Ultra-high efficiency intelligent explosion-proof motor dynamic regulation system based on edge computing

Abstract

The application discloses an ultra-high-efficiency intelligent anti-explosion motor dynamic regulation and control system based on edge computing, relates to the motor dynamic regulation and control technical field, and comprises an edge data acquisition module, a data preprocessing module, a real-time load prediction module, a dynamic compensation module, an energy efficiency safety game control module, a cluster cooperative optimization module and an execution module; the edge data acquisition module is used for collecting motor operation state parameters and environmental safety parameters by adopting a multi-sensor fusion technology to obtain original sensing data; the data preprocessing module is used for carrying out noise reduction processing on the original sensing data by adopting a sliding window filtering method, and extracting a vibration signal characteristic frequency band by adopting a wavelet transform method to obtain standardized time series data; and the real-time load prediction module is used for modeling and analyzing the standardized time series data by adopting a space-time convolution network (TCN) to obtain an initial load prediction value.

Description

Technical Field

[0001] This invention relates to the field of dynamic motor control technology, and in particular to an ultra-efficient intelligent explosion-proof motor dynamic control system based on edge computing. Background Technology

[0002] The field of dynamic motor control technology primarily focuses on how to achieve real-time monitoring, analysis, and control of motor operating status through advanced sensing, communication, computing, and control technologies. This field encompasses everything from basic data acquisition, signal processing, and pattern recognition to advanced machine learning, optimization algorithms, and the design and application of distributed control systems. Therefore, how to utilize advanced technologies to improve the intelligence level and safety of ultra-efficient intelligent explosion-proof motor dynamic control systems has become one of the most pressing issues to be addressed.

[0003] In the field of motor dynamic control, raw sensor data usually contains a lot of noise. If the noise is not effectively processed, it will directly affect the quality of data analysis, making the results of fault diagnosis and load prediction inaccurate. Moreover, existing load prediction methods mostly use simple statistical models or rule-based methods, which are difficult to capture complex patterns and long-term and short-term dependencies in time series, resulting in inaccurate prediction results that cannot meet the needs of efficient energy management and fault prevention. At the same time, most traditional control systems use fixed or preset control strategies, lacking the ability to dynamically adjust according to real-time changes, and cannot adapt to complex and ever-changing actual operating environments, limiting the system's flexibility and response speed. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an ultra-efficient intelligent explosion-proof motor dynamic control system based on edge computing to solve the problem that existing load prediction methods often use simple statistical models or rule-based methods, which are difficult to capture complex patterns and long-term and short-term dependencies in time series, resulting in inaccurate prediction results and failing to meet the needs of efficient energy management and fault prevention.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an ultra-efficient intelligent explosion-proof motor dynamic control system based on edge computing, comprising:

[0008] The module includes an edge data acquisition module, a data preprocessing module, a real-time load prediction module, a dynamic compensation module, an energy efficiency and security game control module, a cluster collaborative optimization module, and an execution module.

[0009] The edge data acquisition module is configured to acquire motor operation state parameters and environmental safety parameters by using a multi-sensor fusion technology to obtain original sensing data.

[0010] The data preprocessing module is configured to perform noise reduction processing on the original sensing data by using a sliding window filtering method, and extract a vibration signal characteristic frequency band by using a wavelet transform method to obtain standardized time series data.

[0011] The real-time load prediction module is configured to model and analyze the standardized time series data by using a spatio-temporal convolution network (TCN) to obtain an initial load prediction value.

[0012] The dynamic compensation module is configured to correct the residual error of the initial load prediction value by using a recursive least squares (RLS) method to obtain a high-precision load prediction value.

[0013] The energy efficiency and safety game control module is configured to optimize the high-precision load prediction value and the environmental parameters by using a multi-objective reinforcement learning algorithm, dynamically calculate a critical safety threshold, and generate an optimal energy efficiency and safety control instruction.

[0014] The cluster collaborative optimization module is configured to perform collaborative calculation on the optimal energy efficiency and safety control instruction by using a distributed algorithm (ADMM), introduce a Nash game equilibrium strategy, and output PWM control parameters of each motor.

[0015] The execution module is configured to issue the PWM control parameters to a motor driver to perform physical control.

[0016] As a preferred scheme of the super-high-efficiency intelligent explosion-proof motor dynamic regulation and control system based on edge computing, the multi-sensor fusion technology is used to acquire motor operation state parameters and environmental safety parameters to obtain original sensing data, and the specific steps are as follows:

[0017] Three groups of high-precision industrial sensors are used for data acquisition.

[0018] The three groups of high-precision industrial sensors include a motor state monitoring group, an environmental monitoring group, and a mechanical monitoring group.

[0019] The motor state monitoring group includes a Hall current sensor, a differential voltage sensor, and a MEMS vibration sensor.

[0020] The environmental monitoring group includes an infrared methane sensor and a PT100 temperature sensor.

[0021] The mechanical monitoring group includes a six-axis inertial measurement unit and an acoustic emission sensor.

[0022] A time-stamp-based hard-synchronization acquisition scheme is designed, and a 10MHz synchronization clock signal is generated by a field programmable gate array (FPGA) and distributed to each sensor node to acquire motor operating state parameters and environmental safety parameters and obtain raw sensor data.

[0023] The motor operating state parameters include current, voltage, speed and temperature of the motor.

[0024] The environmental safety parameters include methane concentration and vibration spectrum.

[0025] As a preferred scheme of the edge computing-based ultra-high-efficiency intelligent anti-explosion motor dynamic regulation system, the original sensor data is denoised by using a sliding window filtering method, and the vibration signal characteristic frequency band is extracted by using a wavelet transform method to obtain standardized time series data, and the specific steps are as follows:

[0026] The size of the sliding window is defined, and for the data of each time point, the arithmetic mean of the data points in the sliding window is calculated.

[0027] Morlet wavelet is selected as the mother wavelet function for time-frequency analysis, and the wavelet transform is applied to the denoised time series data to obtain wavelet coefficients under different scales and shifts.

[0028] By analyzing the results of wavelet transform, specific frequency bands closely related to motor faults or operating states are determined.

[0029] The extracted wavelet coefficients of the characteristic frequency band are standardized to obtain standardized time series data.

[0030] As a preferred scheme of the edge computing-based ultra-high-efficiency intelligent anti-explosion motor dynamic regulation system, the standardized time series data is modeled and analyzed by using a time-space convolution network (TCN) to obtain an initial load prediction value, and the specific steps are as follows:

[0031] Based on the characteristics of the motor operating state parameters, the size of the convolution kernel and the number of network layers are defined, and the output of each layer is defined as represents the hidden state of the lth layer and the time point t.

[0032] Residual connections are introduced between each layer to enable information to be directly transmitted from the previous layer to subsequent layers, and the expression is as follows:

[0033]

[0034] wherein, is the hidden state of the lth layer and the time point t in the TCN model, ReLU is an activation function, and Conv is a one-dimensional convolution operation.

[0035] The model performance is evaluated by using mean square error loss function L, and the expression is:

[0036]

[0037] Where y i represents the actual load value, is the load value predicted by the TCN model;

[0038] The Adam optimization algorithm is selected to adjust the model parameters to minimize the above loss function, and the expression is:

[0039]

[0040] Where θ t is the model parameter, and α is the learning rate, and are the first and second matrix estimates of the gradient, respectively;

[0041] The normalized time series data is input into the trained TCN model to obtain the load prediction value corresponding to each time point.

[0042] As a preferred scheme of the edge computing-based ultra-high efficiency intelligent explosion-proof motor dynamic regulation and control system, the recursive least squares (RLS) method is used to correct the residual of the initial load prediction value to obtain a high-precision load prediction value, and the specific steps are as follows:

[0043] Set the initial parameters for the RLS algorithm;

[0044] Set the initial weight vector of the initial estimate to be a zero vector, indicating no prior knowledge, and set the initial covariance matrix;

[0045] The initial load prediction value obtained by processing the normalized time series data through the TCN model is used as the input of the RLS algorithm;

[0046] The RLS algorithm is used to correct the residual of the initial load prediction value, and for each time point t, the residual e(t) is calculated, and the expression is:

[0047]

[0048] Where y(t) is the actual load value, and φ(t) is the regression vector containing the current and past load information, is the weight vector estimated at the previous time;

[0049] According to the update rule of the RLS algorithm, first calculate the gain vector, then update the weight vector, and finally update the covariance matrix;

[0050] The updated weight vector The new input data is predicted to obtain a high-precision load prediction value.

[0051] As a preferred scheme of the edge computing-based super-high-efficiency intelligent explosion-proof motor dynamic regulation system, the multi-objective reinforcement learning algorithm is used to optimize the high-precision load prediction value and the environmental parameters, the critical safety threshold is dynamically calculated, and the optimal energy efficiency safety control instruction is generated.

[0052] The multi-objective reinforcement learning algorithm MORL is used to optimize the high-precision load prediction value and the environmental parameters.

[0053] The environmental parameters include methane concentration C(t) and temperature T(t).

[0054] A multi-objective function is defined, and the expression is:

[0055] J(θ)=w1R eff (β)+w2R safe (θ);

[0056] Wherein, R eff (β) represents the energy efficiency return, R safe (θ) represents the safety return, and w1 and w2 are weight coefficients.

[0057] The critical safety threshold S th (t) is determined by a dynamic calculation method, and the expression is:

[0058] S th (t)=αC(t)+βT(t)+γ;

[0059] Wherein, α, β and γ are constants obtained based on experimental data and historical analysis, used to adjust the influence degree of different factors on the safety threshold.

[0060] The MORL algorithm is used to generate the optimal control instruction.

[0061] As a preferred scheme of the edge computing-based super-high-efficiency intelligent explosion-proof motor dynamic regulation system, the distributed algorithm ADMM is used to collaboratively calculate the optimal energy efficiency safety control instruction, the Nash game equilibrium strategy is introduced, and the PWM regulation parameters of each motor are output, and the expression is:

[0062] The state set S of all motors is defined.

[0063] The initial Lagrange multiplier in the ADMM algorithm is set to a zero vector, and the penalty parameter is used to adjust the strictness of the constraint condition.

[0064] For each motor, the corresponding control instruction is set by using the optimal energy efficiency safety control instruction.

[0065] For each iteration, for each motor, its local variables are updated by solving a minimization problem expressed as:

[0066]

[0067] where f i (x i ) represents the cost function of the i-th motor, related to energy efficiency and safety objectives, x j denotes the variables of neighboring motors, determined by the network topology, is the Lagrange multiplier at the current iteration;

[0068] The update rule is used to adjust the Lagrange multiplier, and the Nash equilibrium concept is used to optimize the interaction between motors;

[0069] At the end of each iteration, it is checked whether a Nash equilibrium is reached, i.e. no motor can achieve a better result by changing its strategy alone;

[0070] When a Nash equilibrium is reached, where is the current strategy, x i ′ is any possible alternative strategy;

[0071] The final determined control instruction set is used as input to generate the PWM control parameters of each motor.

[0072] As a preferred scheme of the edge computing-based ultra-high efficiency intelligent explosion-proof motor dynamic regulation system, wherein: the PWM control parameters are sent to the motor driver for physical control, expressed as:

[0073] The communication protocol between the edge computing platform and the motor driver is used;

[0074] The PWM control parameters are converted into a form suitable for processing by the target driver;

[0075] The converted PWM control parameters are sent to the motor driver using the selected communication protocol;

[0076] After receiving the data packet, the motor driver parses it to extract the PWM control parameters and adjusts the working state of the motor according to the parameters;

[0077] The motor driver controls the motor to operate according to the new PWM signal.

[0078] In a second aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and wherein the computer program, when executed by the processor, implements any step of the ultra-high efficiency intelligent explosion-proof motor dynamic regulation and control system based on edge computing according to the first aspect of the present application.

[0079] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any step of the ultra-high efficiency intelligent explosion-proof motor dynamic regulation and control system based on edge computing according to the first aspect of the present application.

[0080] The present application has the beneficial effects that: the original sensing data is denoised by the sliding window filtering method, and the vibration signal feature band is extracted by the wavelet transform method, the standardized time series data is obtained, the noise interference is effectively removed, the data quality is improved, the standardized time series data is modeled and analyzed by the space-time convolution network TCN, the initial load prediction value is obtained, the advanced machine learning model is used to capture the complex patterns in the time series, the load prediction is more accurate, which not only helps to identify possible load fluctuations in advance, but also supports dynamic adjustment of the working state of the motor to optimize energy efficiency, the initial load prediction value is corrected by the recursive least squares RLS, the high-precision load prediction value is obtained, the RLS algorithm is used to continuously update the prediction model and correct the prediction error, further improving the accuracy of load prediction, which can ensure that even in complex and variable actual operating environments, highly accurate load prediction can be maintained, and more precise control strategies can be implemented. BRIEF DESCRIPTION OF DRAWINGS

[0081] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0082] Figure 1 The schematic diagram of the ultra-high efficiency intelligent explosion-proof motor dynamic regulation and control system based on edge computing in Example 1. DETAILED DESCRIPTION

[0083] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings of the specification.

[0084] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. However, it will be appreciated that the present application can be practiced in a variety of ways beyond the specific details set forth herein, having regard to the contents of the whole patent document, and that the present application can be practiced in other but essentially similar ways. Accordingly, the present application is not limited in scope to the specific implementations disclosed herein. Embodiments of the application will be described and explained with additional specificity and detail through the use of the accompanying drawings in which:

[0085] Secondly, the "one embodiment" or "an embodiment" as used herein means a specific implementation, feature or aspect of the application that can be implemented and that includes particular features, structures or characteristics within the scope of the application. The appearances of "in one embodiment" or "in an embodiment" in various places in the specification are not necessarily all referring to the same embodiment nor are separate or alternative embodiments mutually exclusive of other embodiments.

[0086] Embodiment 1, Reference Figure 1 For the first embodiment of the present application, the embodiment provides an ultra-high-efficiency intelligent explosion-proof motor dynamic regulation and control system based on edge computing, comprising:

[0087] an edge data acquisition module, a data preprocessing module, a real-time load prediction module, a dynamic compensation module, an energy efficiency and safety game control module, a cluster collaborative optimization module, and an execution module;

[0088] The edge data acquisition module is configured to acquire motor operating state parameters and environmental safety parameters by using a multi-sensor fusion technology to obtain raw sensor data.

[0089] Further, three groups of high-precision industrial-grade sensor arrays are used for data acquisition.

[0090] The three groups of high-precision industrial-grade sensors include a motor state monitoring group, an environmental monitoring group, and a mechanical monitoring group.

[0091] The motor state monitoring group includes a Hall current sensor, a differential voltage sensor, and a MEMS vibration sensor.

[0092] The environmental monitoring group includes an infrared methane sensor and a PT100 temperature sensor.

[0093] The mechanical monitoring group includes a six-axis inertial measurement unit and an acoustic emission sensor.

[0094] A hard-synchronization acquisition scheme based on a timestamp is designed, and a 10MHz synchronization clock signal is generated by a field programmable gate array (FPGA) and distributed to each sensor node to acquire motor operating state parameters and environmental safety parameters and obtain raw sensor data.

[0095] The motor operating state parameters include the current, voltage, speed, and temperature of the motor.

[0096] The environmental safety parameters include methane concentration and vibration spectrum.

[0097] It should be noted that by adopting three groups of high-precision industrial sensor arrays for data acquisition and designing a time-stamp-based hard-synchronization acquisition scheme, the time consistency between the sensor nodes is ensured, thereby improving the accuracy and reliability of the original sensor data, which not only effectively captures the changes of the motor operating state and environmental safety parameters, but also provides a solid foundation for subsequent data processing and analysis, and helps to achieve more accurate motor control and fault diagnosis.

[0098] The data preprocessing module is configured to adopt a sliding window filtering method to perform noise reduction processing on the original sensor data, and extract a characteristic frequency band of the vibration signal by using a wavelet transform method to obtain standardized time series data.

[0099] Further, the size of the sliding window is defined, and for the data of each time point, the arithmetic mean of the data points in the sliding window is calculated.

[0100] The Morlet wavelet is selected as the mother wavelet function for time-frequency analysis, and the wavelet transform is applied to the noise-reduced time series data to obtain wavelet coefficients at different scales and shifts.

[0101] By analyzing the results of the wavelet transform, a specific frequency band closely related to the motor fault or operating state is determined.

[0102] The extracted wavelet coefficients of the characteristic frequency band are standardized to obtain standardized time series data.

[0103] It should be noted that the sliding window filtering method and wavelet transform technology are used for noise reduction processing and feature extraction of the original sensor data, which can significantly reduce noise interference and improve signal quality. By selecting the Morlet wavelet as the mother wavelet function for time-frequency analysis, the specific frequency band closely related to the motor fault or operating state can be effectively identified, which not only improves the usability of the data, but also provides high-quality input data for subsequent load prediction, and helps to improve the stability and reliability of the entire system.

[0104] The real-time load prediction module is configured to model and analyze the standardized time series data using a spatio-temporal convolutional network (TCN) to obtain an initial load prediction value.

[0105] Further, based on the characteristics of the motor operating state parameters, the size of the convolution kernel and the number of network layers are defined, and the output of each layer is defined as where hlt represents the hidden state at the lth layer and time point t.

[0106] Residual connections are introduced between each layer to enable information to be directly passed from the previous layer to subsequent layers, and the expression is:

[0107]

[0108] wherein, is the hidden state of the l-th layer and time point t in the TCN model, ReLU is an activation function, and Conv is a one-dimensional convolution operation;

[0109] The mean square error loss function L is used to evaluate the performance of the model, and the expression is:

[0110]

[0111] wherein, y i represents the actual load value, is the load value predicted by the TCN model;

[0112] The Adam optimization algorithm is selected to adjust the model parameters to minimize the above loss function, and the expression is:

[0113]

[0114] wherein, θ t is the model parameter, and α is the learning rate, and are the first and second matrix estimates of the gradient, respectively;

[0115] The standardized time series data is input into the trained TCN model to obtain the load prediction value corresponding to each time point;

[0116] It should be noted that using the spatio-temporal convolution network TCN to model and analyze the standardized time series data can capture the complex patterns and long and short term dependencies in the time series, thereby providing more accurate initial load prediction values. The TCN model combines the design of residual connection, further enhancing the learning ability and stability of the model, making the prediction result more reliable. It is of great significance for early warning of potential load fluctuations and optimization of motor operation efficiency, and helps to reduce energy consumption and prolong the service life of equipment.

[0117] The dynamic compensation module is used to use the recursive least squares method RLS to correct the residual of the initial load prediction value to obtain a high-precision load prediction value;

[0118] Further, the initial parameters of the RLS algorithm are set;

[0119] The weight vector of the initial estimate is set to a zero vector, indicating that there is no prior knowledge, and the initial covariance matrix is set;

[0120] The initial load prediction value obtained by processing the standardized time series data through the TCN model is used as the input of the RLS algorithm;

[0121] The RLS algorithm is used to correct the residual error of the initial load prediction value. For each time point t, the residual error e(t) is calculated, and the expression is as follows:

[0122]

[0123] where y(t) is the actual load value, φ(t) is the regression vector containing current and past load information, is the weight vector estimated at the previous time;

[0124] According to the update rule of the RLS algorithm, the gain vector is first calculated, then the weight vector is updated, and finally the covariance matrix is updated;

[0125] The updated weight vector is used to predict the new input data and obtain a high-precision load prediction value;

[0126] It should be noted that the recursive least squares method RLS can dynamically adjust the prediction model to adapt to changes in actual operation by correcting the residual error of the initial load prediction value, thereby obtaining a higher-precision load prediction value. Not only can it quickly respond to real-time changes in load, but also can continuously improve the prediction accuracy, providing a solid basis for subsequent energy efficiency and safety game control, which helps to maintain efficient and stable motor operation in complex and variable operating environments, and improves the overall system performance.

[0127] The energy efficiency and safety game control module is used to optimize the high-precision load prediction value and environmental parameters using a multi-objective reinforcement learning algorithm, and to generate optimal energy efficiency and safety control instructions by dynamically calculating the critical safety threshold;

[0128] Furthermore, the multi-objective reinforcement learning algorithm MORL is used to optimize the high-precision load prediction value and environmental parameters;

[0129] The environmental parameters include methane concentration C(t) and temperature T(t);

[0130] A multi-objective function is defined, and the expression is as follows:

[0131] J(θ)=w1R eff (θ)+w2R safe (θ);

[0132] where R eff (θ) represents the energy efficiency return, R safe (θ) represents the safety return, and w1 and w2 are weight coefficients;

[0133] The critical safety threshold S th (t) is determined using a dynamic calculation method, and the expression is as follows:

[0134] Sth (t) = aC(t) + bT(t) + g;

[0135] where a, b and g are constants derived from experimental data and historical analysis to adjust the degree of influence of different factors on the safety threshold;

[0136] The MORL algorithm is used to generate optimal control instructions;

[0137] It should be noted that the multi-objective reinforcement learning algorithm MORL combined with the method of dynamically calculating the critical safety threshold can maximize the energy efficiency return while ensuring system safety. By defining a multi-objective function that includes energy efficiency and safety return, and determining the critical safety threshold based on experimental data and historical analysis, the generation of optimal energy efficiency and safety control instructions is achieved, which not only improves the safety of the system, but also optimizes the energy efficiency performance, helping to reduce energy consumption while meeting safety requirements.

[0138] The cluster collaborative optimization module is used to use the distributed algorithm ADMM to collaboratively calculate the optimal energy efficiency and safety control instructions, introduce the Nash equilibrium strategy, and output the PWM control parameters of each motor;

[0139] Further, the state set S of all motors is defined;

[0140] The initial Lagrange multiplier in the ADMM algorithm is set to a zero vector, and the penalty parameter is used to adjust the strictness of the constraint condition;

[0141] Using the optimal energy efficiency and safety control instructions, the corresponding control instructions for each motor are set;

[0142] For each iteration, for each motor, its local variable is updated by solving the minimization problem, and the expression is:

[0143]

[0144] where f i (x i ) represents the cost function of the i-th motor related to energy efficiency and safety goals, x j represents the variables of adjacent motors determined by the network topology, is the Lagrange multiplier at the current iteration;

[0145] The update rule is used to adjust the Lagrange multiplier, and the Nash equilibrium concept is used to optimize the interaction between motors;

[0146] At the end of each iteration, it is checked whether a Nash equilibrium is reached, i.e. no motor can achieve a better result by changing its strategy alone;

[0147] When the following conditions are met, the algorithm converges: a Nash equilibrium is reached, where is the current strategy, x i ′ is any possible alternative strategy;

[0148] The final determined control instruction set is taken as input to generate the PWM control parameters of each motor;

[0149] It should be noted that the combination of the distributed algorithm ADMM and the Nash game equilibrium strategy can effectively coordinate the interaction between multiple motors, realize the global optimal PWM control parameter output, and ensure that each motor can run under its optimal strategy by introducing the Lagrange multiplier and the penalty parameter and checking whether the Nash equilibrium state is reached, avoiding the problem of local optimal solution, not only improving the collaborative efficiency of the system, but also enhancing the robustness and stability of the overall system.

[0150] The execution module is used to issue the PWM control parameters to the motor driver for physical control;

[0151] Further, a communication protocol between the edge computing platform and the motor driver is adopted;

[0152] The PWM control parameters are converted into a form suitable for processing by the target driver;

[0153] The converted PWM control parameters are sent to the motor driver using the selected communication protocol;

[0154] After receiving the data packet, the motor driver parses it to extract the PWM control parameters and adjusts the working state of the motor according to the parameters;

[0155] The motor driver controls the motor to run according to the new PWM signal;

[0156] It should be noted that the PWM control parameters are issued to the motor driver for physical control through the communication protocol between the edge computing platform and the motor driver, realizing seamless connection from data analysis to actual operation. By converting the PWM control parameters into a form suitable for processing by the target driver and using the selected communication protocol for transmission, the accuracy and real-time performance of the control instructions are ensured, not only improving the precision of motor control, but also enhancing the flexibility and response speed of the system, which is helpful for efficient motor management and maintenance.

[0157] The embodiment also provides a computer device suitable for the case of the ultra-high-efficiency intelligent explosion-proof motor dynamic regulation and control system based on edge computing, which comprises a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions to realize the ultra-high-efficiency intelligent explosion-proof motor dynamic regulation and control system based on edge computing proposed in the above embodiment.

[0158] The computer device can be a terminal, which includes a processor, a memory, a communication interface, a display screen and an input device connected by a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is configured to perform wired or wireless communication with an external terminal. The wireless communication can be achieved by Wi-Fi, an operator network, NFC (Near Field Communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0159] The embodiment also provides a storage medium having a computer program stored thereon, the program being executed by a processor to implement the dynamic regulation and control system of the ultra-high efficient intelligent explosion-proof motor based on edge computing as described in the above embodiment. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk or an optical disk.

[0160] To sum up, the original sensing data is denoised by the sliding window filtering method, and the vibration signal characteristic frequency band is extracted by the wavelet transform method, the standardized time series data is obtained, the noise interference is effectively removed, the data quality is improved, the standardized time series data is modeled and analyzed by the space-time convolution network TCN, the initial load prediction value is obtained, the advanced machine learning model is used to capture the complex pattern in the time series, the load prediction is more accurate, which not only helps to identify possible load fluctuations in advance, but also supports dynamic adjustment of the working state of the motor to optimize energy efficiency, the initial load prediction value is corrected by the recursive least square method RLS, the high-precision load prediction value is obtained, the prediction model is continuously updated by the RLS algorithm, the prediction error is corrected, and the accuracy of the load prediction is further improved, which can ensure that even in a complex and variable actual operation environment, highly accurate load prediction can be maintained, and more fine control strategy implementation is promoted.

[0161] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it, although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. An ultra-high efficiency intelligent explosion-proof motor dynamic regulation system based on edge computing, characterized in that: The application relates to an energy efficiency and safety game control method for motor clusters. The method comprises the following steps: An edge data acquisition module, a data preprocessing module, a real-time load prediction module, a dynamic compensation module, an energy efficiency and safety game control module, a cluster collaborative optimization module, and an execution module are provided. The edge data acquisition module is used for collecting motor operation state parameters and environmental safety parameters by adopting a multi-sensor fusion technology to obtain original sensing data. The data preprocessing module is used for performing noise reduction processing on the original sensing data by adopting a sliding window filtering method, and extracting a vibration signal characteristic frequency band by adopting a wavelet transform method to obtain standardized time series data. The real-time load prediction module is used for modeling and analyzing the standardized time series data by adopting a time-space convolution network (TCN) to obtain an initial load prediction value. The dynamic compensation module is used for correcting the initial load prediction value by adopting a recursive least square (RLS) method to obtain a high-precision load prediction value. The energy efficiency and safety game control module is used for optimizing the high-precision load prediction value and environmental parameters by adopting a multi-objective reinforcement learning algorithm, dynamically calculating a critical safety threshold value, and generating optimal energy efficiency and safety control instructions. The cluster collaborative optimization module is used for collaboratively calculating the optimal energy efficiency and safety control instructions by adopting a distributed algorithm (ADMM), introducing a Nash game equilibrium strategy, and outputting motor PWM control parameters.

2. The edge computing based super-high efficient intelligent anti-explosion motor dynamic regulation and control system of claim 1, wherein: The execution module is used for issuing the PWM control parameters to a motor driver to perform physical control. The multi-sensor fusion technology is used to collect motor operation state parameters and environmental safety parameters to obtain original sensing data, and the specific steps are as follows: Three groups of high-precision industrial sensors are used for data collection. The three groups of high-precision industrial sensors comprise a motor state monitoring group, an environmental monitoring group, and a mechanical monitoring group. The motor state monitoring group comprises a Hall current sensor, a differential voltage sensor, and a MEMS vibration sensor. The environmental monitoring group comprises an infrared methane sensor and a PT100 temperature sensor. The mechanical monitoring group comprises a six-axis inertial measurement unit and an acoustic emission sensor. A hard synchronization collection scheme based on a timestamp is designed, and a 10MHz synchronization clock signal is generated by a field programmable gate array (FPGA) and distributed to each sensor node to collect motor operation state parameters and environmental safety parameters and obtain original sensing data. The motor operation state parameters comprise motor current, voltage, speed, and temperature.

3. The ultra-high efficiency intelligent explosion-proof motor dynamic regulation and control system based on edge computing of claim 2, wherein: The environmental safety parameters comprise methane concentration and vibration spectrum. The specific steps of using the sliding window filtering method to perform noise reduction processing on the original sensing data and using the wavelet transform method to extract a vibration signal characteristic frequency band to obtain standardized time series data are as follows: The size of the sliding window is defined, and the arithmetic mean value of each time point data is calculated by using all data points in the sliding window. Morlet wavelet is selected as a mother wavelet function for time-frequency analysis, and wavelet transform is applied to the noise-reduced time series data to obtain wavelet coefficients under different scales and shifts. The results of the wavelet transform are analyzed to determine specific frequency bands closely related to motor faults or operation states. The wavelet coefficients of the extracted feature band are standardized to obtain standardized time series data.

4. The ultra-high efficiency intelligent explosion-proof motor dynamic regulation and control system based on edge computing of claim 3, wherein: The TCN is used to model and analyze the standardized time series data to obtain an initial load prediction value, and the specific steps are as follows: Based on the characteristics of motor operating state parameters, the size of the convolution kernel and the number of network layers are defined, and the output of each layer is defined as denotes the hidden state at the lth layer and time point t; Residual connections are introduced between each layer to enable information to be directly passed from the previous layer to subsequent layers, and the expression is as follows: wherein, is the hidden state of the l-th layer and time point t in the TCN model, ReLU is an activation function, and Conv is a one-dimensional convolution operation. The mean square error loss function L is used to evaluate the performance of the model, and the expression is as follows: where y i represents the actual load value, is the load value predicted by the TCN model; The Adam optimization algorithm is used to adjust the model parameters to minimize the loss function, and the expression is as follows: where θ t is the model parameter, α is the learning rate, and are the first and second matrix estimates of the gradient, respectively; The standardized time series data is input into the trained TCN model to obtain the load prediction value corresponding to each time point.

5. The edge computing based super-high efficient intelligent explosion-proof motor dynamic regulation and control system of claim 4, wherein: The RLS is used to correct the residual of the initial load prediction value to obtain a high-precision load prediction value, and the specific steps are as follows: Initial parameters are set for the RLS algorithm; The weight vector of the initial estimate is set to a zero vector, indicating no prior knowledge, and the initial covariance matrix is set; The initial load prediction value obtained by processing the standardized time series data through the TCN model is used as the input of the RLS algorithm; The RLS algorithm is used to correct the residual of the initial load prediction value, and for each time point t, the residual e(t) is calculated, and the expression is as follows: where y(t) is the actual load value, φ(t) is a regression vector containing current and past load information, is the weight vector estimated at the previous time instant; According to the update rule of the RLS algorithm, the gain vector is first calculated, then the weight vector is updated, and finally the covariance matrix is updated; adopting the updated weight vector The new input data is predicted to obtain a high-precision load prediction value. 6.The edge computing based super-high efficient intelligent anti-explosion motor dynamic regulation and control system according to claim 5, characterized in that: The MORL is used to optimize the high-precision load prediction value and environmental parameters, dynamically calculate the critical safety threshold, and generate optimal energy efficiency and safety control instructions, and the specific steps are as follows: The MORL is used to optimize the high-precision load prediction value and environmental parameters; The environmental parameters include methane concentration C(t) and temperature T(t); A multi-objective function is defined, and the expression is as follows: J(θ) = w1R eff (θ) + w2R safe (θ); wherein R eff (θ) represents the energy efficiency return, R safe (θ) represents the safety return, and w1 and w2 are weight coefficients. A dynamic calculation method is used to determine the critical safety threshold S th (t), expressed by the formula: S th (t) = aC(t) + bT(t) + g Where α, β, and γ are constants based on experimental data and historical analysis, used to adjust the influence degree of different factors on the safety threshold; The MORL algorithm is used to generate optimal control instructions.

7. The ultra-high efficiency intelligent explosion-proof motor dynamic regulation and control system based on edge computing of claim 6, wherein: The ADMM is used to collaboratively calculate the optimal energy efficiency and safety control instructions, introduce the Nash equilibrium strategy, and output the PWM control parameters of each motor, and the expression is as follows: The state set S of all motors is defined; The initial Lagrange multiplier in the ADMM algorithm is set to a zero vector, and the penalty parameter is set to adjust the strictness of the constraint condition; For each motor, the corresponding control instruction is set using the optimal energy efficiency and safety control instruction; For each iteration, the local variable of each motor is updated by solving the minimization problem, and the expression is as follows: where f i (x i ) represents the cost function of the i-th motor, related to energy efficiency and safety targets, x j represents variables of adjacent motors, determined by the network topology, is the Lagrange multiplier at the current iteration number; The update rule is used to adjust the Lagrange multiplier, and the Nash equilibrium concept is used to optimize the interaction between motors; At the end of each iteration, it is checked whether the Nash equilibrium is reached, i.e., no motor can obtain a better result by changing its strategy alone; When the conditions are met then a Nash equilibrium is reached, where is the current policy, x i ′ is any possible alternative policy; The final determined control instruction set is used as input to generate the PWM control parameters of each motor.

8. The ultra-high efficiency intelligent explosion-proof motor dynamic regulation and control system based on edge computing of claim 7, wherein: The PWM control parameters are sent to the motor driver to perform physical control, and the expression is as follows: The communication protocol between the edge computing platform and the motor driver is used; convert the PWM control parameters into a form suitable for processing by the target driver; send the converted PWM control parameters to the motor driver using a selected communication protocol; after the motor driver receives the data packet, parse it to extract the PWM control parameters and adjust the motor's operating state according to the parameters; use the motor driver to control the motor to operate according to the new PWM signal. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is characterized in that: The processor executes the computer program to implement the steps of the edge computing-based ultra-high-efficiency intelligent explosion-proof motor dynamic regulation and control system of any one of claims 1-7.

10. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to implement the steps of the edge computing-based ultra-high-efficiency intelligent explosion-proof motor dynamic regulation and control system of any one of claims 1-7.

Images