Gas path pressure control method and system for liquid sampling device

Abstract

The invention provides a gas path pressure control method and system for a liquid sampling device, and the method comprises the steps: obtaining the real-time pressure data of a gas path system, and the environment temperature, gas flow and valve opening data, and constructing an input data set of gas path pressure control; according to the input data set, a BP neural network is adopted to establish a nonlinear mapping relation model between the gas path pressure and all the influence factors; and in the training process of the BP neural network model, an Adam optimization algorithm is adopted to carry out adaptive adjustment on the weight and bias parameters of the BP neural network model, and the trained BP neural network model is obtained through the optimization objective function of the minimum prediction error.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method and system for controlling gas path pressure of a liquid sampling device. Background Art

[0002] In liquid injection devices, precise control of gas line pressure is critical to injection accuracy and repeatability.

[0003] At present, conventional PID control algorithms are difficult to achieve high-precision pressure control when faced with complex gas path pressure changes.

[0004] How to use advanced machine learning algorithms, such as the Adam optimization algorithm, to adaptively adjust the parameters of the BP neural network to achieve precise control of the gas path pressure is a technical problem that needs to be solved urgently. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for controlling the gas circuit pressure of a liquid injection device in order to solve the problem of accurate control of the gas circuit pressure.

[0006] The technical solution of the present invention is:

[0007] The present invention provides a method for controlling the gas path pressure of a liquid injection device, the method comprising:

[0008] Obtain the real-time pressure data, ambient temperature, gas flow rate, and valve opening data of the gas circuit system to build the input data set for gas circuit pressure control;

[0009] According to the input data set, a BP neural network is used to establish a nonlinear mapping relationship model between the gas circuit pressure and various influencing factors;

[0010] In the training process of the BP neural network model, an optimization algorithm is used to adaptively adjust the weight and bias parameters of the BP neural network model, and a trained BP neural network model is obtained by minimizing the optimization objective function of the prediction error;

[0011] The gas circuit pressure data and influencing factor data collected in real time are input into the BP neural network model. According to the predicted value of the gas circuit pressure output by the trained BP neural network model, combined with the control algorithm, the valve opening and gas flow parameters of the gas circuit system are dynamically adjusted to achieve precise control of the gas circuit pressure.

[0012] Furthermore, the acquisition of real-time pressure data and ambient temperature, gas flow rate, and valve opening data of the gas circuit system includes:

[0013] Obtain the real-time pressure, ambient temperature, gas flow, valve opening and other data of the gas system, and pre-process the acquired data, including denoising and normalization operations, to obtain a standardized data set;

[0014] According to the physical characteristics and control objectives of the gas circuit system, a mathematical model of gas circuit pressure control is established, and the input and output variables and control parameters of the model are determined.

[0015] Furthermore, the BP neural network is used to establish a nonlinear mapping relationship model between the gas path pressure and various influencing factors, including:

[0016] Obtain real-time data of gas circuit pressure and influencing factors, and pre-process the data, including data cleaning and normalization operations, to obtain a standardized training data set;

[0017] According to the nonlinear mapping relationship between gas path pressure and influencing factors, the network structure of the BP neural network model is designed, including the number of nodes in the input layer, hidden layer and output layer, as well as the selection of activation function;

[0018] The BP neural network model is trained using the preprocessed training data set, and the network weights and thresholds are adjusted through the back propagation algorithm to minimize the error between the model output and the actual gas path pressure value.

[0019] Furthermore, during the training process of the BP neural network model, a gradient descent optimization algorithm is used to adaptively adjust the weight and bias parameters of the BP neural network by minimizing the optimization objective function;

[0020] The optimization objective function is used to measure the training effect and convergence degree of the model;

[0021] Determine whether the preset training termination condition is met. If not, continue to iterate the model training. If the training termination condition is met, use the current weight and bias parameters as the final model parameters to obtain the trained BP neural network model.

[0022] Furthermore, the step of inputting the real-time collected gas path pressure data and influencing factor data into the BP neural network model includes:

[0023] Acquire the real-time gas path pressure data of the gas path system and the data of related factors affecting the gas path pressure, and input the acquired data into the pre-trained BP neural network model;

[0024] According to the gas circuit pressure prediction value output by the BP neural network model, it is judged whether the current gas circuit pressure is within the preset target pressure range. If it is not within the target range, the gas circuit pressure control process is triggered.

[0025] Furthermore, the combined control algorithm dynamically adjusts the valve opening and gas flow parameters of the gas circuit system, including:

[0026] The PID control algorithm is used to calculate the control parameters for adjusting the valve opening of the gas circuit system and the gas flow rate according to the deviation between the predicted value of the gas circuit pressure and the target pressure;

[0027] The calculated valve opening control parameter is transmitted to the valve actuator of the gas circuit system, and the gas flow rate is changed by adjusting the valve opening, thereby affecting the gas circuit pressure;

[0028] The calculated gas flow control parameters are transmitted to the gas flow controller of the gas circuit system, and the gas circuit pressure is affected by adjusting the gas flow.

[0029] Furthermore, after adjusting the valve opening and gas flow, real-time gas circuit pressure data is continuously obtained and input into the BP neural network model to determine whether the adjusted gas circuit pressure has reached the target pressure range;

[0030] If the adjusted gas circuit pressure is within the target range, the current gas circuit pressure control is completed. If it has not reached the target range, the gas circuit pressure control process is continued until the gas circuit pressure is stabilized within the target range.

[0031] A gas circuit pressure control system, comprising:

[0032] The acquisition module is used to obtain the real-time pressure data of the gas system and the ambient temperature, gas flow rate, and valve opening data to build the input data set for gas pressure control;

[0033] A BP neural network model is used to establish a nonlinear mapping relationship model between the gas path pressure and various influencing factors using a BP neural network according to an input data set; during the training process of the BP neural network model, an optimization algorithm is used to adaptively adjust the weight and bias parameters of the BP neural network model, and a trained BP neural network model is obtained by minimizing the optimization objective function of the prediction error;

[0034] The dynamic adjustment module is used to input the real-time collected gas circuit pressure data and influencing factor data into the BP neural network model, and dynamically adjust the valve opening and gas flow parameters of the gas circuit system according to the gas circuit pressure prediction value output by the trained BP neural network model in combination with the control algorithm to achieve precise control of the gas circuit pressure.

[0035] A gas circuit pressure control device, comprising:

[0036] A memory storing executable instructions;

[0037] A processor runs the executable instructions in the memory to implement the gas line pressure control method of the liquid injection device.

[0038] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the gas path pressure control method of a liquid injection device is implemented.

[0039] Beneficial effects of the present invention:

[0040] The present invention discloses a method for precise control of gas circuit pressure. The method acquires the real-time pressure, temperature, flow and valve opening data of the gas circuit system, constructs an input data set, and uses a BP neural network to establish a nonlinear mapping relationship model between the gas circuit pressure and various influencing factors. During the BP neural network training process, an optimization algorithm is used to adaptively adjust the weight and bias parameters to minimize the prediction error. The real-time data is input into the trained BP neural network model, and the valve opening and gas flow are dynamically adjusted according to the output pressure prediction value in combination with the control algorithm to achieve precise control of the gas circuit pressure. The present invention effectively improves the accuracy and stability of gas circuit pressure control by establishing a nonlinear mapping model between gas circuit pressure and influencing factors, and dynamically adjusting in combination with real-time data, and is suitable for gas circuit systems with high pressure control accuracy requirements.

[0041] Other features and advantages of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present invention.

[0043] Figure 1 A flow chart of a method for controlling gas path pressure of a liquid injection device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0044] The preferred embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the preferred embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0045] like Figure 1 The gas path pressure control method of the liquid injection device of this embodiment may specifically include:

[0046] Step S101, obtain the real-time pressure data and ambient temperature, gas flow, and valve opening data of the gas circuit system, and construct an input data set for gas circuit pressure control, specifically:

[0047] Acquire the real-time pressure, ambient temperature, gas flow, valve opening and other data of the gas system, and pre-process the acquired data, including denoising, normalization and other operations, to obtain a standardized data set; establish a mathematical model of gas pressure control based on the physical characteristics and control objectives of the gas system, and determine the input and output variables and control parameters of the model;

[0048] Step S102, based on the acquired real-time pressure data and influencing factor data, a BP neural network is used to establish a nonlinear mapping relationship model between the gas path pressure and each influencing factor, specifically:

[0049] The real-time data of gas circuit pressure and influencing factors are obtained, and the data are preprocessed, including data cleaning, normalization and other operations, to obtain a standardized training data set. According to the nonlinear mapping relationship between gas circuit pressure and influencing factors, the network structure of the BP neural network model is designed, including the number of nodes in the input layer, hidden layer and output layer, and the selection of activation function. The BP neural network model is trained using the preprocessed training data set, and the network weights and thresholds are adjusted through the back propagation algorithm to minimize the error between the model output and the actual gas circuit pressure value.

[0050] Furthermore, during the training process of the BP neural network model, an optimization algorithm is used to adaptively adjust the weight and bias parameters of the BP neural network model, and the trained BP neural network model is obtained by minimizing the optimization objective function of the prediction error.

[0051] During specific training, according to the pre-established BP neural network model structure, the initialized weights and bias parameters are obtained as the starting point of model training; for the training data set, the forward propagation algorithm is used to calculate the predicted output of the BP neural network, and compared with the true label to obtain the prediction error; according to the prediction error, an optimization objective function is constructed to measure the training effect and convergence degree of the model; by using optimization algorithms such as gradient descent, the weights and bias parameters of the BP neural network are adaptively adjusted by minimizing the optimization objective function; it is determined whether the preset training termination condition is met, if not, the second step is returned to continue the model training iteration; if the training termination condition is met, the current weights and bias parameters are used as the final model parameters to obtain the trained BP neural network model; for the test data set, the trained BP neural network model is used for prediction, and the generalization performance and actual application effect of the model are evaluated.

[0052] Step S103, input the gas circuit pressure data and influencing factor data collected in real time into the BP neural network model, and dynamically adjust the valve opening and gas flow of the gas circuit system according to the gas circuit pressure prediction value output by the trained BP neural network model in combination with the control algorithm to achieve precise control of the gas circuit pressure.

[0053] The real-time gas circuit pressure data of the gas circuit system and the data of related factors affecting the gas circuit pressure are obtained, and the obtained data are input into the pre-trained BP neural network model; according to the gas circuit pressure prediction value output by the BP neural network model, it is judged whether the current gas circuit pressure is within the preset target pressure range, and if it is not within the target range, the gas circuit pressure control process is triggered; the PID control algorithm is used to calculate the control parameters for adjusting the valve opening and gas flow of the gas circuit system according to the deviation between the gas circuit pressure prediction value and the target pressure; the calculated valve opening control parameters are transmitted to the valve actuator of the gas circuit system, and the gas flow is changed by adjusting the valve opening, thereby affecting the gas circuit pressure; the calculated gas flow control parameters are transmitted to the gas flow controller of the gas circuit system, and the gas pressure is affected by adjusting the gas flow; after adjusting the valve opening and gas flow, the real-time gas circuit pressure data is continuously obtained and input into the BP neural network model to judge whether the adjusted gas circuit pressure has reached the target pressure range; if the adjusted gas circuit pressure is within the target range, the current gas circuit pressure control is completed, and if it has not reached the target range, the gas circuit pressure control process is returned to continue to be executed until the gas circuit pressure is stabilized within the target range.

[0054] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A method for controlling gas line pressure of a liquid injection device, characterized in that: The method includes: Obtain the real-time pressure data, ambient temperature, gas flow rate, and valve opening data of the gas circuit system to build the input data set for gas circuit pressure control; According to the input data set, a BP neural network is used to establish a nonlinear mapping relationship model between the gas circuit pressure and various influencing factors; In the training process of the BP neural network model, an optimization algorithm is used to adaptively adjust the weight and bias parameters of the BP neural network model, and a trained BP neural network model is obtained by minimizing the optimization objective function of the prediction error; The gas circuit pressure data and influencing factor data collected in real time are input into the BP neural network model. According to the predicted value of the gas circuit pressure output by the trained BP neural network model, combined with the control algorithm, the valve opening and gas flow parameters of the gas circuit system are dynamically adjusted to achieve precise control of the gas circuit pressure.

2. The method for controlling the gas line pressure of a liquid injection device according to claim 1, characterized in that: The acquisition of real-time pressure data and ambient temperature, gas flow rate, and valve opening data of the gas circuit system includes: Obtain the real-time pressure, ambient temperature, gas flow, valve opening and other data of the gas system, and pre-process the acquired data, including denoising and normalization operations, to obtain a standardized data set; According to the physical characteristics and control objectives of the gas circuit system, a mathematical model of gas circuit pressure control is established, and the input and output variables and control parameters of the model are determined.

3. The method for controlling the gas circuit pressure of a liquid injection device according to claim 1, characterized in that: The BP neural network is used to establish a nonlinear mapping relationship model between the gas path pressure and various influencing factors, including: Obtain real-time data of gas circuit pressure and influencing factors, and pre-process the data, including data cleaning and normalization operations, to obtain a standardized training data set; According to the nonlinear mapping relationship between gas path pressure and influencing factors, the network structure of the BP neural network model is designed, including the number of nodes in the input layer, hidden layer and output layer, as well as the selection of activation function; The BP neural network model is trained using the preprocessed training data set, and the network weights and thresholds are adjusted through the back propagation algorithm to minimize the error between the model output and the actual gas path pressure value.

4. The method for controlling the gas line pressure of a liquid injection device according to claim 3, characterized in that: During the training process of the BP neural network model, a gradient descent optimization algorithm is used to adaptively adjust the weight and bias parameters of the BP neural network by minimizing the optimization objective function; The optimization objective function is used to measure the training effect and convergence degree of the model; Determine whether the preset training termination condition is met. If not, continue to iterate the model training. If the training termination condition is met, use the current weight and bias parameters as the final model parameters to obtain the trained BP neural network model.

5. The method for controlling the gas line pressure of a liquid injection device according to claim 1, characterized in that: The step of inputting the real-time collected gas circuit pressure data and influencing factor data into the BP neural network model comprises: Acquire the real-time gas path pressure data of the gas path system and the data of related factors affecting the gas path pressure, and input the acquired data into the pre-trained BP neural network model; According to the gas circuit pressure prediction value output by the BP neural network model, it is judged whether the current gas circuit pressure is within the preset target pressure range. If it is not within the target range, the gas circuit pressure control process is triggered.

6. The method for controlling the gas line pressure of a liquid injection device according to claim 5, characterized in that: The combined control algorithm dynamically adjusts the valve opening and gas flow parameters of the gas circuit system, including: The PID control algorithm is used to calculate the control parameters for adjusting the valve opening of the gas circuit system and the gas flow rate according to the deviation between the predicted value of the gas circuit pressure and the target pressure; The calculated valve opening control parameter is transmitted to the valve actuator of the gas circuit system, and the gas flow rate is changed by adjusting the valve opening, thereby affecting the gas circuit pressure; The calculated gas flow control parameters are transmitted to the gas flow controller of the gas circuit system, and the gas circuit pressure is affected by adjusting the gas flow.

7. The method for controlling the gas path pressure of a liquid injection device according to claim 6, characterized in that: After adjusting the valve opening and gas flow, continue to obtain real-time gas circuit pressure data and input it into the BP neural network model to determine whether the adjusted gas circuit pressure has reached the target pressure range; If the adjusted gas circuit pressure is within the target range, the current gas circuit pressure control is completed. If it has not reached the target range, the gas circuit pressure control process is continued until the gas circuit pressure is stabilized within the target range.

8. A gas circuit pressure control system for a liquid sampling device, characterized in that: include: The acquisition module is used to obtain the real-time pressure data of the gas system and the ambient temperature, gas flow rate, and valve opening data to build the input data set for gas pressure control; A BP neural network model is used to establish a nonlinear mapping relationship model between the gas path pressure and various influencing factors using a BP neural network according to an input data set; In the training process of the BP neural network model, an optimization algorithm is used to adaptively adjust the weight and bias parameters of the BP neural network model, and a trained BP neural network model is obtained by minimizing the optimization objective function of the prediction error; The dynamic adjustment module is used to input the real-time collected gas circuit pressure data and influencing factor data into the BP neural network model, and dynamically adjust the valve opening and gas flow parameters of the gas circuit system according to the gas circuit pressure prediction value output by the trained BP neural network model in combination with the control algorithm to achieve precise control of the gas circuit pressure.

9. A gas line pressure control device for a liquid sampling device, characterized in that: include: A memory storing executable instructions; A processor, wherein the processor runs the executable instructions in the memory to implement the gas path pressure control method of the liquid sampling device of any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for controlling the gas path pressure of a liquid sampling device of any one of claims 1 to 7 is implemented.

Images