Intelligent air compressor control system and method based on deep learning and optimization algorithm

Abstract

The embodiment of the invention provides an intelligent air compressor control system and method based on deep learning and an optimization algorithm, and the system comprises a pressure sensing device which is configured to obtain pressure feature data; the electric quantity monitoring equipment is configured to obtain power consumption data of the air compressor; the temperature sensing equipment is configured to obtain temperature data of the air compressor; and the processor is configured to periodically determine operation parameters of the air compressor based on the adjusting period. And in one adjusting period, the processor is configured to determine operation parameters of one air compressor through an air compression model based on the pressure characteristic data, the air compressor power consumption data and the air compressor temperature data.

Description

Technical Field

[0001] This specification relates to the field of intelligent control, and in particular to an intelligent air compressor control system and method based on deep learning and optimization algorithms. Background Art

[0002] In modern industry, air compressors are widely used as air source equipment in pneumatic tools, mechanical equipment and other links of gas production lines. With the increasing requirements for energy conservation and emission reduction, the energy consumption of air compressors has attracted widespread attention.

[0003] Traditional air compressor control methods, such as proportional-integral-differential (PID) controllers, are widely used in air compressor control due to their simple structure and easy implementation. However, PID controllers cannot adapt to complex nonlinear working conditions and dynamic changes, and cannot achieve optimal control of air compressor energy consumption.

[0004] Therefore, it is desired to provide an intelligent air compressor control system and method based on deep learning and optimization algorithm, which can realize flexible control of the air compressor under various working conditions, thereby reducing energy consumption and improving production efficiency. Summary of the invention

[0005] One or more embodiments of the present specification provide an intelligent air compressor control system based on deep learning and optimization algorithms, characterized in that it includes a pressure sensing device, a power monitoring device, a temperature sensing device and a processor; the pressure sensing device is connected to the gas supply pipeline, the gas supply pipeline is connected to at least one of the air compressor, the gas-using equipment and the gas-using production line, and the pressure sensing device is configured to obtain pressure characteristic data; the pressure characteristic data includes gas supply pressure data and / or gas consumption pressure data; the power monitoring device is configured to obtain air compressor power consumption data; the temperature sensing device is arranged on the air compressor and is configured to obtain air compressor temperature data; the processor is configured to periodically determine the operating parameters of the air compressor based on the adjustment cycle; within an adjustment cycle, the processor is configured to: for an air compressor, based on the pressure characteristic data, the air compressor power consumption data and the air compressor temperature data, determine the operating parameters of the air compressor through the air compression model.

[0006] One of the embodiments of the present specification provides an intelligent air compressor control method based on deep learning and optimization algorithm, characterized in that the method is executed by a processor, including: obtaining pressure characteristic data through a pressure sensing device, the pressure characteristic data including gas supply pressure data and / or gas consumption pressure data; obtaining air compressor power consumption data through a power monitoring device; obtaining air compressor temperature data through a temperature sensing device; and based on an adjustment cycle, periodically determining the operating parameters of the air compressor through an air compression model according to the pressure characteristic data, the air compressor power consumption data and the air compressor temperature data.

[0007] One or more embodiments of the present specification provide an intelligent air compressor control device based on deep learning and optimization algorithms, characterized in that the device includes at least one processor and at least one memory; at least one memory is used to store computer instructions; at least one processor is used to execute at least part of the computer instructions to implement the method described above.

[0008] One or more embodiments of the present specification provide a computer-readable storage medium, wherein the storage medium stores computer instructions, and when a computer reads the computer instructions in the storage medium, the computer executes the method described above.

[0009] Beneficial effects: The intelligent air compressor control system and method based on deep learning and optimization algorithm of the present invention can determine the operating parameters of the air compressor through the air compressor model, and can solve the optimal strategy through the reinforcement learning method. It can reduce the energy consumption of the air compressor while ensuring the stability of the air supply, ensure that the air supply pressure and air supply volume meet the production needs, and improve product quality. By periodically determining the operating parameters of the air compressor based on the adjustment cycle, the control strategy can be adjusted in real time based on the current working conditions, and stable control of the air compressor can be achieved under various working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0011] Figure 1 is an exemplary module diagram of an intelligent air compressor control system based on deep learning and optimization algorithm according to some embodiments of this specification;

[0012] Figure 2 is an exemplary flow chart of an intelligent air compressor control method based on deep learning and optimization algorithm according to some embodiments of this specification;

[0013] Figure 3 is an exemplary schematic diagram of an air compression model according to some embodiments of this specification;

[0014] Figure 4 is an exemplary schematic diagram of determining operating parameters of an air compressor and / or cooling parameters of a cooling device according to some embodiments of this specification;

[0015] Figure 5 It is an exemplary schematic diagram of generating alarm information according to some embodiments of this specification. DETAILED DESCRIPTION

[0016] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following is a brief introduction to the drawings required for the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of this specification. For ordinary technicians in this field, this specification can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0017] It should be understood that the "system", "device", "unit" and / or "module" used herein are a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

[0018] As shown in this specification and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "comprise" and "include" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0019] Flowcharts are used in this specification to illustrate the operations performed by the system according to the embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed precisely in order. Instead, the steps may be processed in reverse order or simultaneously. At the same time, other operations may also be added to these processes, or one or more operations may be removed from these processes.

[0020] Figure 1 It is an exemplary module diagram of an intelligent air compressor control system based on deep learning and optimization algorithm shown in some embodiments of this specification.

[0021] In some embodiments, Figure 1 As shown, the intelligent air compressor control system 100 based on deep learning and optimization algorithm includes a pressure sensing device 110, a power monitoring device 120, a temperature sensing device 130 and a processor 140.

[0022] The pressure sensing device 110 refers to a device that can sense pressure signals and convert the pressure signals into electrical signals. The pressure sensing device 110 includes an analog air pressure sensor, a digital air pressure sensor, and the like.

[0023] In some embodiments, pressure sensing device 110 is configured to acquire pressure characteristic data.

[0024] In some embodiments, the pressure sensing device 110 is connected to an air supply pipeline, and the air supply pipeline is connected to at least one of an air compressor, a gas-using device, and a gas-using production line.

[0025] The gas supply pipeline is the passage for gas transportation. The gas supply pipeline includes gas pipelines, control valves, filters, etc. The input end of the gas supply pipeline is the air compressor, and the output end is the gas-using equipment or gas-using production line.

[0026] An air compressor is a device used to compress gas. For example, air compressors include positive displacement compressors, piston compressors, rotary compressors, etc.

[0027] In some embodiments, the air compressor inhales low-pressure air from the air inlet and discharges high-pressure air from the air outlet. The air in the air supply pipeline flows from the high-pressure area to the low-pressure area under the action of pressure. In other words, the high-pressure air discharged by the air compressor flows to the air-consuming equipment or air-consuming production line through the air supply pipeline.

[0028] In some embodiments, the air supply pipeline is mechanically connected to the air compressor, for example, by threaded connection, flange connection, etc.

[0029] Gas-using equipment refers to equipment that needs to use high-pressure gas, such as pneumatic picks, pneumatic rock drills, and riveting machines.

[0030] Gas production lines refer to production lines that require the use of high-pressure gas. For example, in the food or pharmaceutical industry, high-pressure gas is needed to stir slurries. In the chemical industry, some gases are used to increase the pressure through air compressors to facilitate synthesis and polymerization.

[0031] In some embodiments, a plurality of pressure sensing devices 110 may be disposed at various locations of the gas supply pipeline. For example, the pressure sensing device 110 disposed at the input end of the gas supply pipeline is configured to obtain gas supply pressure data, and the pressure sensing device 110 disposed at the output end of the gas supply pipeline is configured to obtain gas usage pressure data.

[0032] The power monitoring device 120 is a device used to monitor the power consumption of electrical equipment. The power monitoring device 120 includes an electronic power meter, a multi-function power meter, and the like.

[0033] In some embodiments, the power monitoring device 120 is disposed in a current loop where the air compressor is located, and is configured to obtain power consumption data of the air compressor.

[0034] The temperature sensing device 130 refers to a device that can sense temperature signals and convert the temperature signals into electrical signals. The temperature sensing device 130 includes a resistance sensor, a thermocouple sensor, and the like.

[0035] In some embodiments, the temperature sensing device 130 is disposed on the air compressor and is configured to obtain temperature data of the air compressor.

[0036] The processor 140 can process data and / or information related to the intelligent air compressor control system 100 based on deep learning and optimization algorithms. The processor 140 can execute program instructions based on these data, information and / or processing results to perform one or more functions described in this application.

[0037] In some embodiments, the processor 140 may include one or more sub-processing devices (e.g., a single-core processing device or a multi-core multi-core processing device). As an example only, the processor may include a central processing unit (CPU), a digital signal processor (DSP), an embedded processor, etc., or any combination thereof.

[0038] In some embodiments, the processor 140 is configured to periodically determine the operating parameters of the air compressor based on the adjustment cycle. In one adjustment cycle, the processor is configured to: for one air compressor, determine the operating parameters of the air compressor through the air compression model based on the pressure characteristic data, the air compressor power consumption data and the air compressor temperature data.

[0039] In some embodiments, Figure 1 As shown, the processor 140 includes an edge processor 141 , a remote device 142 , multiple service databases 143 and a control center 144 .

[0040] The edge processor 141 refers to a processor that is closer to the location where data is generated. In some embodiments, the edge processor 141 is set in the gas production line and is configured to: for an air compressor, based on the pressure characteristic data, the air compressor power consumption data, the air compressor temperature data, and the intake air temperature, determine the operating parameters of the air compressor and / or the cooling parameters of the cooling device through the air compression model; in response to the operating parameters of the air compressor and / or the cooling parameters of the cooling device meeting the update conditions within a preset time period, send an update request to the control center.

[0041] The remote device 142 refers to a device located at a remote location from where data is generated. In some embodiments, the remote device 142 is disposed at a remote location that is in communication with the gas production line, and is configured to periodically train the air compression model based on an update cycle.

[0042] The service database 143 is used to store data and / or information related to the intelligent air compressor control system 100 based on deep learning and optimization algorithms. In some embodiments, multiple service databases are set in the edge processor 141 and / or the remote device 142, and are configured to store at least one of pressure characteristic data, air compressor power consumption data, and air compressor temperature data.

[0043] In some embodiments, data may be stored in multiple service databases 143 according to preset rules. For example, each service database 143 is configured to store storage pressure characteristic data, air compressor power consumption data, and air compressor temperature data generated by the nearest gas production line. For another example, each service database 143 is configured to store storage pressure characteristic data, air compressor power consumption data, and air compressor temperature data generated by multiple gas production lines.

[0044] The control center 144 refers to a device for implementing data transmission and instruction execution. In some embodiments, the control center 144 is configured to periodically transmit the model parameters trained by the remote device to the edge processor based on an update cycle.

[0045] In some embodiments, Figure 1 As shown, the intelligent air compressor control system 100 based on deep learning and optimization algorithm also includes a cooling device 150, an environmental monitoring device 160, a sound monitoring device 170 and an image monitoring device 180.

[0046] The cooling device 150 is a device for realizing a temperature reduction function. For example, the cooling device 150 includes an air-cooled cooler, a water-cooled cooler, a double-cooled cooler, and the like.

[0047] In some embodiments, the cooling device 150 is disposed on the air compressor and is configured to reduce the temperature of the air compressor.

[0048] The environment monitoring device 160 is a device for monitoring the working environment of the air compressor.

[0049] In some embodiments, the environment monitoring device 160 is configured to obtain the temperature of the intake air. In this case, the environment monitoring device 160 may be a temperature sensor disposed at the air inlet of the air compressor.

[0050] The sound monitoring device 170 is a device for collecting sound data, for example, an audio collector or the like.

[0051] In some embodiments, the sound monitoring device 170 is disposed on the gas production line and is configured to collect sound data on the gas production line.

[0052] The image monitoring device 180 is a device for collecting image data, for example, the image monitoring device 180 includes a camera, a video camera, etc.

[0053] In some embodiments, the image monitoring device 180 is disposed on the gas production line and is configured to collect image data on the gas production line.

[0054] In some embodiments of the present specification, various parameters of the air compressor during operation are collected through various monitoring devices, and the operating parameters of the air compressor are adjusted based on the various parameters. This can integrate data from different sensors to achieve efficient use and in-depth analysis of information.

[0055] It should be understood that Figure 1 The illustrated system and its modules may be implemented in various ways.

[0056] It should be noted that the above description of the intelligent air compressor control system 100 and its modules based on deep learning and optimization algorithm is only for the convenience of description and cannot limit this specification to the scope of the embodiments. It is understandable that for those skilled in the art, after understanding the principle of the system, it is possible to arbitrarily combine the modules or form a subsystem to connect with other modules without deviating from this principle. In some embodiments, Figure 1 The pressure sensing device 110, the power monitoring device 120, the temperature sensing device 130 and the processor 140 disclosed in the specification can be different modules in a system, or one module can realize the functions of two or more modules mentioned above. For example, each module can share a storage module, or each module can have its own storage module. Such variations are within the scope of protection of this specification.

[0057] Figure 2 is an exemplary flow chart of an intelligent air compressor control method based on deep learning and optimization algorithm according to some embodiments of this specification. Figure 2 As shown, the process 200 includes the following steps. In some embodiments, the process 200 may be executed by a processor (eg, the processor 140).

[0058] Step 210, obtaining pressure characteristic data through a pressure sensing device.

[0059] Pressure characteristic data refers to data related to the pressure at various locations in the gas supply pipeline. Pressure characteristic data includes gas supply pressure data and / or gas consumption pressure data.

[0060] The air supply pressure data refers to the pressure at the connection between the air compressor and the air supply pipeline. In some embodiments, the processor can determine the pressure sensing data collected by the pressure sensing device closest to the air compressor in the air supply pipeline as the air supply pressure data corresponding to the air compressor.

[0061] Gas pressure data refers to the pressure at the connection between the gas-using equipment or gas production line and the gas supply pipeline. In some embodiments, the processor can determine the pressure sensor data collected by the pressure sensor device closest to the gas-using equipment or gas production line in the gas supply pipeline as the gas pressure data corresponding to the gas-using equipment or gas production line.

[0062] For more information about gas supply pipelines, air compressors, pressure sensing equipment, gas equipment and gas production lines, see Figure 1 And related instructions.

[0063] Step 220, obtaining the power consumption data of the air compressor through the power monitoring device.

[0064] The air compressor power consumption data refers to data related to the power consumption of the air compressor. For example, the air compressor power consumption data includes the power consumption of the air compressor, etc.

[0065] In some embodiments, the processor may collect the power consumption of the air compressor through a power monitoring device, and determine the power consumption of the air compressor as the power consumption data of the air compressor.

[0066] For more information about power monitoring devices, see Figure 1 And related instructions.

[0067] Step 230, obtaining air compressor temperature data through a temperature sensing device.

[0068] The air compressor temperature data refers to the temperature of the air compressor. In some embodiments, the processor can collect the temperature of the air compressor through a temperature sensor device disposed on the air compressor, thereby obtaining the air compressor temperature data.

[0069] For more information on temperature sensing devices, see Figure 1 And related instructions.

[0070] Step 240, based on the adjustment cycle, according to the pressure characteristic data, the air compressor power consumption data and the air compressor temperature data, the air compressor model is used to periodically determine the operating parameters of the air compressor.

[0071] The adjustment cycle refers to the cycle in which the processor adjusts the operating parameters of the air compressor. In some embodiments, the adjustment cycle is a system preset value.

[0072] In some embodiments, the processor determines the adjustment cycle corresponding to the air compressor of the gas production line based on the production plan of the gas production line.

[0073] The production planning of the gas production line refers to the plan for producing products on the gas production line. The production planning of the gas production line includes at least the expected output of the product. The more the expected output of the product, the shorter the corresponding adjustment cycle. The production planning of the gas production line can be determined based on obtaining user input.

[0074] Operating parameters refer to parameters related to the operation of the air compressor. For example, operating parameters include valve opening, motor power of the air compressor, compressor start and stop, etc.

[0075] In some embodiments, within an adjustment cycle, for an air compressor, the processor determines the operating parameters of the air compressor through an air compression model based on pressure characteristic data, air compressor power consumption data and air compressor temperature data.

[0076] The air compressor model is a model used to determine the operating parameters of the air compressor. In some embodiments, the air compressor model is a reinforcement learning (RL) model.

[0077] Figure 3 is an exemplary schematic diagram of an air compression model according to some embodiments of this specification. Figure 3 As shown, the input of the air compression model 320 includes working state information 310, and the output of the air compression model 320 includes optimal parameters 330. The air compression model 320 includes a working module 321 and an optimal action determination module 322.

[0078] The working state information 310 refers to information related to the working state of the air compressor. The working state information 310 includes a pressure characteristic sequence, a power consumption sequence, and a temperature sequence corresponding to the air compressor. Among them, the pressure characteristic sequence is a sequence obtained by arranging the pressure characteristic data at multiple time points in chronological order, the power consumption sequence is a sequence obtained by arranging the power consumption data of the air compressor at multiple time points in chronological order, and the temperature sequence is a sequence obtained by arranging the temperature data of the air compressor at multiple time points in chronological order.

[0079] In some embodiments, before constructing the working status information 310, the processor may clean and normalize the collected data to improve the quality and usability of the data, thereby improving the efficiency of model processing. Data cleaning includes but is not limited to one or more of data deduplication, missing value processing, and outlier processing. Normalization methods include but are not limited to minimum and maximum normalization or Zscore normalization methods.

[0080] The optimal parameters 330 are the preferred operating parameters of the air compressor in the next adjustment cycle.

[0081] In some embodiments, when determining the optimal parameters 330 based on the air compression model 320, the working state information 310 can be input into the air compression model 320. In the model, the working state information 310 is input into the working module 321, and the working module 321 outputs a set of optional actions. The working state information 310 and the set of optional actions are input into the optimal action determination module 322, and the optimal action determination module 322 outputs the optimal optional action 323. The operating parameters of the air compressor corresponding to the optimal optional action 323 output by the optimal action determination module 322 are determined as the optimal parameters 330, as the output of the air compression model 320.

[0082] The working module 321 may determine a set of optional actions of the air compressor based on the working status information of the current adjustment cycle.

[0083] The optional action set refers to the set of actions that the air compressor can perform under a certain working state. The actions that the air compressor can perform include adjusting the valve opening, adjusting the motor power, etc. Under different working states, the optional action set of the air compressor can be different. For example, when the valve opening reaches the maximum, the optional action set of the air compressor does not include increasing the valve opening.

[0084] The optimal action determination module 322 can be used to determine the reward value of each optional action in the optional action set based on the working state information. The reward value can be used to evaluate the impact of adjusting the operating parameters of the air compressor according to the optional action on product production.

[0085] In some embodiments, in the trained air compressor model 320, the optimal action determination module 322 stores the reward values ​​corresponding to when the air compressor performs various optional actions under various working status information, wherein the reward value corresponding to when an optional action is performed under a working status information can be determined based on a set formula.

[0086] For example, the reward value is related to the energy efficiency coefficient and the motor power after the air compressor performs the corresponding action, and the optimal action determination module 322 can determine the reward value based on the first incentive function.

[0087] Exemplarily, the first excitation function can be expressed by the following formula (1).

[0088] G1=e×p (1)

[0089] Among them, e is the energy efficiency coefficient of the air compressor after executing the action, p is the motor power of the air compressor after executing the action, and G is the reward value corresponding to the executed action.

[0090] The energy efficiency coefficient of the air compressor reflects the energy utilization rate of the air compressor. The larger the energy efficiency coefficient, the lower the energy consumption of the air compressor (such as the lower the power consumption). In some embodiments, the energy efficiency coefficient of the air compressor is positively correlated with the air supply pressure data and negatively correlated with the power consumption data of the air compressor.

[0091] In the actual application of the air compression model, after the optimal action determination module 322 obtains the working state information and the optional action set, it can determine the reward value corresponding to each optional action under the working state information.

[0092] The optimal action determination module 322 may determine the optional action with the maximum reward value as the optimal optional action 323 and output it.

[0093] In some embodiments, the optimal action determination module 322 can be a machine learning model, which can be implemented by various methods, such as a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), etc.

[0094] In some embodiments, the air compressor model can be trained based on a reinforcement learning method, such as Deep Q-Learning Network (DQN), Double Deep Q-Learning Network (DDQN), etc. The first training sample can be the historical working state information of the sample air compressor, and the first label is the optimal optional action corresponding to the sample air compressor under the historical working state information. The first training sample can be obtained based on historical data. The first label can be obtained by a reinforcement learning method, for example, the first label can be the optimal optional action with the highest reward value determined based on formula (1).

[0095] As an example only, the processor can construct a training data set based on the historical working state information of the sample air compressor at the first historical time point, and the operating parameters at the second historical time point after executing the corresponding action, and train the air compressor model based on the training data set combined with the reinforcement learning method. Based on the training data set and the actual operating state of the sample air compressor at the second historical time point, the initial air compressor model can determine the reward value corresponding to the air compressor after executing each action under various working state information based on an incentive function such as formula (1). The processor can further update the parameters of the working module and the optimal action determination module of the initial air compressor model based on the reward value and the loss function until the training stop condition is met. The training stop condition may include meeting the number of iterative training times, etc.

[0096] In some embodiments, before constructing a training data set, the processor may clean and normalize the collected historical data to improve the quality and availability of the data, thereby improving the model training effect.

[0097] In some embodiments, Figure 1As shown, the training of the air compression model is performed by the remote device. After the preliminary training of the air compression model is completed, the edge processor determines the operating parameters of the air compressor through the air compression model based on the pressure characteristic data, the air compressor power consumption data and the air compressor temperature data.

[0098] During the use of the air compression model, the remote device periodically trains the air compression model based on an update cycle, and the control center periodically transmits the model parameters trained by the remote device to the edge processor based on the update cycle.

[0099] In some embodiments, the update period is a system default value, for example, the update period is one week.

[0100] In some embodiments, the update period is related to the data volume of the production data of the gas production line. The larger the data volume of the production data, the shorter the update period.

[0101] Production data refers to data related to the production of products. Production data includes pressure characteristic data, air compressor power consumption data and air compressor temperature data.

[0102] In some embodiments, the processor may determine the amount of production data based on the memory usage of the service database. The more memory usage of the service database, the larger the amount of production data.

[0103] The larger the amount of production data, the greater the production volume of the gas production line, and the more variable the situation of the gas production line. In order to meet the accuracy requirements in the production process, it is necessary to shorten the update cycle of the model to improve the accuracy of the model.

[0104] For more information about remote devices, edge processors, service databases, and control centers, see Figure 1 And related instructions.

[0105] In some embodiments, in response to the operating parameters of the air compressor and / or the cooling parameters of the cooling device satisfying the first update condition within a preset period, the edge processor sends a first update request to the control center. The preset period may be a system default value.

[0106] The first update condition is used to determine whether the control center transmits the trained model parameters to the edge processor.

[0107] In some embodiments, the first update condition is that the fluctuation of the operating parameters of the air compressor within a preset period is greater than a first fluctuation threshold, or the fluctuation of the cooling parameters of the cooling device is greater than a second fluctuation threshold. The first fluctuation threshold and the second fluctuation threshold can be set based on experience.

[0108] For more information on cooling parameters of cooling equipment, see Figure 4 And related instructions.

[0109] In some embodiments, the processor calculates the coefficient of variation of each operating parameter of the air compressor within a preset period, and determines the mean of the coefficients of variation of the multiple operating parameters as the fluctuation value of the operating parameter of the air compressor within the preset period. The processor can obtain multiple values ​​of an operating parameter within a preset period, and then determine the standard deviation and mean of the values ​​of the operating parameter within the preset period, and then use the ratio of the standard deviation and mean of the operating parameter as the coefficient of variation of the operating parameter. The processor can determine the fluctuation value of the cooling parameter of the cooling device in the same way.

[0110] In some embodiments, the first fluctuation threshold and the second fluctuation threshold are related to the update cycle and the type of product produced by the gas production line. The shorter the update cycle, the lower the first fluctuation threshold and the second fluctuation threshold; the higher the quality requirements of the product produced by the gas production line, the lower the first fluctuation threshold and the second fluctuation threshold.

[0111] When the first update condition is met, the product quality fluctuates greatly, which may be caused by the model parameters being no longer applicable. By sending the first update request, the model parameters can be actively updated to improve product quality.

[0112] The first update request is to request the control center to transmit the trained model parameters to the edge processor. In response to receiving the first update request, the control center transmits the trained model parameters to the edge processor.

[0113] Remote devices have a long response time and require more computing resources, while edge processors have a short response time and require fewer computing resources. Since model training takes a long time and requires more computing resources, choosing remote devices for model training and edge processors for model application can not only meet the requirements of rapid response of gas production lines, but also meet the accuracy requirements of model training. By integrating data from different sensors, efficient use and in-depth analysis of information can be achieved.

[0114] In some embodiments, within one adjustment cycle, the processor can simultaneously determine the operating parameters of multiple air compressors based on the pressure characteristic data, power consumption data, and temperature data of the multiple air compressors through the air compression model 320. At this time, the working state information 310 includes a sequence combination of working state information of the multiple air compressors, and the optimal parameters 330 include a sequence combination of optimal operating parameters of the multiple air compressors.

[0115] For example, the working status information 310 can be represented as {[P1, E1, T1], [P2, E2, T2], ..., [P n ,E n ,T n ]}. Among them, P n is the pressure characteristic sequence of the nth air compressor, E nis the power consumption sequence of the nth air compressor, T n is the temperature sequence of the nth air compressor. The optimal parameter 330 can be expressed as [A1, A2, …, A n ], where A n Represents the optimal operating parameters corresponding to the nth air compressor.

[0116] In some embodiments, within a regulation cycle, for an air compressor, the processor determines the operating parameters of the air compressor and / or the cooling parameters of the cooling device through the air compression model based on the pressure characteristic data, the air compressor power consumption data, the air compressor temperature data, and the intake air temperature. Figure 4 And related instructions.

[0117] In some embodiments of this specification, the operating parameters of the air compressor are determined by the air compression model, and the optimal strategy can be solved by the reinforcement learning method. While ensuring the stability of the air supply, the energy consumption of the air compressor can be reduced, and the air supply pressure and air supply volume can be guaranteed to meet production needs and improve product quality. By periodically determining the operating parameters of the air compressor based on the adjustment cycle, the control strategy can be adjusted in real time based on the current working conditions, and stable control of the air compressor can be achieved under various working conditions.

[0118] It should be noted that the above description of process 200 is only for example and illustration, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to process 200 under the guidance of this specification. However, these modifications and changes are still within the scope of this specification. For example, steps 210-230 can be performed simultaneously.

[0119] Figure 4 It is an exemplary schematic diagram for determining the operating parameters of an air compressor and / or the cooling parameters of a cooling device according to some embodiments of this specification.

[0120] In some embodiments, Figure 4 As shown, within one adjustment cycle, for one air compressor, the processor determines the operating parameters 421 of the air compressor and / or the cooling parameters 422 of the cooling equipment through the air compression model 320 based on the pressure characteristic data 411, the air compressor power consumption data 412, the air compressor temperature data 413, and the intake air temperature 414.

[0121] For more information on air compressors and cooling equipment, see Figure 1 For more information about adjustment cycle, pressure characteristic data, air compressor power consumption data, air compressor temperature data, air compressor model, and air compressor operating parameters, see Figure 2 Related instructions.

[0122] The intake air temperature 414 refers to the temperature of the air inhaled by the air compressor. In some embodiments, the processor obtains the intake air temperature 414 through an environmental monitoring device. For more information about the environmental monitoring device, see Figure 1 And related instructions.

[0123] The cooling parameters of the cooling device refer to parameters related to the cooling of the cooling device. The cooling parameters of the cooling device at least include the power of the cooling device.

[0124] In some embodiments, Figure 4 As shown, while determining the operating parameters 421 of the air compressor, the air compression model 320 can also be used to determine the cooling parameters 422 of the cooling device. Figure 3 The working state information 310 in the figure also includes the intake air temperature 414, the output of the air compression model 320 (i.e. Figure 3 The optimal parameters 330) also include the preferred cooling parameters of the cooling device in the next adjustment cycle.

[0125] In some embodiments, when the air compression model 320 is used to simultaneously determine the operating parameters 421 of the air compressor and the cooling parameters 422 of the cooling device, the optimal action determination module 322 can determine the reward value of the optional action based on the second incentive function. In some embodiments, the second incentive function is related to the operating parameters of the air compressor, the cooling parameters of the cooling device, the energy efficiency coefficient of the air compressor, and the energy efficiency coefficient of the cooling device. Exemplarily, the second incentive function can be expressed by the following formula (2).

[0126] G2=α×e×p+β×ec×pc (2)

[0127] Among them, e is the energy efficiency coefficient of the air compressor after performing the optional action, p is the motor power of the air compressor after performing the optional action, ec is the energy efficiency coefficient of the cooling device after performing the optional action, pc is the power of the cooling device after performing the optional action, G2 is the score corresponding to the air compressor and cooling device performing the optional action (that is, the reward value of the optional action), and α and β are weights. For more information about the energy efficiency coefficient of the air compressor, see Figure 3 And related instructions.

[0128] In some embodiments, α and β can be set based on actual needs. The higher the working efficiency requirement of the air compressor, the larger α / β is, and the higher the safety requirement of the air compressor, the smaller α / β is.

[0129] In some embodiments of the present specification, the score corresponding to each combination of candidate operating parameters and candidate cooling parameters is calculated through a second incentive function, and the operating parameters of the air compressor and the cooling parameters of the cooling equipment are determined based on the scores. The optimal solution can be found among multiple combinations of candidate operating parameters and candidate cooling parameters, thereby ensuring the stable operation of the air compressor, improving the safety of the air compressor, and reducing the energy consumption of the air compressor and the cooling equipment.

[0130] For more information about the application and training of the air compression model 320 , see the related description of step 240 .

[0131] In some embodiments, the input of the air compression model 320 (i.e. Figure 3 The working status information 310 in the embodiment also includes abnormal air supply data. The abnormal air supply data refers to data related to abnormal air supply of the air compressor. For more information about abnormal air supply data, see Figure 4 And related instructions.

[0132] In some embodiments of the present specification, operating parameters of the air compressor and cooling parameters of the cooling equipment are determined based on abnormal air supply data. The obtained operating parameters and cooling parameters are safer and more reliable, and can reduce the risk of air compressor failure.

[0133] In some embodiments, within one adjustment cycle, the processor can simultaneously determine the operating parameters of multiple air compressors and the cooling parameters of multiple cooling devices based on the pressure characteristic data of multiple air compressors, the power consumption data of the air compressors, the temperature data of the air compressors, and the intake air temperature through the air compression model 320. At this time, the working state information 310 includes a sequence combination of the working state information of the multiple air compressors, and the optimal parameters 330 include a sequence combination of the optimal operating parameters of the multiple air compressors and the optimal cooling parameters of the multiple cooling devices.

[0134] For example, the working status information 310 can be represented as {[P1, E1, T1, t1], [P2, E2, T2, t2], …, [P n ,E n ,T n ,t n ]}. Among them, P n is the pressure characteristic sequence of the nth air compressor, E n is the power consumption sequence of the nth air compressor, T n is the temperature sequence of the nth air compressor, t n is the suction air temperature of the nth air compressor. The optimal parameter 330 can be expressed as [(A1, A2, …, A n )(B1,B2,…,B m )], where A n represents the optimal operating parameters corresponding to the nth air compressor, B mRepresents the optimal cooling parameters corresponding to the mth cooling device.

[0135] In some embodiments of this specification, the influence of ambient temperature on product production is taken into consideration, and the operating parameters of the air compressor and the cooling parameters of the cooling equipment are determined through an air compression model, so as to ensure that the air supply of the air compressor and the ambient temperature meet the production requirements and improve the product quality.

[0136] Figure 5 It is an exemplary schematic diagram of generating alarm information according to some embodiments of this specification.

[0137] In some embodiments, Figure 5 As shown, the processor determines the abnormal air supply data 530 based on the pressure characteristic data 411 and the air compressor power consumption data 412; in response to the abnormal air supply data 530 meeting the early warning condition, an alarm message is generated and sent to the alarm component.

[0138] For more information about pressure characteristic data and air compressor power consumption data, see Figure 2 Related instructions.

[0139] The gas supply abnormality data 530 refers to data related to the abnormal gas supply of the air compressor. For example, the gas supply abnormality data 530 includes the abnormal location, the temperature of the air compressor when the abnormality occurs, etc. The gas supply abnormality includes but is not limited to the difference between the gas supply pressure data and the gas use pressure data exceeding the preset threshold, the fluctuation value of the gas supply volume at multiple time points exceeding the preset fluctuation threshold, etc.

[0140] In some embodiments, the processor constructs a reference vector based on the pressure characteristic data when the air supply is abnormal and the air compressor power consumption data in the historical data, determines the abnormal position when the air supply is abnormal and the air compressor temperature when the abnormality occurs as the label corresponding to the reference vector, and constructs a reference vector library based on multiple reference vectors and labels.

[0141] The processor constructs a target vector based on the current pressure characteristic data and air compressor power consumption data, matches the target vector in the reference vector library, obtains the reference vector with the highest similarity to the target vector, and determines the label corresponding to the reference vector as the current air supply abnormality data. The similarity can be determined based on the vector distance, which includes but is not limited to the cosine distance.

[0142] In some embodiments, when constructing the reference vector library, the processor may also perform clustering optimization on the reference vectors. For example, the reference vectors recorded in the reference vector library are the reference vectors corresponding to the cluster center of each cluster cluster, so as to obtain more concise and representative reference vectors. Clustering methods include but are not limited to K-Means clustering algorithm, DBSCAN clustering algorithm, etc.

[0143] In some embodiments, for an air compressor, the processor determines the air supply abnormality data 530 through the abnormality model 520 based on the pressure characteristic data 411 , the air compressor power consumption data 412 , the sound data 511 and the image data 512 .

[0144] The abnormal model 520 is a model for determining the gas supply abnormality data 530. In some embodiments, the abnormal model 520 is a machine learning model, for example, the abnormal model 520 is a deep neural network (DNN) model.

[0145] The input of the abnormal model 520 includes the pressure characteristic data 411, the air compressor power consumption data 412, the sound data 511 and the image data 512. The output of the abnormal model 520 is the air supply abnormality data 530.

[0146] The sound data 511 refers to data related to the sound on the gas production line. In some embodiments, the processor collects the sound data 511 through a sound monitoring device.

[0147] Image data 512 refers to data related to an image of a gas production line. In some embodiments, the processor collects image data 512 through an image monitoring device.

[0148] For more information about gas production lines, see Figure 2 For more information about the sound monitoring device and the image monitoring device, see Figure 1 And related instructions.

[0149] In some embodiments, the processor trains an abnormal model 520 based on a plurality of second training samples with a second label. For example, the processor may input a plurality of second training samples into an initial abnormal model, construct a loss function based on the output of the initial abnormal model and the second label, iteratively update the parameters of the initial abnormal model based on the loss function, and terminate the iteration when the iteration completion condition is met to obtain a trained abnormal model. The iterative update method includes but is not limited to the gradient descent method, and the iteration completion condition may be that the loss function converges or the number of iterations reaches a threshold.

[0150] The second training sample includes sample pressure characteristic data, sample air compressor power consumption data, sample sound data and sample image data. The second label is the actual air supply abnormality data of the sample air compressor corresponding to the second training sample. The second training sample and the second label can be constructed based on historical data.

[0151] In some embodiments, the remote device 142 is configured to periodically train the abnormal model based on an update cycle. The control center 144 is configured to periodically transmit the model parameters of the abnormal model trained by the remote device 142 to the edge processor 141 based on the update cycle. The edge processor 141 is configured to: determine the gas supply abnormal data through the abnormal model based on the data in the service database 143; in response to the gas supply abnormal data within the preset time period meeting the second update condition, send a second update request to the control center 144.

[0152] For more information about remote devices, edge processors, service databases, and control centers, see Figure 1 For more information about the update cycle, see Figure 3 And related instructions.

[0153] The second update condition is used to determine whether the control center transmits the model parameters of the trained abnormal model to the edge processor.

[0154] In some embodiments, the second update condition is that the difference between the number of alarms and the actual number of faults within a preset period is greater than a preset number threshold. For more information about alarms, see below.

[0155] In some embodiments, after receiving the alarm information, the maintenance personnel will check the air compressor to determine whether the air compressor has a fault, and then obtain the actual number of faults.

[0156] In some embodiments, the preset quantity threshold may be determined based on the amount of resources (manpower, material resources) required to handle the fault. The greater the amount of resources, the greater the preset quantity threshold.

[0157] The second update request is to request the control center to transmit the model parameters of the trained anomaly model to the edge processor. In response to receiving the second update request, the control center transmits the model parameters of the trained anomaly model to the edge processor.

[0158] In some embodiments of the present specification, the abnormal model is trained by a remote device, and the edge processor performs calculations, which can quickly and accurately evaluate the condition of the air compressor and effectively ensure the safety of the production process.

[0159] In some embodiments of the present specification, the working status of the air compressor is evaluated through an abnormal model based on data in multiple modes such as images and audios, which can more comprehensively analyze the status of the air compressor and more easily detect abnormalities of the air compressor.

[0160] The early warning condition is used to determine whether the air compressor fails. In some embodiments, the early warning condition includes that any one of the abnormal position is a preset risk position, the air compressor temperature is greater than a temperature threshold, etc. is satisfied. The temperature threshold can be set based on experience.

[0161] The preset risk position refers to the key position set by the user in advance according to the importance of each link in the production process. For example, the preset position is the air outlet of the air compressor.

[0162] Alarm information refers to information related to the alarm of the alarm component. For example, the alarm information includes the faulty air compressor, the fault location, the severity, etc. In some embodiments, the processor determines the abnormal location as the fault location, and determines the severity based on the air compressor temperature. The higher the air compressor temperature, the greater the severity.

[0163] In some embodiments, after receiving the alarm information, the alarm component warns the user in the form of one or more of text prompts, sound prompts, voice prompts, vibrations, flashing lights, etc.

[0164] In some embodiments of this specification, the product production process is monitored in real time. When data is abnormal, an alarm is sounded through an alarm component installed in the gas production line. When equipment is abnormal, maintenance personnel can be reminded to perform repairs in a timely manner to ensure safe production of products.

[0165] One or more embodiments of the present specification provide an intelligent air compressor control device based on deep learning and optimization algorithms, characterized in that the device includes at least one processor and at least one memory; at least one memory is used to store computer instructions; at least one processor is used to execute at least part of the computer instructions to implement the method described above.

[0166] One or more embodiments of the present specification provide a computer-readable storage medium, wherein the storage medium stores computer instructions, and when a computer reads the computer instructions in the storage medium, the computer executes the method described above.

[0167] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only for example and does not constitute a limitation of this specification. Although not explicitly stated here, those skilled in the art may make various modifications, improvements and corrections to this specification. Such modifications, improvements and corrections are suggested in this specification, so such modifications, improvements and corrections still belong to the spirit and scope of the exemplary embodiments of this specification.

[0168] At the same time, this specification uses specific words to describe the embodiments of this specification. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of this specification. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures or characteristics in one or more embodiments of this specification can be appropriately combined.

[0169] In addition, unless explicitly stated in the claims, the order of the processing elements and sequences described in this specification, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of this specification. Although the above disclosure discusses some invention embodiments that are currently considered useful through various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the essence and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing server or mobile device.

[0170] Similarly, it should be noted that in order to simplify the description disclosed in this specification and thus help understand one or more embodiments of the invention, in the above description of the embodiments of this specification, multiple features are sometimes combined into one embodiment, figure or description thereof. However, this disclosure method does not mean that the features required by the subject matter of this specification are more than the features mentioned in the claims. In fact, the features of the embodiments are less than all the features of the single embodiment disclosed above.

[0171] In some embodiments, numbers describing the number of components and attributes are used. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise specified, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may change according to the required features of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of this specification are approximate values, in specific embodiments, the setting of such numerical values ​​is as accurate as possible within the feasible range.

[0172] Each patent, patent application, patent application publication, and other materials, such as articles, books, specifications, publications, documents, etc., cited in this specification are hereby incorporated by reference in their entirety. Except for application history documents that are inconsistent with or conflicting with the contents of this specification, documents that limit the broadest scope of the claims of this specification (currently or later attached to this specification) are also excluded. It should be noted that if the descriptions, definitions, and / or use of terms in the materials attached to this specification are inconsistent or conflicting with the contents described in this specification, the descriptions, definitions, and / or use of terms in this specification shall prevail.

[0173] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, as an example and not a limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly introduced and described in this specification.

Claims

1. An intelligent air compressor control system based on deep learning and optimization algorithm, characterized in that: Including pressure sensing devices, power monitoring devices, temperature sensing devices and processors; The pressure sensing device is connected to a gas supply pipeline, the gas supply pipeline is connected to at least one of an air compressor, a gas-using device, and a gas-using production line, and the pressure sensing device is configured to obtain pressure characteristic data; the pressure characteristic data includes gas supply pressure data and / or gas-using pressure data; The power monitoring device is configured to obtain power consumption data of the air compressor; The temperature sensing device is disposed on the air compressor and is configured to obtain temperature data of the air compressor; The processor is configured to periodically determine an operating parameter of the air compressor based on an adjustment period; In one of the adjustment cycles, the processor is configured to: For one of the air compressors, the operating parameters of the air compressor are determined through an air compression model based on the pressure characteristic data, the air compressor power consumption data and the air compressor temperature data, and the air compression model is a reinforcement learning model.

2. The system according to claim 1, characterized in that The system further comprises a cooling device and an environmental monitoring device, wherein the environmental monitoring device is configured to obtain an intake air temperature, wherein the intake air temperature is an air temperature of air inhaled by the air compressor; In one of the adjustment cycles, the processor is further configured to: For one of the air compressors, based on the pressure characteristic data, the air compressor power consumption data, the air compressor temperature data, and the intake air temperature, the operating parameters of the air compressor and / or the cooling parameters of the cooling equipment are determined through the air compression model.

3. The system according to claim 2, characterized in that The processor includes an edge processor, a remote device, a plurality of service databases and a control center; The multiple service databases are arranged in the edge processor and / or the remote device, and the multiple service databases are configured to store at least one of the pressure characteristic data, the air compressor power consumption data, and the air compressor temperature data; The remote device is arranged at a remote end communicating with the gas production line, and is configured to periodically train the air compression model based on an update cycle; the control center is configured to periodically transmit the model parameters trained by the remote device to the edge processor based on the update cycle; The edge processor is disposed in the gas production line and is configured as follows: For one of the air compressors, based on the pressure characteristic data, the power consumption data of the air compressor, the temperature data of the air compressor, and the intake air temperature, the operating parameters of the air compressor and / or the cooling parameters of the cooling device are determined through the air compression model; In response to the operating parameters of the air compressor and / or the cooling parameters of the cooling device satisfying an update condition within a preset time period, an update request is sent to the control center.

4. The system according to claim 1, characterized in that The system further comprises an alarm component, which is disposed in the gas production line and is configured to give an alarm based on the alarm information; The processor is further configured to: Determining air supply abnormality data based on the pressure characteristic data and the air compressor power consumption data; In response to the abnormal gas supply data satisfying the early warning condition, an alarm message is generated and sent to the alarm component.

5. The system according to claim 4, characterized in that The system further comprises a sound monitoring device and an image monitoring device, wherein the sound monitoring device and the image monitoring device are arranged in the gas production line, the sound monitoring device is configured to obtain sound data, and the image monitoring device is configured to obtain image data; The processor is further configured to: For one of the air compressors, the air supply abnormality data is determined based on the pressure characteristic data, the air compressor power consumption data, the sound data and the image data through an abnormality model, and the abnormality model is a machine learning model.

6. An intelligent air compressor control method based on deep learning and optimization algorithm, characterized in that: The method is executed by a processor of the intelligent air compressor control system based on deep learning and optimization algorithm according to claim 1, comprising: Acquiring pressure characteristic data through a pressure sensing device, wherein the pressure characteristic data includes gas supply pressure data and / or gas consumption pressure data; Obtain the power consumption data of the air compressor through the power monitoring equipment; Obtain air compressor temperature data through temperature sensing equipment; and Based on the adjustment cycle, according to the pressure characteristic data, the air compressor power consumption data and the air compressor temperature data, the operating parameters of the air compressor are periodically determined through an air compression model, and the air compression model is a reinforcement learning model.

7. The method according to claim 6, characterized in that The periodic determination of the operating parameters of the air compressor based on the adjustment cycle, according to the pressure characteristic data, the air compressor power consumption data and the air compressor temperature data, through the air compression model, includes: Based on the pressure characteristic data, the air compressor power consumption data, the air compressor temperature data, and the intake air temperature, the operating parameters of the air compressor and / or the cooling parameters of the cooling equipment are determined through the air compression model.

8. The method according to claim 6, characterized in that The method further comprises: Determining air supply abnormality data based on the pressure characteristic data and the air compressor power consumption data; In response to the abnormal gas supply data satisfying the early warning condition, an alarm message is generated and sent to the alarm component.

9. An intelligent air compressor control device based on deep learning and optimization algorithm, characterized in that: The apparatus comprises at least one processor and at least one memory; The at least one memory is used to store computer instructions; The at least one processor is configured to execute at least part of the computer instructions to implement the method according to any one of claims 6 to 8.

10. A computer-readable storage medium, characterized in that: The storage medium stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the method according to any one of claims 6 to 8.

Images