model, which can be directly used for subsequent rolling optimization calculations.

[0039] By constructing a computable nonlinear state equation for a single type of energy storage step by step, clarifying the specific calculation methods for each loss term and the unified rules for multi-energy flow dimensions, and combining standardized efficiency fitting methods and node power balance rules, the dynamic nonlinear characteristics of multi-energy flow conversion and storage processes can be accurately characterized, providing a directly callable model calculation basis for subsequent coordinated operation control.

[0040] Furthermore, the method provided in this application embodiment includes: Collect new energy power output forecast data and load demand forecast data at multiple time scales; perform outlier removal, missing value completion, and time sequence alignment on the new energy power output forecast data and the load demand forecast data in sequence to obtain a standardized forecast dataset, wherein the forecast dataset includes new energy power generation forecast data and load forecast data.

[0041] Optionally, a power measurement and control terminal and a data acquisition gateway are first used to deploy acquisition units at the grid connection points of new energy power generation and load incoming nodes. Multi-timescale new energy output forecast data and load demand forecast data for the corresponding nodes are obtained through a station-level energy management system. During the acquisition process, the timestamp and power value corresponding to each set of forecast data are recorded synchronously. The acquisition timescale can include three typical scales: 15 minutes, 1 hour, and 24 hours. Examples of acquired power parameters include photovoltaic power generation active power, wind power active power, and comprehensive electricity load active power. The data is transmitted to the control and processing unit via industrial Ethernet to complete the aggregation and storage of the raw forecast data.

[0042] Next, preprocessing operations are performed on the aggregated raw forecast data. First, outlier removal is carried out, using the 3σ criterion as the basis for judging outliers. Continuous data sequences within the current scheduling period are selected as statistical samples, for example, with a statistical period of 24 hours. The arithmetic mean and standard deviation of data sequences at the same time scale within this period are calculated. Data points whose values ​​deviate from the mean by more than three times the standard deviation are judged as outliers and directly removed. The remaining valid data that conforms to the deviation range are retained.

[0043] After outlier removal, missing values ​​are filled in for the data sequence caused by outlier removal or transmission interruption. Linear interpolation is used for the filling calculation. Two adjacent valid data points before and after the missing point are selected as interpolation endpoints. The corresponding filling value is calculated based on the time interval and numerical difference between the two endpoints, combined with the time position of the missing point. The filling value is then filled into the missing position to form a continuous single-class data sequence.

[0044] After completing the missing values ​​of the two types of data, time-series alignment processing is performed. Using a preset standard time base as a unified time axis, linear resampling interpolation is used to adjust the new energy output forecast data and load demand forecast data to the same time interval and time starting point. For example, they are uniformly adjusted to a data sequence of equal duration with a 15-minute interval, so that the corresponding new energy forecast value and load forecast value exist at the same time stamp, and finally a standardized forecast dataset is formed.

[0045] By deploying a multi-node, multi-scale data acquisition process and performing a three-level standardized preprocessing operation, a standardized prediction dataset with time-series matching and complete data can be obtained, providing a stable and reliable data input foundation for the accurate calculation of the subsequent new energy absorption capacity index.

[0046] Furthermore, the method provided in this application embodiment includes: Based on a standardized forecast dataset, the power surplus between the forecast output of new energy sources and the upper limit of grid-connected transmission power is calculated, as well as the total adjustable capacity that each energy storage unit can provide under the current energy storage state. The ratio of the power surplus to the total adjustable capacity is used as the new energy absorption capacity index. When the new energy absorption capacity index is less than or equal to 1, it indicates that the system has full absorption capacity, and when it is greater than 1, it indicates that there is a risk of energy curtailment.

[0047] Specifically, the system extracts the predicted power output values ​​of various new energy sources at the same timestamp from the standardized prediction dataset, and sums the predicted power output of photovoltaic power generation and wind power generation to obtain the total predicted power output of new energy sources. It then retrieves the system's preset upper limit parameter for grid-connected transmission power, which is the maximum allowed transmission power setting at the grid connection point. This parameter can be set according to the grid-connected dispatch protocol; in this example, it is set to 90% of the rated transmission power at the grid connection point. The difference between the total predicted power output of new energy sources and the upper limit of grid-connected transmission power is calculated point by point. When the difference is greater than 0, this difference represents the power surplus at the current moment. When the difference is less than or equal to 0, the power surplus at the current moment is set to 0. This process is repeated for each time point within the entire prediction time domain, forming a power surplus sequence corresponding to the given time period.

[0048] Next, the current operating status parameters and rated parameters of the equipment for electrochemical energy storage, thermal energy storage, and hydrogen energy storage are read respectively. For electrochemical energy storage, the current state of charge (SOC) and its upper limit are read. If the current SOC is lower than the upper limit, the adjustable absorption capacity is taken as the unit's maximum charging power. If the current SOC reaches or exceeds the upper limit, the adjustable absorption capacity is 0. For thermal energy storage, the current heat storage and its upper limit are read. If the current heat storage is lower than the upper limit, the adjustable absorption capacity is taken as the unit's maximum heat charging power. If the current heat storage reaches or exceeds the upper limit, the adjustable absorption capacity is 0. For hydrogen energy storage, the current hydrogen storage and its upper limit are read. If the current hydrogen storage is lower than the upper limit, the adjustable absorption capacity is taken as the unit's maximum hydrogen production power. If the current hydrogen storage reaches or exceeds the upper limit, the adjustable absorption capacity is 0. Finally, the adjustable absorption capacity values ​​of the three types of energy storage are added together to obtain the total adjustable capacity of the system at the current moment.

[0049] Then, the current power surplus is used as the numerator, and the current total adjustable capacity is used as the denominator. The ratio of the two is calculated to obtain the renewable energy absorption capacity index. When the total adjustable capacity is 0, the renewable energy absorption capacity index is directly determined to be greater than 1. After completing the index calculation for each time point in the full time domain, the maximum value of the renewable energy absorption capacity index in the current prediction time domain is selected as the parameter configuration basis for the subsequent hierarchical model predictive control architecture. After the index value is determined, a state determination is performed. When the renewable energy absorption capacity index is less than or equal to 1, it indicates that the adjustable capacity of energy storage can cover all surplus power, and the system has full absorption capacity. When the renewable energy absorption capacity index is greater than 1, it indicates that the surplus renewable energy power exceeds the adjustable capacity of energy storage, and the system faces the risk of energy curtailment.

[0050] By calculating the ratio of power surplus to total adjustable capacity of multiple energy storage points, a quantitative absorption capacity index is obtained, and the value selection rules in the prediction time domain are clarified. This can intuitively and accurately reflect the new energy absorption capacity of the system at each moment, providing a clear and reliable quantitative basis for the dynamic configuration of parameters in the subsequent hierarchical control architecture.

[0051] Furthermore, the method provided in this application embodiment includes: A hierarchical model predictive control architecture comprising a strategic layer, a coordination layer, and an execution layer is constructed. Based on the new energy absorption capacity index, the optimization time domain, constraints, and decision weights for each layer are set to complete the architecture configuration with the joint optimization goal of smoothing grid-connected power fluctuations and reducing system curtailment rate.

[0052] Specifically, a hierarchical model predictive control architecture is built within the control and computation unit, divided into three logical computational layers from top to bottom: a strategy layer, a coordination layer, and an execution layer. Each layer employs a sequential, progressive computational connection relationship. The optimization results output from the upper layer directly serve as the constraint boundaries for the lower layer's computations. Real-time operational status data collected from the lower layer is synchronously fed back to the upper layer, forming a top-down instruction transmission channel and a bottom-up status feedback channel. Each layer shares the pre-constructed nonlinear coupling model and standardized prediction dataset as the computational basis, and uses the same system time benchmark for iterative calculations to ensure consistent computational timing across all layers. The specific construction steps are as follows: A strategic layer computation module is constructed, positioned as the global energy planning level. New energy forecast data and load forecast data across the entire forecast time domain, along with system grid connection transmission constraints, are used as input parameters for the strategic layer. The long-term charge / discharge plan curves of each energy storage unit and the grid-connected power baseline value are used as output results. The strategic layer employs a linear programming solver to perform optimization calculations, with a typical optimization time domain set to 24 hours and a computation cycle of 1 hour. It focuses on the global optimization of the total system energy curtailment across the entire time domain, ensuring the rationality of energy allocation in long-term operation.

[0053] A coordination layer computation module is constructed, positioned as a multi-energy storage collaborative scheduling level. The long-term charging / discharging plan and grid-connected power benchmark value output from the strategic layer are used as the input constraints of the coordination layer, while the mid-term power command range and operation mode determination results of various energy storage units are used as the output results. The coordination layer employs a quadratic programming solver to perform optimization calculations, with a typical optimization time domain set to 2 hours and a computation cycle set to 15 minutes. The focus is on smoothing grid-connected power fluctuations and balancing the load of multiple energy storage units within the time domain, ensuring that the scheduling scheme adapts to short- and medium-term fluctuations in source and load.

[0054] An execution layer computation module is constructed, positioned as the real-time power control level. The mid-time-domain power command range output from the coordination layer is used as the input constraint boundary for the execution layer, outputting real-time charging and discharging power commands for each energy storage device. These commands are directly sent to field execution devices such as electrochemical energy storage converters, thermal energy storage controllers, and hydrogen energy storage-to-hydrogen power generation controllers. The execution layer employs a model prediction feedforward combined with proportional-integral feedback correction to complete real-time computation. First, a power tracking baseline adjustment is calculated based on a nonlinear coupled model as the feedforward value. Then, the deviation between the actual operating power and the command value is collected, and the feedback correction is calculated through a proportional-integral loop. The feedforward value and the feedback correction are superimposed to obtain the final output command. The typical optimization time domain for the execution layer is set to 15 minutes, and the computation cycle is set to 1 minute, the shortest of the three levels, emphasizing the tracking accuracy and response speed of control commands within a short time domain to ensure that the control effect adapts to real-time changes in the system's state.

[0055] Next, the renewable energy absorption capacity index within the current prediction time domain is obtained, and the parameters for each level are configured according to the index values. The joint optimization objective adopts a weighted summation objective function. The calculated objective function value equals the curtailment rate weight multiplied by the system curtailment rate index plus the power fluctuation weight multiplied by the grid-connected power fluctuation index. The initial default values ​​for both weights are set to 0.5. For example, when the renewable energy absorption capacity index is greater than 1, corresponding to a system with curtailment risk, the optimization time domain of the strategic layer is extended to 36 hours, the curtailment rate weight is adjusted to 0.7, the power fluctuation weight is adjusted to 0.3, and the upper limit constraint of grid-connected transmission power is tightened, reducing the allowable fluctuation range of grid-connected power within a single 15-minute interval from 10% to 5% of the rated grid-connected power; the optimization time domains of the coordination layer and the execution layer are shortened simultaneously to improve the response speed of power regulation.

[0056] For example, when the renewable energy absorption capacity index is less than or equal to 1, corresponding to the system having full absorption capacity, the optimization time domain of the strategic layer is shortened to 12 hours, the curtailment rate weight is adjusted to 0.3, and the power fluctuation weight is adjusted to 0.7. At the same time, the fluctuation range constraint of grid-connected power is relaxed, and the allowable fluctuation range of grid-connected power within a single 15-minute interval is relaxed from 10% to 15% of the rated grid-connected power. Simultaneously, the optimization time domain of the coordination layer is extended to improve the operational stability of multi-energy storage scheduling. After completing the configuration of the optimization time domain, constraints, and decision weights of each level, the corresponding parameters are written into the objective function and constraint condition modules of each level solver, completing the configuration of the entire hierarchical model predictive control architecture.

[0057] By constructing a three-level progressive hierarchy, defining the computation cycle and time-domain parameter configuration, and combining dynamic weights and constraint adaptation based on the absorption capacity index, a hierarchical control architecture that balances global long-term optimization and real-time precise adjustment can be built, providing a complete and executable computational framework for the rolling optimization solution of multi-type energy storage collaborative operation.

[0058] Furthermore, the method provided in this application embodiment includes: The system updates real-time operating status and prediction data in each control cycle, and solves the optimization problem at each level in the order of strategic layer, coordination layer and execution layer. After solving at each level, the optimization results are passed down to the next level as constraint boundaries. The charging and discharging power and operating sequence of each type of energy storage are output through rolling iteration, and the charging and discharging coordinated control commands are output.

[0059] Specifically, using the 1-minute computation cycle of the execution layer as the basic rolling unit, at the beginning of each control cycle, the real-time operating status and predicted data of the system are updated first. Real-time operating values ​​of electrochemical energy storage state of charge, thermal energy storage heat storage, and hydrogen energy storage hydrogen storage are collected through the field power measurement and control terminal. The collected state parameters are written into the nonlinear coupling model as the initial state values ​​for the current solution cycle. At the same time, a rolling window update is performed on the standardized prediction dataset. The rolling step size is consistent with the computation cycle of the execution layer. Each rolling step removes the historical data at the front of the window and adds the latest predicted data corresponding to the end of the window, forming prediction data sequences that match the optimization time domain length of the strategy layer, coordination layer, and execution layer, providing a unified and up-to-date input basis for optimization solutions at each level.

[0060] Every 60 execution-level control cycles completed corresponds to the trigger time of the 1-hour strategic-level computation cycle, initiating the strategic-level optimization solution. The updated full-time-domain prediction data, grid-connected transmission constraint parameters, and the system's real-time initial state are substituted into the strategic-level linear programming solver. The configured weighted summation joint optimization objective function is used as the optimization target, and the upper and lower limits of capacity and rated charging / discharging power for each energy storage unit are used as basic constraints. The linear programming solver is invoked to complete the global optimization calculation. The solution process uses a convergence criterion of the objective function calculation value being less than one-thousandth between two adjacent iterations. If a constraint conflict occurs and there is no solution, the grid-connected power fluctuation constraint is relaxed first, and the solution is re-solved. The solution yields the charging / discharging plan curves and grid-connected power benchmark values ​​for each energy storage unit throughout the entire strategic optimization time domain. After the strategic-level solution is completed, the output global programming results are passed down to the coordination layer as the global constraint boundary for the coordination layer's computation.

[0061] Every 15 execution-level control cycles completed corresponds to the trigger moment of the 15-minute operation cycle at the coordination level, initiating optimization at the coordination level. The charging / discharging plan curves issued by the strategic level and the grid-connected power benchmark value are used as boundary constraints. Combined with updated mid-time domain prediction data and the system's real-time initial state, these are substituted into the quadratic programming solver of the coordination level. Using the configured joint optimization objective function as the optimization target, optimization calculations are performed within the allowable power fluctuation range given by the strategic level. The solution process uses a convergence criterion of a deviation of less than 0.5% between adjacent iterations of the objective function calculation values. If a constraint conflict arises and there is no solution, the allowable range is slightly expanded based on the power range given by the strategic level, and the solution is recalculated. The solution yields the upper and lower limits of the mid-time domain power command and the operation mode determination results for each energy storage unit. After the coordination level completes its solution, the output power command range is passed down to the execution level as the constraint range for real-time power control.

[0062] Each execution-level control cycle initiates a solution operation at the execution level. Using the power command range issued by the coordination layer as the constraint boundary, and substituting the current real-time system state as the initial value into the nonlinear coupling model, the power tracking reference adjustment is calculated and used as the feedforward value. Simultaneously, the actual operating power of each energy storage device from the previous control cycle is collected, and the deviation between the actual operating value and the target command value is calculated. The feedback correction is then calculated through a proportional-integral control loop. The feedforward value and the feedback correction are superimposed and constrained within the power command range given by the coordination layer to obtain the charging and discharging power control command for each energy storage unit at the current moment. The control command is directly issued to field execution devices such as the electrochemical energy storage converter, thermal energy storage heat exchange controller, and hydrogen energy storage hydrogen production and power generation controller, completing a single rolling solution. As the control cycle progresses, the state update and layer-by-layer solution process is continuously repeated, forming a continuous closed-loop control with rolling time-domain iteration.

[0063] By employing clearly defined multi-level nested triggering rules, quantified rolling window update step size, convergence judgment, and fault tolerance degradation mechanisms, combined with top-down hierarchical progressive solution iteration, the system can continuously and stably output multi-type energy storage collaborative control commands that adapt to the real-time operating conditions of the system, ensuring that the multi-energy complementary energy storage system maintains stable grid-connected power and efficient new energy consumption throughout the entire process.

[0064] Furthermore, the method provided in this application embodiment includes: Collect actual operating feedback data from the system, compare the deviation between the actual operating parameters and the model prediction results; based on the deviation, correct and compensate the charging and discharging coordinated control command for the next control cycle, and continuously update the characteristic parameters and constraint boundaries of the nonlinear coupling model.

[0065] In one embodiment, at the end of each control cycle, the actual charge / discharge power and actual state of charge of the electrochemical energy storage unit, the actual charge / discharge power and actual heat storage of the thermal energy storage unit, the actual operating power and actual hydrogen storage of the hydrogen energy storage unit, and the actual transmission power data of the grid connection point are synchronously collected through the power measurement and control terminal and state sensing unit deployed on-site. All collected data are marked with corresponding timestamps, corresponding one-to-one with the time nodes of the control cycle. The corresponding predicted values ​​output by the nonlinear coupling model in the previous control cycle are retrieved, and the difference between the actual value and the predicted value of each operating parameter is calculated to obtain the deviation of each state quantity and power quantity. The deviation is calculated by subtracting the model's predicted value from the actual operating value, forming a deviation data sequence corresponding to each parameter. The power quantity deviation is directly used for real-time correction of the control command in the next control cycle, while the state quantity deviation is used to verify the state tracking accuracy of the model and provide data support for subsequent constraint boundary adjustments and model parameter updates.

[0066] Subsequently, a proportional compensation method is used to correct and compensate the control commands for the next control cycle. A compensation coefficient for each energy storage unit is pre-set, with a typical value range of 0.2 to 0.5. The power deviations calculated in the previous cycle are multiplied by the corresponding compensation coefficient to obtain the command correction for the next control cycle. This correction is then superimposed on the baseline control command value for the next cycle. The superimposed command value is then limited to the power command range issued by the coordination layer to prevent exceeding the equipment's rated operating range. Finally, the corrected charge-discharge coordinated control command is obtained for equipment operation control in the next cycle.

[0067] Using a 24-hour update cycle, full-condition operating data is collected within one update cycle. First, steady-state operating data is screened. The criterion is that if the power fluctuation within 5 consecutive minutes does not exceed 2% of the corresponding equipment's rated power, the operating data for that period is considered steady-state data. After screening the operating data points of each energy conversion device and energy storage unit under steady-state conditions, the polynomial coefficients of each efficiency function are refitted using the least squares method. The characteristic parameters of the electrochemical energy storage charge / discharge efficiency function, thermal storage heat transfer efficiency function, electrolytic hydrogen production efficiency function, and fuel cell power generation efficiency function are updated. The updated model characteristic parameters officially take effect in the first strategic-layer solution cycle after the update is completed, replacing the original parameters in the nonlinear coupling model, completing the rolling correction of the model characteristic parameters, and adapting to changes in equipment performance over operating time.

[0068] Before each strategic-level optimization solution is initiated, the actual available capacity boundaries and power operating boundaries of each energy storage unit are verified based on the latest collected system operating status data. For example, when the actual available capacity of an energy storage unit deviates from the originally set upper limit by more than 5%, a capacity constraint adjustment is triggered, with a single adjustment step not exceeding 3% of the original set value, corresponding to lowering or raising the upper limit constraint of the state variables of that energy storage unit in the model. When the actual maximum operating power of the equipment deviates from the original rated value by more than 5%, a power constraint adjustment is triggered, with a single adjustment step not exceeding 3% of the original rated value, corresponding to adjusting the upper and lower limits of charging and discharging power constraints in the model. The updated constraint boundaries officially take effect in this strategic-level solution cycle and are simultaneously distributed to solvers at all levels as the constraint basis for the next stage of optimization solution.

[0069] By collecting real-time operational feedback data to calculate prediction deviations, combining proportional compensation to correct control commands, and using quantified steady-state screening rules and constraint adjustment thresholds with clearly defined update activation nodes, the error between model predictions and actual operation can be continuously reduced, improving the tracking accuracy of collaborative control commands and ensuring the long-term reliability and control effectiveness of multi-energy complementary energy storage systems.

[0070] In summary, the collaborative operation control method for a multi-energy complementary energy storage system provided in this application has the following technical effects: This application constructs a nonlinear coupled model of electric, thermal, and hydrogen energy storage, collects and standardizes source-load prediction data to calculate the new energy consumption index, configures a three-layer hierarchical model prediction and control architecture based on the index, generates energy storage collaborative control commands through layered rolling solutions, collects operational feedback to calculate deviation correction commands and updates model parameter boundaries, and realizes multi-energy storage collaborative regulation, smooths grid-connected power fluctuations, reduces system curtailment rate, and achieves the technical effect of dynamic collaborative adjustment of multiple energy storage devices, stabilizing grid-connected output, and reducing new energy waste.

[0071] Example 2, as Figure 2As shown, based on the same inventive concept as in Embodiment 1 above, this application provides a multi-energy complementary energy storage system collaborative operation control system, the system comprising: The coupling model construction module 1 is used to construct a nonlinear coupling model of a multi-energy complementary energy storage system, which is used to characterize the dynamic nonlinear characteristics of the multi-energy flow conversion and storage process.

[0072] Predictive data acquisition module 2 is used to acquire new energy power generation prediction data and load prediction data, and calculate the new energy absorption capacity index.

[0073] The control architecture construction module 3 is used to construct a hierarchical model predictive control architecture by configuring the corresponding parameters of the hierarchical model predictive control architecture according to the new energy absorption capacity index, with the joint optimization objectives of smoothing grid-connected power fluctuations and reducing system curtailment rate.

[0074] The control command generation module 4 is used to perform rolling time-domain solution of the nonlinear coupled model through the hierarchical model predictive control architecture to generate charging and discharging coordinated control commands for various types of energy storage.

[0075] Furthermore, the coupling model construction module 1 is used to perform the following steps: Nonlinear mathematical descriptions of the energy conversion, storage, and loss processes corresponding to electrochemical energy storage, thermal energy storage, and hydrogen energy storage are established respectively. By coupling and associating the nonlinear mathematical descriptions with the energy conversion efficiency function and power balance constraint relationship between the electric, thermal, and hydrogen multi-energy flows, a multi-energy-dimensional nonlinear coupling model is constructed.

[0076] Furthermore, the prediction data acquisition module 2 is used to perform the following steps: Collect new energy power output forecast data and load demand forecast data at multiple time scales; perform outlier removal, missing value completion, and time sequence alignment on the new energy power output forecast data and the load demand forecast data in sequence to obtain a standardized forecast dataset, wherein the forecast dataset includes new energy power generation forecast data and load forecast data.

[0077] Furthermore, the prediction data acquisition module 2 is used to perform the following steps: Based on a standardized forecast dataset, the power surplus between the forecast output of new energy sources and the upper limit of grid-connected transmission power is calculated, as well as the total adjustable capacity that each energy storage unit can provide under the current energy storage state. The ratio of the power surplus to the total adjustable capacity is used as the new energy absorption capacity index. When the new energy absorption capacity index is less than or equal to 1, it indicates that the system has full absorption capacity, and when it is greater than 1, it indicates that there is a risk of energy curtailment.

[0078] Furthermore, the control architecture construction module 3 is used to perform the following steps: A hierarchical model predictive control architecture comprising a strategic layer, a coordination layer, and an execution layer is constructed. Based on the new energy absorption capacity index, the optimization time domain, constraints, and decision weights for each layer are set to complete the architecture configuration with the joint optimization goal of smoothing grid-connected power fluctuations and reducing system curtailment rate.

[0079] Furthermore, the control command generation module 4 is used to perform the following steps: The system updates real-time operating status and prediction data in each control cycle, and solves the optimization problem at each level in the order of strategic layer, coordination layer and execution layer. After solving at each level, the optimization results are passed down to the next level as constraint boundaries. The charging and discharging power and operating sequence of each type of energy storage are output through rolling iteration, and the charging and discharging coordinated control commands are output.

[0080] Furthermore, the control command generation module 4 is used to perform the following steps: Collect actual operating feedback data from the system, compare the deviation between the actual operating parameters and the model prediction results; based on the deviation, correct and compensate the charging and discharging coordinated control command for the next control cycle, and continuously update the characteristic parameters and constraint boundaries of the nonlinear coupling model.

[0081] The multi-energy complementary energy storage system collaborative operation control system provided in the embodiments of the present invention can execute the multi-energy complementary energy storage system collaborative operation control method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0082] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0083] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for coordinated operation control of a multi-energy complementary energy storage system, characterized in that, Includes the following steps: A nonlinear coupling model of a multi-energy complementary energy storage system is constructed to characterize the dynamic nonlinear characteristics of the multi-energy flow conversion and storage process; Obtain new energy power generation forecast data and load forecast data, and calculate the new energy absorption capacity index; With the joint optimization goal of smoothing grid-connected power fluctuations and reducing system curtailment rate, a hierarchical model predictive control architecture is constructed by configuring the corresponding parameters of the hierarchical model predictive control architecture according to the new energy absorption capacity index. The nonlinear coupled model is solved in the rolling time domain by the hierarchical model predictive control architecture to generate charging and discharging coordinated control commands for various types of energy storage.

2. The method for coordinated operation control of a multi-energy complementary energy storage system as described in claim 1, characterized in that, Constructing a nonlinear coupling model for a multi-energy complementary energy storage system, including: Nonlinear mathematical descriptions of the energy conversion, storage, and loss processes for electrochemical energy storage, thermal energy storage, and hydrogen energy storage are established respectively. By coupling and associating the energy conversion efficiency function and power balance constraint relationship between the electric, thermal, and hydrogen multi-energy flows, the nonlinear mathematical description is constructed into a multi-energy-dimensional nonlinear coupled model.

3. The method for coordinated operation control of a multi-energy complementary energy storage system as described in claim 1, characterized in that, Obtain renewable energy power generation forecast data and load forecast data, including: Collect new energy output forecast data and load demand forecast data at multiple time scales; The new energy power output forecast data and the load demand forecast data are sequentially processed by outlier removal, missing value completion, and time sequence alignment to obtain a standardized forecast dataset, which includes new energy power generation forecast data and load forecast data.

4. The method for coordinated operation control of a multi-energy complementary energy storage system as described in claim 1, characterized in that, The calculated renewable energy absorption capacity index includes: Based on a standardized forecast dataset, the power surplus between the forecast output of new energy sources and the upper limit of grid-connected transmission power is calculated, as well as the total adjustable capacity that each energy storage unit can provide under the current energy storage state. The ratio of the power surplus to the total adjustable capacity is used as the new energy absorption capacity index. When the new energy absorption capacity index is less than or equal to 1, it indicates that the system has full absorption capacity; when it is greater than 1, it indicates that there is a risk of energy curtailment.

5. The method for coordinated operation control of a multi-energy complementary energy storage system as described in claim 1, characterized in that, Constructing a hierarchical model predictive control architecture, including: Construct the hierarchical model predictive control architecture comprising a strategy layer, a coordination layer, and an execution layer; Based on the new energy absorption capacity index, the optimization time domain, constraints and decision weights of each level are set respectively, and the architecture configuration is completed with the joint optimization goal of smoothing grid-connected power fluctuations and reducing system curtailment rate.

6. The method for coordinated operation control of a multi-energy complementary energy storage system as described in claim 1, characterized in that, The nonlinear coupled model is solved in the rolling time domain using the hierarchical model predictive control architecture, including: The system updates its real-time operating status and forecast data within each control cycle, and solves the optimization problem at each level in the order of strategic layer, coordination layer, and execution layer. After solving each layer, the optimization results are passed down to the next layer as constraint boundaries. The charging and discharging power and operating sequence of various types of energy storage are output through rolling iterations, which are the charging and discharging coordinated control commands.

7. The method for coordinated operation control of a multi-energy complementary energy storage system as described in claim 1, characterized in that, After generating charging and discharging coordinated control commands for various types of energy storage, the following are also included: Collect actual operating feedback data from the system and compare the deviation between the actual operating parameters and the model prediction results; The charging and discharging coordinated control command for the next control cycle is corrected and compensated based on the deviation, and the characteristic parameters and constraint boundaries of the nonlinear coupling model are updated on a rolling basis.

8. A collaborative operation control system for a multi-energy complementary energy storage system, characterized in that, A method for implementing the collaborative operation control of a multi-energy complementary energy storage system according to any one of claims 1-7, the system comprising: The coupling model construction module is used to construct a nonlinear coupling model of a multi-energy complementary energy storage system, which is used to characterize the dynamic nonlinear characteristics of the multi-energy flow conversion and storage process; The forecast data acquisition module is used to acquire forecast data of new energy power generation and load forecast data, and calculate the new energy absorption capacity index. The control architecture construction module is used to construct a hierarchical model predictive control architecture by configuring the corresponding parameters of the hierarchical model predictive control architecture according to the new energy absorption capacity index, with the joint optimization objectives of smoothing grid-connected power fluctuations and reducing system curtailment rate. The control command generation module is used to perform rolling time-domain solution of the nonlinear coupled model through the hierarchical model predictive control architecture to generate charging and discharging coordinated control commands for various types of energy storage.

Images