Compressed air energy storage system control method and system considering parameter time-varying characteristics

Abstract

The invention relates to the technical field of compressed air energy storage, and provides a compressed air energy storage system control method and system considering parameter time-varying characteristics, and the method comprises the steps: constructing a compressed air energy storage system simulation model, collecting the operation data of the simulation model at each stage, and carrying out the preprocessing; constructing a neural state space model, and outputting a variable prediction value according to the preprocessed simulation model operation data; and establishing a compressed air energy storage system control model, and outputting a control strategy based on the variable predicted value by taking the minimum variable predicted value deviation and the control signal variation as targets. The objective of the invention is to solve the problem of control performance reduction caused by parameter time-varying characteristics and system complexity of a compressed air energy storage system.

Description

Technical Field

[0001] The present invention relates to the technical field of compressed air energy storage, and in particular to a method and system for controlling a compressed air energy storage system taking into account time-varying parameters. Background Art

[0002] In the context of dual carbon emissions, compressed air energy storage (CAES) has become a key technology for promoting renewable energy consumption and enhancing grid flexibility, thanks to its significant advantages, including large-scale energy storage, high efficiency, low cost, and environmental friendliness. During the energy storage phase, a compressor compresses air to a high pressure, while a heat exchange system recovers the heat energy generated during the compression process and stores it in a dedicated heat storage system. During the energy release phase, the high-pressure air is released and the stored heat energy is used to drive an expander to generate electricity. CAES is an effective means of smoothing out fluctuations in renewable energy, improving energy utilization, and providing rapid peak-shaving capabilities.

[0003] Since the compressed air energy storage system is a complex multi-input and multi-output system, its components are highly coupled, and the system parameters change over time, making it difficult for the control strategy of the traditional PID control system to meet its control needs. The operating parameters of the compressed air energy storage system change with time and external conditions, which leads to the instability of the system state and increases the operating risk. The traditional PID control system is difficult to effectively deal with such parameter time-varying problems; on the other hand, in order to control the different components of the system, the traditional PID control system needs to adopt multiple PID controllers. The coupling effect between these controllers may lead to a decrease in control performance, especially when the system complexity increases. Based on this, a compressed air energy storage system control method and system considering the time-varying parameters are proposed. Summary of the Invention

[0004] In order to solve the problem of degraded control performance caused by time-varying parameters and system complexity of compressed air energy storage systems, the present invention aims to provide a method and system for controlling compressed air energy storage systems that take time-varying parameters into account. The technical solutions adopted are as follows:

[0005] The first technical solution is a method for controlling a compressed air energy storage system taking into account time-varying parameters, the method comprising:

[0006] Build a simulation model of the compressed air energy storage system, collect and pre-process the simulation model operation data at each stage;

[0007] Construct a neural state space model and output variable predictions based on the preprocessed simulation model run data;

[0008] A control model for a compressed air energy storage system is established, and a control strategy is output based on variable prediction values ​​with the goal of minimizing the deviation of variable prediction values ​​and the change in control signals.

[0009] Furthermore, collecting the simulation model operation data at each stage and preprocessing it includes:

[0010] The operation data of the simulation model at each stage are divided into state variables, disturbance variables, control variables and result variables; the state variables include at least the heat exchanger outlet temperature, the gas storage temperature and the oil storage tank temperature; the disturbance variables include at least the compressor inlet temperature and the heat exchanger cold flow temperature; the control variables include at least the heat exchanger cold flow velocity and the opening and closing degree of the gas storage inlet valve.

[0011] Furthermore, collecting the simulation model operation data at each stage and preprocessing it also includes:

[0012] Define the state equation, which is expressed as:

[0013]

[0014] Where, represents the rate of change of the state variable at time t; Q represents the state transfer matrix; s(t) represents the state variable at time t; W represents the control influence matrix; u(t) represents the control variable at time t; E represents the disturbance influence matrix; d(t) represents the disturbance variable at time t;

[0015] Define the output equation, which is expressed as:

[0016] y(t)=Rx(t)

[0017] Where y(t) represents the result variable at time t; R represents the mapping matrix from state variables to result variables; and x(t) represents the state variable at time t.

[0018] Furthermore, a neural state space model is constructed, and the output variable prediction values ​​based on the preprocessed simulation model running data include:

[0019] Input the state variables, disturbance variables and control variables of the previous period, and use the neural state space model to predict the state variables of the next period;

[0020] Output the outcome variable for the next period based on the predicted state variable.

[0021] Furthermore, using the neural state space model to predict the state variables for the next period includes:

[0022] Input the preprocessed simulation model running data into the first layer of the MLP network;

[0023] According to the preset time step, the hidden layer output is calculated as the input of the next time step;

[0024] For each time step, the weights and biases of the current time step are linearly combined with the hidden layer output of the previous time step;

[0025] Use the hyperbolic tangent function as the activation function to introduce nonlinear features;

[0026] The output of the last layer of the MLP network is used as the predicted value of the state variable.

[0027] Furthermore, the objective function is expressed as:

[0028]

[0029] Where, U represents the control time domain; u i Indicates the control variable value of the i-th control device in the current control time domain; y i represents the predicted value of the state variable of the i-th control device; r i represents the preset value of the state variable of the i-th control device; λ represents the adjustment parameter; k represents the k-th time node; N represents the length of the control time domain.

[0030] Furthermore, the compressed air energy storage system is controlled based on the control strategy, and the control strategy is dynamically adjusted according to the prediction error through a closed-loop feedback mechanism.

[0031] Furthermore, a first constraint condition is included, and the first constraint condition is positioned as the state variable being within a preset upper and lower limit range of the state variable.

[0032] Furthermore, a second constraint condition is included, and the second constraint condition is defined as the control variable being within a preset upper and lower limit range of the control variable.

[0033] The second technical solution is a compressed air energy storage system control system that takes into account time-varying parameters, characterized in that the system includes:

[0034] a data acquisition module configured to use the compressed air energy storage system simulation model to collect operation data of the simulation model at each stage and perform preprocessing;

[0035] Prediction module, configured to output variable prediction values ​​based on preprocessed simulation model running data

[0036] A control module is configured to output a control strategy based on the variable prediction value and with the goal of minimizing the deviation of the variable prediction value and the change of the control signal;

[0037] The feedback module is configured to control the compressed air energy storage system based on the control strategy and dynamically adjust the control strategy according to the prediction error through a closed-loop feedback mechanism.

[0038] The present invention has the following beneficial effects:

[0039] The present invention provides a compressed air energy storage system control method and system that takes into account the time-varying nature of parameters. Through rolling time domain optimization, the system can better cope with the time-varying nature of parameters and external disturbances. The neural state space model can capture the complex nonlinear relationships within the system, so that the control system can adapt to the time-varying nature of parameters and the complexity of the system, thereby improving the stability of the system; by minimizing the deviation between the output variable and the set value, it is ensured that the system output tracks the set value more accurately; by penalizing the change in the control signal, a smooth control sequence is output, reducing unnecessary energy loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0041] Figure 1 A flow chart of a method for controlling a compressed air energy storage system taking into account time-varying parameters provided by one embodiment of the present invention;

[0042] Figure 2 A schematic structural diagram of a compressed air energy storage system control system considering time-varying parameters provided by one embodiment of the present invention;

[0043] Figure 3 A flow chart of the control algorithm for the energy storage process of a compressed air energy storage system provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0044] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail a method and system for controlling a compressed air energy storage system that takes into account the time-varying parameters proposed by the present invention, its specific implementation, structure, features and effects. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0045] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0046] The following describes in detail a method and system solution for controlling a compressed air energy storage system taking into account time-varying parameters provided by the present invention with reference to the accompanying drawings.

[0047] See also Figure 1 , which shows a method flow chart of a method for controlling a compressed air energy storage system taking into account time-varying parameters provided by an embodiment of the present invention,

[0048] The first technical solution is a method for controlling a compressed air energy storage system taking into account time-varying parameters, the method comprising:

[0049] Step S100: constructing a compressed air energy storage system simulation model, collecting and preprocessing the simulation model operation data at each stage;

[0050] Step S100 specifically includes:

[0051] Step S110: Divide the simulation model operation data of each stage into state variables, disturbance variables, control variables and result variables; the state variables include at least the heat exchanger outlet temperature, the gas storage temperature and the oil storage tank temperature; the disturbance variables include at least the compressor inlet temperature and the heat exchanger cold flow temperature; the control variables include at least the heat exchanger cold flow velocity and the opening and closing degree of the gas storage inlet valve; the result variable is the result of the system output, such as the gas storage inlet temperature; since the collected compressed air energy storage system data has problems such as multiple data types, poor data quality and data imbalance, it is also necessary to remove or process missing values ​​and outliers to improve data reliability; unify the sampling frequencies of different sensors to facilitate data integration and analysis; and convert the data to the same numerical range to help improve the training efficiency and generalization ability of the model.

[0052] Step S120: define the state equation, which is expressed as:

[0053]

[0054] Where, represents the rate of change of the state variable at time t; Q represents the state transfer matrix; s(t) represents the state variable at time t; W represents the control influence matrix; u(t) is the control variable at time t; E represents the disturbance influence matrix; d(t) represents the disturbance variable at time t; specifically, the state transfer matrix is ​​determined according to the dynamic characteristics of the system and is used to reflect the interaction between the state variables; the control influence matrix is ​​determined according to the influence of the control variable on the state variable and is used to reflect the effect of the control variable; the disturbance influence matrix is ​​determined according to the influence of the disturbance variable on the state variable and is used to reflect the effect of the disturbance variable;

[0055] Step S130: define the output equation, which is expressed as:

[0056] y(t)=Rx(t)

[0057] Where y(t) represents the result variable at time t; R represents the mapping matrix from state variables to result variables; x(t) represents the state variable at time t; the mapping matrix is ​​determined based on the relationship between the system output and the state variable, and is used to reflect how the state variable is mapped to the output variable;

[0058] Step S200: Construct a neural state space model and output variable prediction values ​​based on the preprocessed simulation model running data; specifically, the neural state space model predicts the behavior of the compressed air energy storage system in the future, taking into account the time delay and dynamic interaction between input and output. The prediction module uses the collected data to fit the neural state space model, and uses the input state variables, disturbance variables and control variables of the previous prediction time domain to predict the output result variables of the current prediction time domain.

[0059] Step S200 specifically includes:

[0060] Step S210: input the state variables, disturbance variables and control variables of the previous time period, and use the neural state space model to predict the state variables of the next time period; specifically, input data preparation: state variables: state variables s(t-1) of the previous time period, such as heat exchanger outlet temperature, gas storage temperature, oil storage tank temperature, etc.; disturbance variables: disturbance variables d(t-1) of the previous time period, such as compressor intake temperature, heat exchanger cold flow temperature, etc.; control variables: control variables u(t-1) of the previous time period, such as heat exchanger cold flow velocity, gas storage inlet valve opening and closing degree, etc.; use multilayer perceptron (MLP) as the neural state space model, input the state variables, disturbance variables and control variables of the previous time period into the neural network, and the neural network outputs the state variables of the next time period.

[0061] Step S210 specifically includes:

[0062] Step S211: Input the preprocessed simulation model operation data into the first layer of the MLP network; organize the state variables, disturbance variables, and control variables of the previous period to ensure that the data has been preprocessed (resampling, normalization, etc.), and organize these data into a vector or tensor form to serve as the first layer input of the neural network;

[0063] Step S212: Calculate the hidden layer output as the input for the next time step according to the preset time step;

[0064] Step S213: For each time step, use the weights and biases of the current time step and the hidden layer output of the previous time step to perform a linear combination;

[0065] Step S214: using the hyperbolic tangent function as an activation function to introduce nonlinear features;

[0066] Step S215: The output of the last layer of the MLP network is used as the predicted value of the state variable;

[0067] The expression of the above neural state space model is:

[0068] h (t) =tanh(W (t) h (t-1) +b (t) )

[0069] Where h (t) Represents the prediction result, tanh represents the activation function, W (t) Indicates the preset weight, h (t-1) represents the output of the previous hidden layer, b (t) is bias;

[0070] This neural state space model uses a neural network with a tanh activation function to better capture the nonlinear relationship of the system and improve prediction accuracy; it uses data from the previous period to predict the state of the next period, taking into account time delays and dynamic interactions, thereby improving prediction accuracy. By capturing the dynamic behavior of the system through the neural network, it can better predict the future state even if the parameters vary with time; and rolling optimization control strategies within each prediction time domain enable the system to better cope with parameter time variability and external disturbances.

[0071] Step S220: Output the result variable of the next time period based on the predicted state variable; specifically, obtain the state variable of the next time period predicted in the previous step (step S210), and use the output equation to map the state variable to the result variable, that is, according to the output equation, use the mapping matrix to perform a linear combination of the predicted state variables to obtain the result variable; calculate the result variable based on the predicted value of the state variable, so that the model can have better generalization ability under different working conditions, predict the state variable through the neural network, and then calculate the result variable through the output equation. The whole process can dynamically adapt to changes in system parameters.

[0072] Step S300: Establish a control model for the compressed air energy storage system, and output a control strategy based on the variable prediction value and with the goal of minimizing the deviation of the variable prediction value and the change in the control signal; specifically, the expression of the objective function is:

[0073]

[0074] Where, U represents the control time domain; u i Indicates the control variable value of the i-th control device in the current control time domain; y i represents the predicted value of the state variable of the i-th control device; r irepresents the preset value of the state variable of the i-th control device; λ represents the adjustment parameter; k represents the k-th time node; N represents the length of the control time domain; through the above objective function, ensure that the difference between the output result and the set value is minimized and avoid too frequent or drastic control actions. By solving the above objective function, find the optimal control sequence so that the output result y is within the control time domain. i As close as possible to the preset value r i , while the control signal u i The change in y is as small as possible, and the objective function can be solved using gradient descent, sequential quadratic programming (SQP), interior point method or other optimization algorithms. Specifically, an initial control sequence can be selected and the control sequence can be gradually adjusted using the optimization algorithm. The number of iterations can be determined based on the change in the objective function value or the preset maximum number of iterations. The control strategy is continuously adjusted based on the latest predicted values ​​and actual measured values ​​to cope with parameter time variability and external disturbances. By minimizing the output result y i With the default value r i The deviation between them can ensure that the system output tracks the set value more accurately; by punishing the control signal u i The change in , encourages the controller to output a smooth control sequence.

[0075] In order to avoid a large difference between the control results and the physical world and reduce the amount of calculation, the first constraint and the second constraint are added to the neural model predictive controller. The first constraint is defined as the state variable being within the preset upper and lower limits of the state variable, and the second constraint is defined as the control variable being within the preset upper and lower limits of the control variable. The expression is:

[0076] s min ≤s(t)≤s max

[0077] u min ≤u(t)≤u max

[0078] Where s min Indicates the lower limit of the state variable; s max Indicates the upper limit of the state variable; u min Indicates the lower limit of the control variable; u max Indicates the upper limit of the control variable;

[0079] In summary, through rolling optimization, this method enables the control system to adjust the control strategy in real time, ensuring that the difference between the output result and the set value is minimized, avoiding overly frequent or drastic control actions, and improving the stability of the system; at the same time, it ensures that the state variables and control variables remain within a reasonable range, avoiding exceeding the actual limitations of the physical system; the dynamic behavior of the system is captured by the neural state space model, and even if the parameters vary in time, the future state variables can be better predicted. By predicting the state variables and calculating the result variables, it can better cope with the impact of external disturbances.

[0080] It also includes step S400: controlling the compressed air energy storage system based on the control strategy, and dynamically adjusting the control strategy according to the prediction error through a closed-loop feedback mechanism; due to the influence of the time-varying parameters of the compressed air energy storage system, as the system operating time changes and external disturbances occur, the objective function used in the optimal module changes, and the system error will gradually accumulate. The commonly used PID control cannot solve this type of problem, and a feedback module is added to perform rolling time domain optimization to solve this problem. The feedback module determines whether to perform feedback correction based on the relationship between the prediction error and the threshold. If a correction is made, the optimal control value will continue to be solved. If no correction is made, the control variable value of the current control time domain is output to suppress the impact of external disturbances on the system.

[0081] Specifically, according to the optimal control sequence obtained in step S300, the compressed air energy storage system is controlled, the control variables in the optimal control sequence are applied to each control device in the system, the state variables and result variables of the system are obtained in real time through sensors, and the prediction error between the actually measured result variables and the predicted result variables is compared, and the prediction error is compared with a preset threshold; if the prediction error exceeds the threshold, it is considered that correction is required. If correction is required, a new optimal control sequence is continued to be solved; if no correction is required, the control variable value of the current control time domain is output to suppress the influence of external disturbances on the system; this embodiment re-solves the control sequence in each time step to adapt to the time variability of system parameters and external disturbances, and continuously adjusts the control strategy according to the latest predicted values ​​and actual measured values ​​to ensure stable operation of the system. Through the implementation of step S400, real-time control based on the optimal control strategy can be achieved in the compressed air energy storage system, and the control strategy can be dynamically adjusted according to the prediction error through the closed-loop feedback mechanism. This method not only improves the stability and control accuracy of the system, but also enhances the adaptability and robustness of the system, making the system more reliable and efficient in practical applications. Through rolling time domain optimization and closed-loop feedback mechanism, the system can better cope with parameter time-varying and external disturbances, thereby achieving efficient and stable operation.

[0082] Below, taking the control process of the energy storage process as an example, the energy storage control algorithm flow is as follows: Figure 3 As shown:

[0083] First, preset the system parameters: Preset parameters: Gas storage inlet temperature setting value: Set the target value of the gas storage inlet temperature.

[0084] Upper and lower limits of heat exchanger cold flow velocity: set the allowable range of heat exchanger cold flow velocity; Upper and lower limits of gas storage inlet valve opening: set the allowable range of gas storage inlet valve opening; Prediction time domain: set the length of the prediction time domain, that is, how many time steps are predicted in the future; Control time domain: set the length of the control time domain, that is, how many time steps are considered in one control cycle;

[0085] Then obtain the data of the previous prediction time domain: Actual value of the gas storage inlet temperature: obtain the actual measured value of the gas storage inlet temperature in the previous prediction time domain from the sensor; Actual value of the heat exchanger cold flow velocity: obtain the actual measured value of the heat exchanger cold flow velocity in the previous prediction time domain from the sensor; Actual value of the gas storage inlet valve opening: obtain the actual measured value of the gas storage inlet valve opening in the previous prediction time domain from the sensor; Actual value of the first-stage compressor inlet temperature: obtain the actual measured value of the first-stage compressor inlet temperature from the sensor; Actual value of the heat exchanger cold flow temperature: obtain the actual measured value of the heat exchanger cold flow temperature from the sensor;

[0086] Finally, the acquired data is input into the neural model predictive controller, and the neural state space model (such as MLP network) is used for prediction. The predicted value of the gas storage inlet temperature in the current prediction time domain is calculated through the forward propagation of the neural network; according to the optimization objective function, the optimal control sequence in the current control time domain is cyclically solved, and the control variables in the solved optimal control sequence are input into the heat exchanger oil pump controller and the gas storage valve controller, thereby changing the cold flow temperature and air flow in the system; finally, the difference between the actual value and the set value of the gas storage inlet temperature in the current control time domain is calculated, and the actual value of the gas storage inlet temperature is obtained in real time through the sensor. The difference threshold between the actual value and the set value is calculated: if the difference is less than the preset threshold, the operation continues; if the difference is greater than the threshold, the correction step is entered.

[0087] See also Figure 2 The second technical solution is a compressed air energy storage system control system that takes into account the time-varying parameters, characterized in that the system includes:

[0088] a data acquisition module configured to use the compressed air energy storage system simulation model to collect operation data of the simulation model at each stage and perform preprocessing;

[0089] Prediction module, configured to output variable prediction values ​​based on preprocessed simulation model running data

[0090] A control module is configured to output a control strategy based on the variable prediction value and with the goal of minimizing the deviation of the variable prediction value and the change of the control signal;

[0091] The feedback module is configured to control the compressed air energy storage system based on the control strategy and dynamically adjust the control strategy according to the prediction error through a closed-loop feedback mechanism.

[0092] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0093] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A method for controlling a compressed air energy storage system considering time-varying parameters, characterized in that: The method comprises: Construct a simulation model of the compressed air energy storage system, collect and pre-process the operation data of the simulation model at each stage; Construct a neural state space model and output variable prediction values ​​based on the preprocessed simulation model running data; A control model for a compressed air energy storage system is established, and a control strategy is output based on variable prediction values ​​with the goal of minimizing the deviation of variable prediction values ​​and the change in control signals.

2. The method for controlling a compressed air energy storage system taking into account the time-varying nature of parameters according to claim 1, characterized in that: Collecting simulation model operation data at each stage and preprocessing it includes: The operation data of the simulation model at each stage are divided into state variables, disturbance variables, control variables and result variables; the state variables include at least the heat exchanger outlet temperature, the gas storage reservoir temperature and the oil storage tank temperature; the disturbance variables include at least the compressor inlet temperature and the heat exchanger cold flow temperature; the control variables include at least the heat exchanger cold flow velocity and the opening degree of the gas storage reservoir inlet valve.

3. The method for controlling a compressed air energy storage system taking into account the time-varying nature of parameters according to claim 2, characterized in that: Collecting simulation model operation data at each stage and preprocessing it also includes: Define the state equation, which is expressed as: In the formula, represents the rate of change of state variables at time t; Q represents the state transfer matrix; s(t) represents the state variables at time t; W represents the control influence matrix; u(t) represents the control variables at time t; E represents the disturbance influence matrix; d(t) represents the disturbance variable at time t; Define the output equation, which is expressed as: y(t)=Rx(t) In the formula, y(t) represents the result variable at time t; R represents the mapping matrix from state variables to result variables; x(t) represents the state variable at time t.

4. The method for controlling a compressed air energy storage system taking into account the time-varying nature of parameters according to claim 3, characterized in that: Construct a neural state space model and run the simulation model according to the preprocessed data to output variable prediction values ​​including: Input the state variables, disturbance variables and control variables of the previous period, and use the neural state space model to predict the state variables of the next period; Output the outcome variable for the next period based on the predicted state variable.

5. The method for controlling a compressed air energy storage system taking into account the time-varying nature of parameters according to claim 4, characterized in that: The state variables predicted using the neural state space model for the next period include: Input the preprocessed simulation model running data into the first layer of the MLP network; According to the preset time step, the hidden layer output is calculated as the input of the next time step; For each time step, the weights and biases of the current time step are linearly combined with the hidden layer output of the previous time step; Use the hyperbolic tangent function as the activation function to introduce nonlinear features; The output of the last layer of the MLP network is used as the predicted value of the state variable.

6. The method for controlling a compressed air energy storage system taking into account the time-varying nature of parameters according to any one of claims 1 to 5, characterized in that: The expression of the objective function is: Where, U represents the control time domain; u i represents the control variable value of the i-th control device in the current control time domain; y i represents the predicted value of the state variable of the i-th control device; r i represents the preset value of the state variable of the i-th control device; λ represents the adjustment parameter; k represents the k-th time node; N represents the length of the control time domain.

7. The method for controlling a compressed air energy storage system taking into account the time-varying nature of parameters according to claim 6, characterized in that: The method also includes: controlling the compressed air energy storage system based on a control strategy, and dynamically adjusting the control strategy according to the prediction error through a closed-loop feedback mechanism.

8. The method for controlling a compressed air energy storage system taking into account the time-varying nature of parameters according to claim 6, characterized in that: Including a first constraint condition, the first constraint condition is positioned as the state variable being within a preset state variable upper and lower limit range.

9. The method for controlling a compressed air energy storage system taking into account the time-varying nature of parameters according to claim 8, characterized in that: A second constraint condition is also included, and the second constraint condition is defined as that the control variable is within a preset upper and lower limit range of the control variable.

10. A compressed air energy storage system control system considering time-varying parameters, characterized in that: The system comprises: A data acquisition module is configured to use the compressed air energy storage system simulation model to collect the simulation model operation data of each stage and perform preprocessing; Prediction module, configured to output variable prediction values ​​based on preprocessed simulation model run data A control module configured to output a control strategy based on a variable prediction value and with the goal of minimizing a deviation of the variable prediction value and a change in a control signal; The feedback module is configured to control the compressed air energy storage system based on the control strategy, and dynamically adjust the control strategy according to the prediction error through a closed-loop feedback mechanism.

Images