Intelligent control method and system for accumulator pressure regulation

Abstract

The application discloses an accumulator pressure regulation intelligent control method and system, comprising collecting historical operation data; extracting time sequence characteristics of the data and calculating characteristic data; constructing a physical information neural network and constructing a physical constraint loss function training model; real-time data acquisition and preprocessing, acquiring data in real time through a sensor group; predicting based on the trained physical information neural network to obtain a predicted future pressure sequence; defining a dynamic weight multi-objective optimization objective function and solving by using a numerical optimization algorithm to obtain an optimal control sequence. The accumulator pressure regulation intelligent control method and system ensure that the predicted pressure sequence result conforms to the thermodynamic law, and at the same time, the target function is defined by combining the pressure tracking error, the energy consumption variable and the valve wear variable, the predicted pressure sequence output by the physical information neural network is optimized, the timeliness and accuracy are greatly improved, and the production quality is improved.

Description

Technical Field

[0001] This invention belongs to the field of accumulator pressure regulation, and specifically relates to an intelligent control method and system for accumulator pressure regulation. Background Technology

[0002] In hydraulic systems, accumulators mitigate pressure fluctuations by storing and releasing energy, thereby ensuring stable system operation. Accumulators absorb excess energy in the hydraulic system and release it when needed to maintain pressure balance. This energy management is crucial for improving the efficiency and reliability of hydraulic systems, especially in applications with frequent load changes or requiring rapid response.

[0003] However, traditional accumulator pressure regulation methods have some significant limitations. First, these methods often rely on manual monitoring and adjustment, which not only increases operational complexity but can also lead to response delays because manual operation cannot match the response speed of automated systems. Second, traditional methods have slow response times, which is particularly noticeable when facing rapidly changing load conditions, potentially causing the system to be unable to adjust pressure in a timely manner, thus affecting performance. Furthermore, due to the lack of flexibility in regulation, traditional methods are also limited in their efficiency in energy recovery and reuse, which not only wastes energy but may also increase system operating costs.

[0004] Therefore, the above problems urgently need to be solved. Summary of the Invention

[0005] Purpose of the invention: To overcome the above shortcomings, the purpose of this invention is to provide an intelligent control method and system for accumulator pressure regulation. By constructing a physical constraint loss function to train a physical information neural network, the predicted pressure sequence results are ensured to conform to thermodynamic laws and actual production. At the same time, by combining pressure tracking error, energy consumption variables, and valve wear variables to define an objective function, the predicted pressure sequence output by the physical information neural network is optimized, greatly improving timeliness and accuracy, and enhancing production quality.

[0006] Technical Solution: To achieve the above objectives, this invention provides an intelligent control method for accumulator pressure regulation, comprising: S1): Collect historical operational data and preprocess the collected data; S2): Extract the time-series features of the data to obtain the original time-series data, calculate the theoretical pressure deviation, and fuse it with the original time-series data to obtain the feature data; by extracting the time-series features of the data and combining them with the theoretical pressure deviation, the dynamic characteristics of the system can be better captured.

[0007] Furthermore, it also includes S3): constructing a physical information neural network and constructing a physical constraint loss function to train the model; using feature data as input to complete the training of the physical information neural network; the model does not rely solely on data for training, but can also utilize physical knowledge to achieve better performance.

[0008] Furthermore, S301): The physical constraint loss function embeds the theoretical pressure deviation into the loss function to ensure that the prediction results conform to thermodynamic laws; by using the physical constraint loss function to ensure that the prediction results conform to thermodynamic laws, the prediction accuracy of the model is improved.

[0009] Furthermore, it also includes S4): real-time data acquisition and preprocessing, which acquires data in real time through sensor array monitoring and preprocesses the collected data; through real-time data acquisition and preprocessing, the current status of the system can be obtained in a timely manner.

[0010] Furthermore, it also includes S5): calling S2 to obtain the theoretical pressure deviation, and making predictions based on the trained physical information neural network to obtain the predicted future pressure sequence; predicting the future pressure sequence based on the trained model to provide support for real-time control.

[0011] Furthermore, it also includes S6): defining a dynamic weighted multi-objective optimization objective function and solving it using a numerical optimization algorithm to obtain the optimal control sequence; defining a dynamic weighted multi-objective optimization objective function can dynamically adjust the optimization objective according to the real-time state, thereby realizing a more flexible control strategy; S601): Analyze sensor data in real time and obtain the target pressure sequence based on a static rule base; S602): Calculate the mean square error between the predicted pressure sequence and the target pressure sequence to obtain the pressure tracking error; S603): Set energy consumption variables and valve wear variables, and define the objective function by combining pressure tracking error, energy consumption variables, and valve wear variables; S604): A numerical optimization algorithm is used to solve the problem and obtain the optimal control sequence. During the optimization process, not only pressure tracking error is considered, but also energy consumption and valve wear variables are taken into account, achieving multi-objective optimization. This ensures that the control strategy not only meets the requirements of pressure control but also achieves a balance between energy consumption and equipment lifespan; it finds the optimal trade-off point among multiple objectives, thereby improving overall performance, reducing operating costs, and extending service life.

[0012] Furthermore, the physical constraint loss function Loss formula mentioned in S301 is as follows: Among them, y pred For the model's predicted future stress sequence; ytrue For the actual observed future pressure sequence; p pred V represents the pressure value predicted by the model at the current moment. actual Let n be the actual volume of the gas chamber; nR be the gas constant, where n is the amount of substance of the gas and R is the universal gas constant; T actual λ represents the gas temperature measured by the sensor; λ is the weighting coefficient of the physical constraint term. By incorporating the ideal gas law, a physical law, as a constraint term into the loss function, we can ensure that the model predictions not only conform to the data but also to known physical laws, thereby improving the physical consistency of the model. At the same time, since the model depends not only on the data but is also guided by physical laws, the prediction accuracy and robustness of the model are improved.

[0013] Furthermore, the formula for the pressure tracking error J1 in S602 is as follows: Where, p pred,k p represents the pressure value predicted by the model at step k; target,k Let J be the target pressure value at step k. The pressure tracking error J1 provides a quantitative metric to assess how close the model-predicted pressure is to the target pressure. By calculating the pressure tracking error J1 in real time, prediction errors can be quickly identified.

[0014] Furthermore, the formulas for the energy consumption variable J2 and the valve wear variable J3 in S603 are as follows: Among them, u k The valve opening at step k is obtained through data acquisition; The objective function J is formulated as follows: Here, w1, w2, and w3 represent the weights of the pressure tracking error J1, energy consumption variable J2, and valve wear variable J3, respectively. By introducing the energy consumption variable J2 and the valve wear variable J3, factors related to both the energy utilization efficiency of the hydraulic system and the service life of the equipment can be incorporated into the objective function J, improving the globality of the optimization. Simultaneously, the weights w1, w2, and w3 allow decision-makers to customize the balance between performance (such as pressure tracking accuracy) and cost (such as energy consumption and equipment wear) according to specific circumstances, ensuring the flexibility of the control method.

[0015] Furthermore, the formula for the theoretical pressure deviation Δp in S2 is as follows: Where, p actual The actual pressure measured by the sensor; T actualV represents the actual temperature measured by the sensor. actual This is the gas chamber volume inferred from the liquid volume change. Calculating the theoretical pressure deviation can help predict future pressure change trends in the system, providing a basis for model decision-making.

[0016] Furthermore, the physical information neural network in S3 includes an input layer, two LSTM layers, and an output layer; the input layer, two LSTM layers, and output layer are connected sequentially. The LSTM contains one or more memory units, which can store long-term information and access this information when needed, effectively processing time-series features and improving accuracy.

[0017] Furthermore, S5 also includes setting a threshold; if the theoretical pressure deviation exceeds the threshold, an alarm is triggered. By setting the threshold, pressure deviation can be monitored in real time, ensuring that the pressure is always within a safe and effective range, reducing unexpected downtime caused by equipment failure, and guaranteeing the continuity and stability of production.

[0018] This invention also provides an intelligent control system for accumulator pressure regulation, used to implement the aforementioned intelligent control method for accumulator pressure regulation. The system includes: a data collection module, a pressure prediction module, and an optimization control module. The data collection module is connected to the pressure prediction module, transmitting real-time and historical data to it. The data collection module is also connected to the optimization control module, transmitting real-time data to it. The pressure prediction module is connected to the optimization control module, transmitting pressure prediction results to it. The data collection module focuses on data acquisition and preprocessing, ensuring data quality and usability. The pressure prediction module focuses on data-driven modeling and prediction, using historical and real-time data for accurate pressure prediction. The optimization control module focuses on making optimization decisions based on prediction results and real-time data, generating the optimal control strategy. The three modules have clearly defined functions, allowing each module to be developed and optimized independently, simplifying the system's complexity.

[0019] As can be seen from the above technical solution, the present invention has the following beneficial effects: 1. The present invention provides an intelligent control method and system for regulating the pressure of an accumulator. By constructing a physical constraint loss function to train a physical information neural network, it ensures that the predicted pressure sequence results conform to thermodynamic laws and actual production. 2. The present invention provides an intelligent control method and system for accumulator pressure regulation. By combining pressure tracking error, energy consumption variables and valve wear variables to define an objective function, the predicted pressure sequence output by the physical information neural network is optimized, which greatly improves timeliness and accuracy, and enhances production quality.

[0020] 3. An intelligent control method and system for accumulator pressure regulation is invented. The optimization algorithm based on data and models can provide operators with scientific decision-making basis and can quickly provide the optimal control strategy even in complex situations, thereby improving the system's operating efficiency. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating the steps of an intelligent control method for regulating accumulator pressure according to the present invention. Figure 2 This is an architecture diagram of an intelligent control system for regulating the pressure of an energy storage device according to the present invention. Detailed Implementation

[0022] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention. Example

[0023] In this embodiment, as Figure 1 This invention discloses an intelligent control method for regulating accumulator pressure, comprising: S1): Collect historical operational data and preprocess the collected data; S2): Extract the temporal features of the data to obtain the original time-series data, calculate the theoretical pressure deviation, and fuse it with the original time-series data to obtain the feature data; S3): Construct a physical information neural network and a physical constraint loss function training model; use feature data as input to complete the training of the physical information neural network; S301): The physical constraint loss function embeds the theoretical pressure deviation into the loss function to ensure that the prediction results conform to thermodynamic laws; S4): Real-time data acquisition and preprocessing, which involves real-time monitoring and acquisition of data through a sensor array, and preprocessing the collected data. S5): Call S2 to obtain the theoretical pressure deviation, and make a prediction based on the trained physical information neural network to obtain the predicted future pressure sequence; S6): Define a dynamic weighted multi-objective optimization objective function and solve it using a numerical optimization algorithm to obtain the optimal control sequence; S601): Analyze sensor data in real time and obtain the target pressure sequence based on a static rule base; S602): Calculate the mean square error between the predicted pressure sequence and the target pressure sequence to obtain the pressure tracking error; S603): Set energy consumption variables and valve wear variables, and define the objective function by combining pressure tracking error, energy consumption variables, and valve wear variables; S604): The optimal control sequence is obtained by using a numerical optimization algorithm.

[0024] Specifically, the time-series features in S2 include pressure sequence, temperature sequence, flow rate sequence, and valve opening; the pressure sequence is the sampled value of the past N seconds, such as once per second, for a total historical window of 60 seconds; the temperature sequence is the liquid temperature; the flow rate sequence is the liquid injection rate; and the valve opening is the control signal.

[0025] Specifically, the data preprocessing in S1 and S4 includes data cleaning, sliding window segmentation, and normalization. The data cleaning includes, but is not limited to, removing outliers caused by sensor malfunctions such as a sudden drop in pressure to negative values, and interpolating and filling in missing values. The sliding window segmentation includes setting an input window, an output window, and defining a sliding step size; the input window takes the data from the past 60 seconds as the input feature, starting from a time step t; the output window takes the pressure value from t+1 to t+10 seconds into the next 10 seconds as the label; the sliding step size is 1 second to generate continuous samples; The normalization process performs Min-Max normalization on each feature column.

[0026] Specifically, in S2, data fusion is achieved by splicing the theoretical pressure deviation to the original time series data to obtain feature data.

[0027] In this embodiment, the physical constraint loss function Loss formula mentioned in S301 is as follows: Among them, y pred For the model's predicted future stress sequence; y true For the actual observed future pressure sequence; p pred V represents the pressure value predicted by the model at the current moment. actual Let n be the actual volume of the gas chamber; nR be the gas constant, where n is the amount of substance of the gas and R is the universal gas constant; T actual λ represents the gas temperature measured by the sensor; λ is the weighting coefficient of the physical constraint term.

[0028] Specifically, in the loss function, the theoretical pressure deviation Δp is implicitly embedded; the p predicted by the forced model is also included. pred Satisfy p pred V actual ≈nRT actual To ensure that the prediction results conform to thermodynamic laws; For example, if the model predicts ppred =100 bar, and according to the real-time temperature T actual and volume V actual If the calculated theoretical stress deviates significantly from this value, the physical constraint term in the loss function will penalize this deviation, forcing the model to correct its predictions.

[0029] In this embodiment, the formula for the pressure tracking error J1 in S602 is as follows: Where, p pred,k p represents the pressure value predicted by the model at step k; target,k Let be the target pressure value for the k-th step.

[0030] Specifically, p target,k To obtain the corresponding pressure sequence by analyzing the pressure, temperature and flow data of the sensor in real time and comparing them with the preset target values ​​in the static rule base; the static rule base presets several operating conditions (such as "gear shift", "high load" and "leakage") and corresponding sensor data thresholds.

[0031] In this embodiment, the formulas for the energy consumption variable J2 and the valve wear variable J3 in S603 are as follows: Among them, u k The valve opening at step k is obtained through data acquisition; The objective function J is formulated as follows: Among them, w1, w2 and w3 are the weights of pressure tracking error J1, energy consumption variable J2 and valve wear variable J3, respectively.

[0032] Specifically, the sum of pump power and valve power can be used as the valve opening degree u. k One alternative is the following: the pump power is related to the flow rate and is proportional to the square of the rotational speed; the valve power is related to the valve opening adjustment range, such as frequent large valve movements will increase energy consumption.

[0033] In particular, smooth transitions can be introduced for w1, w2, and w3 as a preferred option to achieve dynamic weight adjustments, increasing computational cost in exchange for higher recommendation accuracy.

[0034] In this embodiment, as Figure 1 The formula for the theoretical pressure deviation Δp in S2 is as follows: Where, p actual The actual pressure measured by the sensor; Tactual V represents the actual temperature measured by the sensor. actual This is the volume of the gas chamber inferred from the change in liquid volume.

[0035] Specifically, the model input includes not only the raw time-series data (pressure, temperature, flow rate, valve opening), but also the theoretical pressure deviation Δp as a key feature. The input dimensions are: [Pressure, temperature, flow rate, valve opening, theoretical pressure deviation].

[0036] In this embodiment, as Figure 1 The physical information neural network in S3 includes an input layer, two LSTM layers, and an output layer; the input layer, two LSTM layers, and output layer are connected in sequence.

[0037] Specifically, the number of neurons in each LSTM layer is set to 64 as an optimal value to ensure computational depth while avoiding excessive computational complexity; the fully connected layer outputs the stress sequence for the next 10 seconds.

[0038] In this embodiment, as Figure 1 S5 also includes setting a threshold, and triggering an alarm if the theoretical pressure deviation exceeds the threshold.

[0039] Specifically, the theoretical pressure deviation Δp is calculated in real time, and a threshold of 2% is set as an optimal value. If the absolute value of the theoretical pressure deviation Δp exceeds the threshold, the following actions are triggered: (1) Alarm: Notify maintenance personnel to check for system leaks or sensor malfunctions; (2) Control rollback: Pause model predictive control and switch to PID control until the deviation is recovered.

[0040] This invention also discloses an intelligent control system for accumulator pressure regulation, used to implement the aforementioned intelligent control method for accumulator pressure regulation. The system is characterized by comprising: a data collection module, a pressure prediction module, and an optimization control module; the data collection module is connected to the pressure prediction module, transmitting real-time and historical data to the pressure prediction module; the data collection module is also connected to the optimization control module, transmitting real-time data to the optimization control module; and the pressure prediction module is connected to the optimization control module, transmitting pressure prediction results to the optimization control module.

[0041] Specifically, the data collection module needs to integrate and deploy sensors such as pressure sensors, temperature sensors, and flow meters to collect data such as pressure, temperature, and flow rate in real time. Data exchange between modules needs to be achieved through standardized interfaces, and detailed API documentation needs to be written for each interface, including the interface name, function, input and output parameters, return values, etc., to ensure data format consistency and transmission reliability. The collected data needs to be stored in a local database for subsequent analysis and use.

[0042] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements can be made without departing from the principle of the present invention, and these improvements should also be considered within the scope of protection of the present invention.

Claims

1. A smart control method for regulating accumulator pressure, characterized in that: include: S1): Collect historical operational data and preprocess the collected data; S2): Extract the temporal features of the data to obtain the original time-series data, calculate the theoretical pressure deviation, and fuse it with the original time-series data to obtain the feature data; S3): Construct a physical information neural network and a physical constraint loss function training model; use feature data as input to complete the training of the physical information neural network; S301): The physical constraint loss function embeds the theoretical pressure deviation into the loss function to ensure that the prediction results conform to thermodynamic laws; The physical constraint loss function Loss formula mentioned in S301 is as follows: Among them, y pred For the model's predicted future stress sequence; y true For the actual observed future pressure sequence; p pred V represents the pressure value predicted by the model at the current moment. actual Let n be the actual volume of the gas chamber; nR be the gas constant, where n is the amount of substance of the gas and R is the universal gas constant; T actual λ represents the gas temperature measured by the sensor; λ is the weighting coefficient of the physical constraint term. S4): Real-time data acquisition and preprocessing, which involves real-time monitoring and acquisition of data through a sensor array, and preprocessing the collected data. S5): Call S2 to obtain the theoretical pressure deviation, and make a prediction based on the trained physical information neural network to obtain the predicted future pressure sequence; S6): Define a dynamic weighted multi-objective optimization objective function and solve it using a numerical optimization algorithm to obtain the optimal control sequence; S601): Analyze sensor data in real time and obtain the target pressure sequence based on a static rule base; S602): Calculate the mean square error between the predicted pressure sequence and the target pressure sequence to obtain the pressure tracking error; S603): Set energy consumption variables and valve wear variables, and define the objective function by combining pressure tracking error, energy consumption variables, and valve wear variables; S604): The optimal control sequence is obtained by using a numerical optimization algorithm.

2. The intelligent control method for regulating accumulator pressure according to claim 1, characterized in that: The formula for the pressure tracking error J1 in S602 is as follows: Where, p pred,k p represents the pressure value predicted by the model at step k; target,k Let be the target pressure value for the k-th step.

3. The intelligent control method for regulating accumulator pressure according to claim 2, characterized in that: The formulas for energy consumption variable J2 and valve wear variable J3 in S603 are as follows: Among them, u k The valve opening at step k is obtained through data acquisition; The objective function J is formulated as follows: Among them, w1, w2 and w3 are the weights of pressure tracking error J1, energy consumption variable J2 and valve wear variable J3, respectively.

4. The intelligent control method for regulating accumulator pressure according to claim 1, characterized in that: The formula for the theoretical pressure deviation Δp in S2 is as follows: Where, p actual The actual pressure measured by the sensor; T actual V represents the actual temperature measured by the sensor. actual This is the volume of the gas chamber inferred from the change in liquid volume.

5. The intelligent control method for regulating accumulator pressure according to claim 1, characterized in that: The physical information neural network in S3 includes an input layer, two LSTM layers, and an output layer; the input layer, two LSTM layers, and output layer are connected sequentially.

6. The intelligent control method for regulating accumulator pressure according to claim 1, characterized in that: The S5 also includes setting a threshold, which triggers an alarm if the theoretical pressure deviation exceeds the threshold.

7. An intelligent control system for regulating accumulator pressure, used to implement the intelligent control method for regulating accumulator pressure as described in any one of claims 1 to 6, characterized in that: include: Data collection module, pressure prediction module, and optimization control module; The data collection module is connected to the pressure prediction module, transmitting real-time and historical data to the pressure prediction module; the data collection module is also connected to the optimization control module, transmitting real-time data to the optimization control module; the pressure prediction module is connected to the optimization control module, transmitting pressure prediction results to the optimization control module.

Images