A method, system, device and medium for regulating the pressure of the main pipe of an air compressor

Abstract

The present invention provides an air compressor header pressure regulation method, system, device and medium. The method includes: using the air compressor header pressure data and the pressure data at the end of each branch pipe as input variables, using the flow rate data at the end of each branch pipe as output variables, and establishing a first prediction model based on the prediction data set in the non-fully measured scenario; obtaining the current prediction data of the output variables based on the first prediction model; using the current flow rate prediction data at the end of each branch pipe output by the first prediction model as input variables, using the air compressor header pressure data as output variables, and establishing a second prediction model based on the prediction data set in the non-fully measured scenario; obtaining the current pressure prediction data of the air compressor header based on the second prediction model; using the current pressure prediction data of the air compressor header output by the second prediction model as the current regulation target value to regulate the current pressure of the air compressor header, effectively improving the reliability of the header pressure control and reducing the energy cost.

Description

Technical Field

[0001] The present invention relates to the field of air compressor regulation, and particularly to a method, system, device and medium for regulating the main pipe pressure of an air compressor. Background Art

[0002] In the air supply process of an air compressor system, there are usually multiple end points (end points of branch pipes). Currently, in the general air supply management scenario of an air compressor, flow meters are installed at all end points. However, in the actual air supply process, due to factors such as high cost or complex installation, it is impossible to configure flow meters at all end points simultaneously. In this case, it is impossible to obtain the flow data of all end points at the same time, and the monitoring and guidance of the air supply process of the air compressor cannot reach the expected level.

[0003] Moreover, currently, the air compressor system generally lacks sufficient regulation ability in the design of the main pipe pressure value. It can only simply set a sufficiently high main pipe pressure threshold (default value) according to experience values, or determine the main pipe pressure threshold according to various actual operating parameters of the air compressor. The influence factors of the flow data of all branch pipe end points on the main pipe pressure control have not been considered yet, resulting in low reliability of the main pipe pressure control and causing problems of energy waste and high cost.

[0004] In view of this problem, the present invention provides a method, system, device and medium for regulating the main pipe pressure of an air compressor to solve the above problems. Summary of the Invention

[0005] In order to solve the problems existing in the prior art, the present invention innovatively proposes a method, system, device and medium for regulating the main pipe pressure of an air compressor, effectively solving the problems of low reliability of the main pipe pressure control caused by the prior art and high energy cost, effectively improving the reliability of the main pipe pressure control and reducing the energy cost.

[0006] In the first aspect of the present invention, a method for regulating the main pipe pressure of an air compressor is provided, including:

[0007] Obtain the main pipe pressure data, main pipe flow data, pressure data of each branch pipe end point, and flow data of each branch pipe end point of the air compressor, and establish a prediction data set under a non-fully measured scenario;

[0008] Taking the main pipe pressure data of the air compressor and the pressure data of each branch pipe end point as input variables, and taking the flow data of each branch pipe end point as output variables, based on the prediction data set under the non-fully measured scenario, establish a first prediction model; input the real-time data of the input variables into the first prediction model to obtain the current prediction data of the output variables;

[0009] Taking the current flow prediction data of each branch pipe end output by the first prediction model as input variables and the air compressor main pipe pressure data as output variables, a second prediction model is established based on the prediction data set in the non-complete measurement scenario; the current flow prediction data of each branch pipe end output by the first prediction model is input into the second prediction model to obtain the current pressure prediction data of the air compressor main pipe;

[0010] Taking the current pressure prediction data of the air compressor main pipe output by the second prediction model as the current control target value, the current pressure of the air compressor main pipe is regulated according to the difference between the currently monitored pressure of the air compressor main pipe and the current pressure prediction data of the air compressor main pipe.

[0011] Optionally, obtaining the air compressor main pipe pressure data, main pipe flow data, pressure data of each branch pipe end, and flow data of each branch pipe end, and establishing a prediction data set in the non-complete measurement scenario specifically includes:

[0012] Install a pressure gauge and a flow meter at the main pipe, and install a pressure gauge at the end of all branch pipes. Only install a flow meter at the end of one branch pipe at the same time;

[0013] Keep all air compressors running normally, obtain the pressure gauge value and flow meter value at the main pipe, and the pressure gauge values at the ends of each branch pipe. Keep the pressure gauge value and flow meter value at the main pipe, and the pressure gauge values at the ends of each branch pipe unchanged, and collect the flow meter value at the end of the currently installed flow meter;

[0014] Only change any one of the pressure gauge value at the main pipe and the pressure gauge values at the ends of each branch pipe each time; collect the flow meter value at the end of the currently installed flow meter again until the flow meter values at the ends of the currently installed flow meter with a preset number of samples are collected;

[0015] Install the flow meter at the end of the next branch pipe, and collect the flow meter values at the ends of the next branch pipe with a preset number of samples until the flow meter values at the ends of all branch pipes with a preset number of samples are collected, obtaining the prediction data set of the data to be trained in the non-complete scenario.

[0016] Furthermore, for the flow meter values at the ends of the branch pipes where the flow meter is not currently installed in the data to be trained in the non-complete scenario, a boolean mask is set to mark the null data or missing value data that is not considered in subsequent training calculations.

[0017] Optionally, when the flow meter value at the end of the branch pipe where the flow meter is currently installed is not less than the flow meter value at the main pipe, keep the current pressure gauge value and flow meter value at the main pipe, and the current pressure gauge values at the ends of each branch pipe unchanged, and re-collect the flow meter value at the end of the currently installed flow meter.

[0018] Further, before establishing a second prediction model based on the prediction dataset in the non-fully measured scenario, with the current flow prediction data of each branch pipe end output by the first prediction model as input variables and the air compressor main pipe pressure data as the output variable, it further includes:

[0019] Determine whether each current flow prediction data of the branch pipe end obtained by inputting the real-time data of the input variable into the first prediction model is less than the current flowmeter value at the main pipe. If the current flow prediction data of a certain branch pipe end is not less than the monitored value of the flowmeter at the current main pipe, then re-input the real-time data of the input variable into the first prediction model to predict the current flow of this branch pipe end, or predict the current flow of this branch pipe end after correcting the structural parameters of the first prediction model until each current flow prediction data of the branch pipe end is less than the current flowmeter value at the main pipe.

[0020] Optionally, the models, configurations, manufacturers, and usage conditions of all air compressors are the same. Optionally, both the first prediction model and the second prediction model are prediction models based on neural networks.

[0021] The second aspect of the present invention provides an air compressor main pipe pressure regulation system, including:

[0022] An acquisition module that acquires air compressor main pipe pressure data, main pipe flow data, pressure data of each branch pipe end, and flow data of each branch pipe end, and establishes a prediction dataset in the non-fully measured scenario;

[0023] A first establishment module that uses the air compressor main pipe pressure data and the pressure data of each branch pipe end as input variables and the flow data of each branch pipe end as output variables, and establishes a first prediction model based on the prediction dataset in the non-fully measured scenario; inputs the real-time data of the input variable into the first prediction model to obtain the current prediction data of the output variable;

[0024] A second establishment module that uses the current flow prediction data of each branch pipe end output by the first prediction model as input variables and the air compressor main pipe pressure data as the output variable, and establishes a second prediction model based on the prediction dataset in the non-fully measured scenario; inputs the current flow prediction data of each branch pipe end output by the first prediction model into the second prediction model to obtain the current pressure prediction data of the air compressor main pipe;

[0025] A regulation module that uses the current pressure prediction data of the air compressor main pipe output by the second prediction model as the current regulation target value, and regulates the current pressure of the air compressor main pipe according to the difference between the current monitored pressure of the air compressor main pipe and the current pressure prediction data of the air compressor main pipe.

[0026] A third aspect of the present invention provides an electronic device, including: a memory for storing a computer program; a processor for implementing the steps of a method for regulating the pressure of the main pipe of an air compressor as described in the first aspect of the present invention when executing the computer program.

[0027] A fourth aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of a method for regulating the pressure of the main pipe of an air compressor as described in the first aspect of the present invention are implemented.

[0028] The technical solution adopted by the present invention includes the following technical effects:

[0029] 1. The technical solution of the present invention uses the main pipe pressure data of the air compressor and the pressure data at the end of each branch pipe as input variables, and the flow data at the end of each branch pipe as output variables. Based on the prediction data set in the non-fully measured scenario, a first prediction model is established; based on the first prediction model, the current prediction data of the output variable is obtained; using the current flow prediction data at the end of each branch pipe output by the first prediction model as input variables and the main pipe pressure data of the air compressor as output variables, based on the prediction data set in the non-fully measured scenario, a second prediction model is established; based on the second prediction model, the current pressure prediction data of the main pipe of the air compressor is obtained; using the current pressure prediction data of the main pipe of the air compressor output by the second prediction model as the current regulation target value, the current pressure of the main pipe of the air compressor is regulated. It can not only achieve comprehensive monitoring without installing flow meters at each end, thereby reducing the overall cost of the system; moreover, based on the flow prediction data at the end of each branch pipe, the current pressure prediction data of the main pipe of the air compressor is obtained, and the current pressure prediction data of the main pipe of the air compressor is used as the current regulation target value (the current main pipe pressure threshold); compared with determining the main pipe pressure threshold according to various operating parameters of the air compressor, the data types involved in this solution are fewer, and it is more convenient to implement, and it can effectively combine the flow monitoring at the end of the branch pipe and the main pipe pressure control; effectively solve the problem of low reliability of main pipe pressure control and high energy cost caused by the prior art, effectively improve the reliability of main pipe pressure control, and reduce the energy cost.

[0030] 2. In the technical solution of the present invention, a pressure gauge and a flow meter are installed at the main pipe, and pressure gauges are installed at the ends of all branch pipes. Only one flow meter is installed at the end of one branch pipe at the same time; the flow meter is successively installed at the end of each branch pipe, and the flow meter value at the end of this branch pipe with a preset number of samples is collected until the flow meter values with a preset number of samples are collected at the ends of all branch pipes, obtaining a prediction data set of the data to be trained in the non-fully measured scenario, reducing the monitoring cost of the flow at the end of the branch pipe.

[0031] 3. In the technical solution of the present invention, when the flowmeter value at the end of the branch pipe where the flowmeter is currently installed in the to-be-trained data under the non-complete scenario is not less than the flowmeter value at the main pipe, the current pressure gauge value and flowmeter value at the main pipe are maintained, and the current pressure gauge value at the end of each branch pipe remains unchanged. The flowmeter value at the end of the branch pipe where the flowmeter is currently installed is re-collected, ensuring the accuracy of the prediction data set under the non-complete scenario, and thus ensuring the accuracy of the data monitoring of the flow at the end of the branch pipe.

[0032] 4. In the technical solution of the present invention, it is judged whether the real-time data of the input variable is input into the first prediction model to obtain the current flow prediction data at the end of each branch pipe and is not less than the current flowmeter value at the main pipe. If the current flow prediction data at the end of a certain branch pipe is not less than the current flowmeter monitoring value at the main pipe, the real-time data of the input variable is re-input into the first prediction model to predict the current flow at the end of this branch pipe, or the structural parameters of the first prediction model are corrected and then the current flow at the end of this branch pipe is predicted until the current flow prediction data at the end of all branch pipes are not less than the current flowmeter value at the main pipe, ensuring the accuracy of the prediction data set under the non-complete scenario, and thus ensuring the accuracy of the data monitoring of the flow at the end of the branch pipe.

[0033] 5. In the technical solution of the present invention, the models, configurations, manufacturers, and usage conditions of all air compressors are the same, further ensuring the accuracy of the prediction data set under the non-complete scenario, and thus further ensuring the accuracy of the data monitoring of the flow at the end of the branch pipe and the pressure control of the main pipe.

[0034] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0036] Figure 1 It is a flowchart of the method in Embodiment 1 of the present invention;

[0037] Figure 2 It is a flowchart of step S1 in the method in Embodiment 1 of the present invention;

[0038] Figure 3 It is a simplified schematic diagram of the air compressor system in the method in Embodiment 1 of the present invention;

[0039] Figure 4 It is another flowchart of the method in Embodiment 1 of the present invention;

[0040] Figure 5 It is a schematic structural diagram of the system in the second embodiment of the solution of the present invention;

[0041] Figure 6 It is a schematic structural diagram of the device in the third embodiment of the solution of the present invention. Detailed implementation manners

[0042] To clearly illustrate the technical features of this solution, the present invention will be elaborated in detail below through specific implementation manners and in combination with its accompanying drawings. The following disclosure provides many different embodiments or examples for implementing different structures of the present invention. To simplify the disclosure of the present invention, the components and settings of specific examples are described below. In addition, the present invention may repeat reference numerals and / or letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed. It should be noted that the components illustrated in the drawings are not necessarily drawn to scale. The present invention omits the description of well-known components and processing technologies and processes to avoid unnecessarily limiting the present invention.

[0043] Embodiment 1

[0044] As Figure 1 shown, the present invention provides a method for regulating the pressure of the main pipe of an air compressor, including:

[0045] S1, obtaining the main pipe pressure data of the air compressor, the main pipe flow data, the pressure data at the end of each branch pipe, and the flow data at the end of each branch pipe, and establishing a prediction data set under a non-fully measured scenario;

[0046] S3, using the main pipe pressure data of the air compressor and the pressure data at the end of each branch pipe as input variables, and using the flow data at the end of each branch pipe as output variables, based on the prediction data set under a non-fully measured scenario, establishing a first prediction model; inputting the real-time data of the input variables into the first prediction model to obtain the current prediction data of the output variables;

[0047] S5, using the current flow prediction data at the end of each branch pipe output by the first prediction model as input variables, and using the main pipe pressure data of the air compressor as output variables, based on the prediction data set under a non-fully measured scenario, establishing a second prediction model; inputting the current flow prediction data at the end of each branch pipe output by the first prediction model into the second prediction model to obtain the current pressure prediction data of the main pipe of the air compressor;

[0048] S7, using the current pressure prediction data of the main pipe of the air compressor output by the second prediction model as the current regulation target value, and regulating the current pressure of the main pipe of the air compressor according to the difference between the currently monitored pressure of the main pipe of the air compressor and the current pressure prediction data of the main pipe of the air compressor.

[0049] Among them, asFigure 2 As shown, in step S1, the main pipe pressure data, main pipe flow data, pressure data at the end of each branch pipe, and flow data at the end of each branch pipe of the air compressor are obtained, and the prediction data set in the non-fully measured scenario is established, which specifically includes:

[0050] S11, install a pressure gauge and a flow meter at the main pipe, and install a pressure gauge at the end of all branch pipes. Only install a flow meter at the end of one branch pipe at the same time.

[0051] As Figure 3 shown, it is a simplified schematic diagram of a certain air compressor system. Air compressor 1, air compressor 2, air compressor 3, and air compressor 4 are 4 air compressors without any difference (the models, configurations, manufacturers, usage conditions, etc. of all air compressors are the same). The main pipe is the output end jointly connected by the four air compressors. A pressure gauge P0 and a flow meter F0 are always installed at the main pipe. End 1, end 2, and end 3 are the three output ends (branch pipe ends) of the main pipe. The order before and after sorting has no influence. Pressure gauges such as P1, P2, and P3 are installed at all ends. However, only one end is installed with a flow meter at the same time. For example, if it is installed at end 1, then at this time, F1 (the flow meter value at end 1) has flow data, and F2 (the flow meter value at end 2) and F3 (the flow meter value at end 3) are empty. The same is true when installed at end 2 or end 3.

[0052] S12, maintain the normal operation of all air compressors, obtain the pressure gauge value and flow meter value at the main pipe, and the pressure gauge value at the end of each branch pipe. Keep the pressure gauge value and flow meter value at the main pipe, and the pressure gauge value at the end of each branch pipe unchanged, and collect the flow meter value at the end of the branch pipe where the current flow meter is installed.

[0053] After installing the flow meter at end 1, maintain the normal operation of the 4 air compressors. By observing the pressure gauge values of the main pipe and ends 1, 2, and 3, keep the pressure gauge value P0 at the main pipe, the pressure gauge value P1 at end 1, the pressure gauge value P2 at end 2, and the pressure gauge value P3 at end 3 unchanged, and collect the flow meter value F1 at end 1 at this time.

[0054] S13, only change any one of the pressure gauge value at the main pipe and the pressure gauge value at the end of each branch pipe each time; collect the flow meter value at the end of the branch pipe where the current flow meter is installed again until the flow meter values at the end of the branch pipe where the current flow meter is installed with the preset number of samples are collected.

[0055] Change the values of the main pipe and end pressures (when changing the pressure at the end of a certain branch pipe, it is default to change the main pipe pressure because in practice, the change of the main pipe pressure often leads to the change of the pressures at each end; when changing the main pipe pressure, the pressures at the branch pipe ends will also change accordingly), collect the value F1 of the flow meter at end 1 again, and repeat the operation until enough sample numbers of F1 values are collected.

[0056] S14. Install the flowmeter at the end of the next branch pipe, and collect the flowmeter values at the end of the next branch pipe for the preset number of samples until the flowmeter values for the preset number of samples are collected at the ends of all branch pipes, obtaining a prediction data set for the training data in the incomplete scenario.

[0057] Then install the flowmeter at end 2, repeat the operation at end 1, and the same applies to end 3, obtaining the training data for the incomplete scenario as shown in Table 1.

[0058] Table 1: Training data table for the incomplete scenario

[0059]

[0060] Among them, for the flowmeter values at the ends of the branch pipes where the flowmeter is not currently installed in the training data for the incomplete scenario, set a boolean mask to mark the null or missing value data that is not considered in subsequent training calculations.

[0061] When the main pipe pressure changes, the main pipe flow data also needs to change. Generally, it is not possible and there is no way to directly adjust the main pipe flow. When predicting the end flow through the pressure data, the main pipe flow is an important measurement standard and boundary value (the end flow must not be greater than or equal to the main pipe flow). For the flowmeter values at the ends of the branch pipes where the flowmeter is currently installed in the training data for the incomplete scenario, when the flowmeter value at the end of the branch pipe is not less than the flowmeter value at the main pipe, maintain the current pressure gauge value and flowmeter value at the main pipe, keep the current pressure gauge value at each end of the branch pipe unchanged, and re-collect the flowmeter value at the end of the branch pipe where the flowmeter is currently installed.

[0062] Among them, in step S3, standardize the data in the training set (prediction data set for the incomplete scenario) in the prediction data set for the incomplete scenario to unify the dimension.

[0063] To make the data easier to process in the first prediction model, the first prediction model standardizes all the data, which means that all the data becomes a unified format. By setting their mean to zero and variance to one, the training speed of the first prediction model can be accelerated, and the accuracy of the first prediction model can be improved. In addition, the standardization process can eliminate the dimension differences of the data, making the training of the first prediction model more stable and efficient, ensuring that the data at different ends has a consistent scale, thereby improving the learning efficiency and prediction ability of the first prediction model.

[0064] Set a custom loss function (for example, it can be the MSE function, that is, the mean square error function), and set a boolean mask for the null values of x in the training data for the incomplete scenario to mark which data are null values not considered in subsequent calculations.

[0065] Because it is impossible to install flow meters at all ends simultaneously, there are missing values in the data. Traditional processing methods may ignore this data or replace it with certain fixed values, but these methods will affect the accuracy and precision of the model. By customizing the loss function, these missing data points are ignored when training the first prediction model, ensuring that the first prediction model only learns valid data, thereby improving the accuracy and robustness of the first prediction model.

[0066] Define the model construction function, initialize the sequential first prediction model, set the input layer, hidden layer, and output layer of the neural network, set the activation function (for example, the ReLU function), add a dropout layer, compile the model, select the optimizer, and set the loss function to the custom loss function set above.

[0067] Build the first prediction model, with the input shape being the number of columns of the standardized input data, define an early stopping callback function to prevent getting stuck, and train the first prediction model, setting the parameters.

[0068] Output the predicted training data, denormalize the predicted values and true values of the training data, calculate the mean square error and the coefficient of determination to evaluate the first prediction model, and obtain a flow identification model (the first prediction model) in a non-complete scenario with an approximate complete data training effect.

[0069] As Figure 4 shown, an embodiment of the present invention also provides a method for regulating the pressure of the main pipe of an air compressor, that is, before step S5, it further includes:

[0070] S4. Determine whether each current flow prediction data at the end of each branch pipe obtained by inputting the real-time data of the input variables into the first prediction model is less than the current flow meter value at the main pipe. If the judgment result is yes, execute step S5; if the judgment result is no, execute step S3, that is, re-input the real-time data of the input variables into the first prediction model to predict the current flow at the end of this branch pipe, or, after correcting the structural parameters of the first prediction model, predict the current flow at the end of this branch pipe until each current flow prediction data at the end of all branch pipes is less than the current flow meter value at the main pipe.

[0071] Among them, in step S5, a model for predicting the main pipe pressure from the flow rate (the second prediction model) is established. The specific process can be:

[0072] Standardize the predicted complete flow rate data set (the current flow prediction data at the end of each branch pipe) obtained through step S3, and use the flow prediction data at the end of each branch pipe as the input end to train a model that can predict the main pipe pressure in real time by inputting the flow rate data.

[0073] To make the data more easily processed in the second prediction model, the second prediction model standardizes all the data, which means that all the data is transformed into a unified format. By setting their mean to zero and variance to one, the training speed of the second prediction model can be accelerated, and the accuracy of the second prediction model can be improved. In addition, the standardization process can eliminate the dimensional differences of the data, making the training of the second prediction model more stable and efficient, ensuring that the data at different ends has a consistent scale, thereby improving the learning efficiency and prediction ability of the second prediction model.

[0074] Set a custom loss function (for example, it can be the MSE function, that is, the mean squared error function), and set a boolean mask for the null values with x in the non-complete scenario training data to mark which data are null values that are not considered in subsequent calculations.

[0075] Since it is impossible to install flow meters at all ends simultaneously, there are missing values in the data. Traditional processing methods may ignore this data or replace it with some fixed values, but these methods will affect the accuracy and precision of the model. By using a custom loss function, these missing data points are ignored when training the second prediction model, ensuring that the second prediction model only learns valid data, thereby improving the accuracy and robustness of the second prediction model.

[0076] Define the model construction function, initialize the sequential second prediction model, set the input layer, hidden layer, and output layer of the neural network, set the activation function (for example, it is the ReLU function), and set the loss function to the custom loss function set above.

[0077] Construct the second prediction model, with the input shape being the number of columns of the standardized input data, define an early stopping callback function to prevent getting stuck, train the second prediction model, and set the parameters.

[0078] Output the predicted training data, denormalize the predicted values and true values of the training data, calculate the mean squared error and the coefficient of determination to evaluate the second prediction model, adjust the parameters to improve the model accuracy, and obtain the flow prediction main pipe pressure model (the second prediction model) in the non-complete scenario with an approximate complete data training effect.

[0079] It should be noted that both the first prediction model and the second prediction model are prediction models based on neural networks. The neural network can be a BP neural network or other types of neural networks, which are not limited in the present invention.

[0080] Among them, in step S7, the current pressure prediction data of the air compressor main pipe output by the second prediction model is used as the current regulation target value, and the current pressure of the air compressor main pipe is regulated according to the difference between the current monitored pressure of the air compressor main pipe and the current pressure prediction data of the air compressor main pipe. For the pressures at the ends 1, 2, and 3, there are generally technological constraint requirements for the lower pressure limit (the same or different). When the end pressure during production is higher than the lower limit requirement, the corresponding end pressure can be reduced to be equal to it to achieve the purpose of energy conservation.

[0081] For example, it has been measured that the current pressure of the main pipe is 0.539531 kPa (the current monitored pressure of the main pipe, and all subsequent pressure units are defaulted to kPa), the maximum threshold of the main pipe pressure is set to 0.53 (the technological constraint requirement of the main pipe pressure), the end pressure 1 is 0.510998 (the current monitored pressure of the branch end 1), the end pressure 2 is 0.530519 (the current monitored pressure of the branch end 2), the end pressure is 0.526239 (the current monitored pressure of the branch end 1), and the end flow 1 is 16.89187 M 3 / s (the current flow prediction data of the branch end 1, and all subsequent flow units are defaulted to M 3 / s), the end flow 2 is 7.751181 (the current flow prediction data of the branch end 2), the end flow 3 is 9.413598 (the current flow prediction data of the branch end 3). By inputting the current flow prediction data of each branch end into the second prediction model, it is obtained that normal production can be carried out when the main pipe pressure is 0.528876 (the current pressure prediction data of the main pipe) at this time. Taking the current pressure prediction data of the main pipe at this time as the regulation target value, there is an optimization space of 0.010665 (0.539531 kPa - 0.528876), and it is sent to the air compressor system for execution. At the same time, the triggering mechanism for the air compressor system to regulate the current pressure of the air compressor main pipe can be real-time, or according to the change of working conditions, or a set triggering period can be set. That is, when the working conditions change, the air compressor system can monitor and measure the end flow prediction data F1, F2, F3. When the working conditions suddenly change, such as when the end flow prediction data increases to reach the preset change threshold, the prediction of the current pressure prediction data of the main pipe is carried out to ensure or restart the optimization calculation and regulation process; or, when the preset triggering period is reached, the air compressor system can monitor and measure the end flow prediction data F1, F2, F3, and carry out the prediction of the current pressure prediction data of the main pipe to ensure or restart the optimization calculation and regulation process, realizing real-time optimization and iteration.

[0082] In practical applications, due to cost or other factor limitations, it is impossible to install flow meters at all ends simultaneously. According to this solution, the problem of being unable to measure the flow of all ends simultaneously can be solved, and it provides a data basis for the subsequent prediction and regulation of the main pipe pressure data.

[0083] Through the prediction of the first prediction model, it is no longer necessary to install flow meters at each end simultaneously. This not only saves costs but also reduces the complexity of installation and maintenance, achieving comprehensive prediction and monitoring of the flow rate at the end of the air compressor while reducing equipment investment. Thus, it not only provides an economical and efficient flow monitoring solution, reduces the use of physical equipment, lowers the overall cost of the system, improves economic efficiency, and realizes energy conservation and consumption reduction; moreover, it can also perform real-time regulation of the main pipe pressure based on the predicted data of the flow rate at the end of the air compressor.

[0084] Using deep learning technology, a black-box model (the first prediction model and the second prediction model) is constructed. The black-box model can understand and learn the relationships between various parts of the system. Even in the absence of comprehensive measurement data, it can calculate the flow rate of the entire system based on the existing data and can train an end-flow rate identification model and a flow prediction main pipe pressure with similar training effects as those of a complete data training set through an incomplete data training set.

[0085] The technical solution of the present invention takes the main pipe pressure data of the air compressor and the pressure data at the end of each branch pipe as input variables, and the flow rate data at the end of each branch pipe as output variables. Based on the prediction data set under the non-complete measurement scenario, a first prediction model is established; based on the first prediction model, the current prediction data of the output variable is obtained; taking the current flow rate prediction data at the end of each branch pipe output by the first prediction model as input variables and the main pipe pressure data of the air compressor as output variables, based on the prediction data set under the non-complete measurement scenario, a second prediction model is established; based on the second prediction model, the current pressure prediction data of the main pipe of the air compressor is obtained; taking the current pressure prediction data of the main pipe of the air compressor output by the second prediction model as the current regulation target value, the current pressure of the main pipe of the air compressor is regulated. Not only can comprehensive monitoring be achieved without installing flow meters at each end, thereby reducing the overall cost of the system; moreover, based on the flow rate prediction data at the end of each branch pipe, the current pressure prediction data of the main pipe of the air compressor is obtained, and the current pressure prediction data of the main pipe of the air compressor is used as the current regulation target value (the current main pipe pressure threshold); compared with determining the current main pipe pressure threshold according to the operating parameters of the air compressor, the types of data involved in this solution are fewer, and it is more convenient to implement, and the effective combination of flow rate monitoring at the end of the branch pipe and main pipe pressure control can be achieved; effectively solving the problem of low reliability of main pipe pressure control and high energy cost caused by the existing technology, effectively improving the reliability of main pipe pressure control and reducing the energy cost.

[0086] In the technical solution of the present invention, a pressure gauge and a flow meter are installed at the main pipe, and pressure gauges are installed at the ends of all branch pipes. Only one flow meter is installed at the end of one branch pipe at the same time. The flow meters are installed at the ends of each branch pipe in turn, and the flow meter values at the ends of the branch pipes with a preset sample number are collected until the flow meter values with a preset sample number are collected at the ends of all branch pipes, obtaining a prediction data set for the training data under the incomplete scenario, reducing the monitoring cost of the flow rate at the end of the branch pipe.

[0087] In the technical solution of the present invention, for the training data to be trained under the incomplete scenario, when the flow meter value at the end of the branch pipe where the current flow meter is installed is not less than the flow meter value at the main pipe, the current pressure gauge value and flow meter value at the main pipe are maintained, and the current pressure gauge value at the end of each branch pipe remains unchanged. The flow meter value at the end of the branch pipe where the current flow meter is installed is re-collected, ensuring the accuracy of the prediction data set under the incomplete scenario, and thus ensuring the accuracy of the data monitoring of the flow rate at the end of the branch pipe. In the technical solution of the present invention, it is judged whether the real-time data of the input variable is input into the first prediction model to obtain the current flow prediction data at the end of each branch pipe, and whether it is not less than the current flow meter value at the main pipe. If the current flow prediction data at the end of a certain branch pipe is not less than the current flow meter monitoring value at the main pipe, the real-time data of the input variable is re-input into the first prediction model to predict the current flow rate at the end of the branch pipe, or the structural parameters of the first prediction model are corrected and then the current flow rate at the end of the branch pipe is predicted until the current flow prediction data at the ends of all branch pipes are not less than the current flow meter value at the main pipe, ensuring the accuracy of the prediction data set under the incomplete scenario, and thus ensuring the accuracy of the data monitoring of the flow rate at the end of the branch pipe.

[0088] In the technical solution of the present invention, the models, configurations, manufacturers, and usage conditions of all air compressors are the same, further ensuring the accuracy of the prediction data set under the incomplete scenario, and thus further ensuring the accuracy of the data monitoring of the flow rate at the end of the branch pipe and the pressure control of the main pipe.

[0089] Embodiment 2

[0090] As Figure 5 shown, the technical solution of the present invention also provides an air compressor main pipe pressure regulation system, including:

[0091] An acquisition module 101, which acquires the air compressor main pipe pressure data, main pipe flow data, pressure data at the end of each branch pipe, and flow data at the end of each branch pipe, and establishes a prediction data set under the incomplete measurement scenario;

[0092] The first establishment module 102 uses the air compressor header pressure data and the pressure data at the end of each branch pipe as input variables, and the flow rate data at the end of each branch pipe as output variables. Based on the prediction dataset in the incomplete measurement scenario, a first prediction model is established. The real-time data of the input variables is input into the first prediction model to obtain the current prediction data of the output variables.

[0093] The second establishment module 103 uses the current flow rate prediction data at the end of each branch pipe output by the first prediction model as input variables, and the air compressor header pressure data as output variables. Based on the prediction dataset in the incomplete measurement scenario, a second prediction model is established. The current flow rate prediction data at the end of each branch pipe output by the first prediction model is input into the second prediction model to obtain the current pressure prediction data of the air compressor header.

[0094] The regulation module 104 uses the current pressure prediction data of the air compressor header output by the second prediction model as the current regulation target value, and regulates the current pressure of the air compressor header according to the difference between the currently monitored pressure of the air compressor header and the current pressure prediction data of the air compressor header.

[0095] It should be noted that the implementation processes of the acquisition module 101, the first establishment module 102, the second establishment module 103, and the regulation module 104 in this embodiment correspond to the method steps in Embodiment 1, and the present invention will not elaborate here.

[0096] The technical solution of the present invention uses the air compressor header pressure data and the pressure data at the end of each branch pipe as input variables, and the flow rate data at the end of each branch pipe as output variables. Based on the prediction dataset in the incomplete measurement scenario, a first prediction model is established. Based on the first prediction model, the current prediction data of the output variables is obtained. Using the current flow rate prediction data at the end of each branch pipe output by the first prediction model as input variables, and the air compressor header pressure data as output variables, a second prediction model is established based on the prediction dataset in the incomplete measurement scenario. Based on the second prediction model, the current pressure prediction data of the air compressor header is obtained. Using the current pressure prediction data of the air compressor header output by the second prediction model as the current regulation target value, the current pressure of the air compressor header is regulated. This can not only achieve comprehensive monitoring without installing flow meters at each end, thereby reducing the overall cost of the system; moreover, based on the flow rate prediction data at the end of each branch pipe, the current pressure prediction data of the air compressor header is obtained, and the current pressure prediction data of the air compressor header is used as the current regulation target value (the current header pressure threshold); compared with determining the current header pressure threshold according to various operating parameters of the air compressor, the data types involved in this solution are fewer, and the implementation is more convenient. It can effectively combine the flow rate monitoring at the end of the branch pipe and the pressure control of the header; effectively solve the problem of low reliability of header pressure control caused by the existing technology, resulting in energy waste and high-cost problems, effectively improve the reliability of header pressure control, and reduce the energy cost.

[0097] In the technical solution of the present invention, a pressure gauge and a flow meter are installed at the main pipe, and pressure gauges are installed at the ends of all branch pipes. Only one flow meter is installed at the end of one branch pipe at the same time. The flow meters are installed at the ends of each branch pipe in sequence, and the flow meter values at the ends of the branch pipes with a preset sample number are collected until the flow meter values with a preset sample number are collected at the ends of all branch pipes, obtaining a prediction data set for the data to be trained in the non-complete scenario, and reducing the monitoring cost of the flow at the end of the branch pipe.

[0098] In the technical solution of the present invention, for the data to be trained in the non-complete scenario, when the flow meter value at the end of the branch pipe where the current flow meter is installed is not less than the flow meter value at the main pipe, the current pressure gauge value and flow meter value at the main pipe are maintained, the current pressure gauge value at the end of each branch pipe remains unchanged, and the flow meter value at the end of the branch pipe where the current flow meter is installed is collected again, ensuring the accuracy of the prediction data set in the non-complete scenario, and thus ensuring the accuracy of the data monitoring of the flow at the end of the branch pipe. In the technical solution of the present invention, it is judged whether the real-time data of the input variable is input into the first prediction model to obtain the current flow prediction data at the end of each branch pipe, and whether it is not less than the current flow meter value at the main pipe. If the current flow prediction data at the end of a certain branch pipe is not less than the current flow meter monitoring value at the main pipe, the real-time data of the input variable is input into the first prediction model again to predict the current flow at the end of the branch pipe, or the structural parameters of the first prediction model are corrected and then the current flow at the end of the branch pipe is predicted until the current flow prediction data at the ends of all branch pipes are not less than the current flow meter value at the main pipe, ensuring the accuracy of the prediction data set in the non-complete scenario, and thus ensuring the accuracy of the data monitoring of the flow at the end of the branch pipe.

[0099] In the technical solution of the present invention, the models, configurations, manufacturers, and usage conditions of all air compressors are the same, further ensuring the accuracy of the prediction data set in the non-complete scenario, and thus further ensuring the accuracy of the data monitoring of the flow at the end of the branch pipe.

[0100] Embodiment III

[0101] As Figure 6 shown, the technical solution of the present invention also provides an electronic device, which is characterized by including: a memory 201 for storing a computer program; a processor 202 for implementing the steps of an air compressor main pipe pressure regulation method as in Embodiment I when executing the computer program.

[0102] The memory 201 in the embodiments of the present application is used to store various types of data to support the operation of the electronic device. Examples of such data include: any computer program for operating on the electronic device. It can be understood that the memory 201 can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a ferromagnetic random access memory (FRAM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM); the magnetic surface memory can be a disk memory or a tape memory. The volatile memory can be a random access memory (RAM), which is used as an external cache.By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), sync link dynamic random access memory (SLDRAM), direct rambus random access memory (DRRAM). The memory 201 described in the embodiments of the present application is intended to include but not limited to these and any other suitable types of memory.

[0103] The method disclosed in the embodiments of the present application can be applied to or implemented by the processor 202. The processor 202 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit in the hardware of the processor 202 or the instructions in the form of software. The above-mentioned processor 202 may be a general-purpose processor, a DSP (Digital Signal Processing, that is, a chip capable of implementing digital signal processing technology), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 202 can implement or execute each method, step, and logic block diagram disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or any conventional processor, etc. Combining the steps of the method disclosed in the embodiments of the present application can be directly embodied as being executed and completed by the hardware decoding processor, or executed and completed by the combination of the hardware and software modules in the decoding processor. The software module may be located in the storage medium, and this storage medium is located in the memory 201. The processor 202 reads the program in the memory 201 and combines its hardware to complete the steps of the foregoing method. When the processor 202 executes the program, it realizes the corresponding processes in each method of the embodiments of the present application. For the sake of brevity, it will not be elaborated here.

[0104] Embodiment 4

[0105] The technical solution of the present invention provides a computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, it realizes the steps of a method for regulating the pressure of the main pipe of an air compressor as in Embodiment 1.

[0106] For example, it includes the memory 201 storing the computer program, and the above computer program can be executed by the processor 202 to complete the steps of the foregoing method. The computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disc, or CD-ROM.

[0107] Those of ordinary skill in the art can understand that all or part of the steps to implement the above method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including those of the above method embodiments. The aforementioned storage medium includes various media that can store program codes, such as removable storage devices, ROM, RAM, magnetic disks, or optical discs. Alternatively, if the above integrated units are implemented in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the embodiments of the present application, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing an electronic device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the various embodiments of the present application. The aforementioned storage medium includes various media that can store program codes, such as removable storage devices, ROM, RAM, magnetic disks, or optical discs.

[0108] Although the specific implementation manners of the present invention have been described above in conjunction with the accompanying drawings, it is not a limitation on the protection scope of the present invention. Those skilled in the art should understand that, based on the technical solution of the present invention, various modifications or deformations that can be made without creative efforts by those skilled in the art are still within the protection scope of the present invention.

Claims

1. A method for regulating the pressure of an air compressor main pipe, characterized in that: include: Obtain the main pipe pressure data, main pipe flow data, pressure data at each branch pipe end, and flow data at each branch pipe end of the air compressor, and establish a prediction data set under an incomplete measurement scenario; wherein, obtaining the main pipe pressure data, main pipe flow data, pressure data at each branch pipe end, and flow data at each branch pipe end of the air compressor, and establishing a prediction data set under an incomplete measurement scenario specifically includes: Install a pressure gauge and flow meter at the main pipe, install a pressure gauge at the end of all branch pipes, and install a flow meter at only one branch pipe end at a time; Maintain the normal operation of all air compressors, obtain the pressure count value and flow count value at the main pipe, the pressure count value at the end of each branch pipe, maintain the pressure count value and flow count value at the main pipe, the pressure count value at the end of each branch pipe remains unchanged, and collect the flow count value at the end of the branch pipe where the flow meter is currently installed; Each time, only one of the pressure count value at the main pipe and the pressure count value at the end of each branch pipe is changed; the flow count value at the end of the branch pipe where the flow meter is currently installed is collected again until a preset number of sample count values ​​at the end of the branch pipe where the flow meter is currently installed are collected; A flow meter is installed at the end of the next branch pipe, and a preset number of sample flow count values ​​of the next branch pipe end are collected, until the preset number of sample flow count values ​​of all branch pipe ends are collected, thereby obtaining a prediction data set of the data to be trained in an incomplete scenario; Taking the pressure data of the main pipe of the air compressor and the pressure data of each branch pipe end as input variables and the flow data of each branch pipe end as output variables, a first prediction model is established based on the prediction data set in the incomplete measurement scenario; the real-time data of the input variables is input into the first prediction model to obtain the current prediction data of the output variables; Taking the current flow prediction data of each branch pipe end output by the first prediction model as input variables and the air compressor main pipe pressure data as output variables, a second prediction model is established based on the prediction data set in the incomplete measurement scenario; the current flow prediction data of each branch pipe end output by the first prediction model is input into the second prediction model to obtain the current pressure prediction data of the air compressor main pipe; The current pressure prediction data of the air compressor main pipe output by the second prediction model is used as the current control target value, and the current pressure of the air compressor main pipe is regulated according to the difference between the current monitored pressure of the air compressor main pipe and the current pressure prediction data of the air compressor main pipe.

2. The method for controlling the pressure of an air compressor main pipe according to claim 1, characterized in that: For the flow count values ​​at the end of the branch pipe where no flow meter is currently installed in the data to be trained in the incomplete scenario, a Boolean mask is set to mark the null value data or missing value data that is not considered for subsequent training calculations.

3. The method for controlling the pressure of an air compressor main pipe according to claim 1, characterized in that: For the data to be trained in the incomplete scenario, when the flow count value at the end of the branch pipe where the flow meter is currently installed is not less than the flow count value at the main pipe, the current pressure count value and flow count value at the main pipe are maintained, the current pressure count value at the end of each branch pipe remains unchanged, and the flow count value at the end of the branch pipe where the flow meter is currently installed is collected again.

4. The method for controlling the pressure of an air compressor main pipe according to claim 3, characterized in that: Taking the current flow prediction data of each branch pipe end output by the first prediction model as the input variable and the main pipe pressure data of the air compressor as the output variable, based on the prediction data set in the incomplete measurement scenario, before establishing the second prediction model, it also includes: Determine whether the current flow prediction data of each branch pipe end obtained by inputting the real-time data of the input variables into the first prediction model is less than the flow count value at the current main pipe. If the current flow prediction data of a branch pipe end is not less than the monitoring value of the flow meter at the current main pipe, then re-input the real-time data of the input variables into the first prediction model to predict the current flow of the branch pipe end, or, after correcting the structural parameters of the first prediction model, predict the current flow of the branch pipe end until the current flow prediction data of all branch pipe ends are less than the flow count value at the current main pipe.

5. A method for controlling the pressure of an air compressor main pipe according to any one of claims 1 to 4, characterized in that: The models, configurations, manufacturers and usage of all air compressors are the same.

6. A method for controlling the pressure of an air compressor main pipe according to any one of claims 1 to 4, characterized in that: Both the first prediction model and the second prediction model are prediction models based on neural networks.

7. An air compressor main pipe pressure control system, characterized in that: include: The acquisition module acquires the air compressor main pipe pressure data, main pipe flow data, pressure data of each branch pipe end, flow data of each branch pipe end, and establishes a prediction data set under an incomplete measurement scenario; wherein, acquiring the air compressor main pipe pressure data, main pipe flow data, pressure data of each branch pipe end, flow data of each branch pipe end, and establishing a prediction data set under an incomplete measurement scenario specifically includes: Install a pressure gauge and flow meter at the main pipe, install a pressure gauge at the end of all branch pipes, and install a flow meter at only one branch pipe end at a time; Maintain the normal operation of all air compressors, obtain the pressure count value and flow count value at the main pipe, the pressure count value at the end of each branch pipe, maintain the pressure count value and flow count value at the main pipe, the pressure count value at the end of each branch pipe remains unchanged, and collect the flow count value at the end of the branch pipe where the flow meter is currently installed; Each time, only one of the pressure count value at the main pipe and the pressure count value at the end of each branch pipe is changed; the flow count value at the end of the branch pipe where the flow meter is currently installed is collected again until a preset number of sample count values ​​at the end of the branch pipe where the flow meter is currently installed are collected; A flow meter is installed at the end of the next branch pipe, and a preset number of sample flow count values ​​of the next branch pipe end are collected, until the preset number of sample flow count values ​​of all branch pipe ends are collected, thereby obtaining a prediction data set of the data to be trained in an incomplete scenario; The first establishment module uses the pressure data of the main pipe of the air compressor and the pressure data of each branch pipe end as input variables, and the flow data of each branch pipe end as output variables, and establishes a first prediction model based on the prediction data set in the incomplete measurement scenario; inputs the real-time data of the input variables into the first prediction model to obtain the current prediction data of the output variables; The second establishment module uses the current flow prediction data of each branch pipe end output by the first prediction model as input variables and the air compressor main pipe pressure data as output variables to establish a second prediction model based on the prediction data set in the incomplete measurement scenario; the current flow prediction data of each branch pipe end output by the first prediction model is input into the second prediction model to obtain the current pressure prediction data of the air compressor main pipe; The control module uses the current pressure prediction data of the air compressor main pipe output by the second prediction model as the current control target value, and controls the current pressure of the air compressor main pipe according to the difference between the current monitored pressure of the air compressor main pipe and the current pressure prediction data of the air compressor main pipe.

8. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, used to implement the steps of an air compressor main pipe pressure control method as described in any one of claims 1 to 6 when executing the computer program.

9. 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 steps of the air compressor main pipe pressure control method according to any one of claims 1 to 6 are implemented.

Images