Self-adaptive intelligent control method and system of PLC control cabinet

Abstract

The invention provides a self-adaptive intelligent control method and system for a PLC control cabinet. The method comprises the following steps: acquiring real-time data and historical data of the system and preprocessing the real-time data and the historical data; according to the invention, by means of an edge computing architecture, the computing burden of the PLC main control unit is reduced, the real-time data processing and control decision optimization capability is improved, and the adaptability of the system to a complex environment is enhanced; a deep reinforcement learning strategy is adopted, a control strategy is dynamically adjusted according to real-time and historical data, and control precision and system stability are guaranteed; by using a mode of combining LSTM and CNN, potential faults are identified and warned in advance, and the reliability of the system is improved; through task management based on priority scheduling and load balancing, resources are efficiently allocated, calculation bottlenecks and task conflicts are avoided, and the working efficiency is improved; and the bottleneck of the traditional technology is broken through, the intelligence and the self-adaptive capability are improved by utilizing the frontier technology, and the system can stably operate in a complex production environment, has remarkable innovativeness and practicability and has a wide market application prospect.

Description

Technical Field

[0001] The present invention relates to the technical field of PLC control cabinets, and particularly to an adaptive intelligent control method and system for a PLC control cabinet. Background Art

[0002] In the booming field of industrial automation, PLC control cabinets play a crucial role with their excellent performance and high reliability, and have a very wide range of applications. From the highly precise automated production lines in automobile manufacturing plants to the complex production process control in chemical enterprises; from the precise handling and storage management of goods in intelligent warehousing and logistics systems to the stable control and monitoring of various electrical equipment in the power system, PLC control cabinets play an irreplaceable role. It can efficiently achieve various functions such as logical control, sequential control, timing control, and counting control, accurately coordinate each link in the industrial production process, and ensure the efficient, stable, and safe progress of production activities. It has become an indispensable core component in the modern industrial automation system. However, there are still some problems:

[0003] I. Insufficient adaptability: Most are based on preset control strategies. When facing changes in load, working conditions, or external environment, the response efficiency is low, and it is difficult to adjust in a timely manner according to the actual operating conditions, resulting in the system operation being unable to maintain the best state;

[0004] II. Low degree of intelligence: Generally rely on fixed programming logic or rule control, lack the ability of self-learning, prediction, and optimization, and cannot perform intelligent optimization for changes in equipment or production processes, restricting the improvement of production efficiency and product quality;

[0005] III. Lack of energy efficiency optimization: Lack a dynamic energy efficiency adjustment mechanism. During peak energy consumption periods, the system load cannot be intelligently adjusted, resulting in both energy waste and affecting the service life of equipment;

[0006] IV. Response speed and stability problems: When dealing with multi-task concurrency, especially in a complex industrial environment, the existing PLC control system is prone to delays and instability, seriously affecting the continuity and reliability of industrial production;

[0007] Therefore, an adaptive intelligent control method and system for a PLC control cabinet are proposed. Summary of the Invention

[0008] In view of this, embodiments of the present invention hope to provide an adaptive intelligent control method and system for a PLC control cabinet to solve or alleviate the technical problems existing in the prior art and at least provide a beneficial option.

[0009] To solve the above technical problems, a technical solution adopted by the present application is: An adaptive intelligent control method for a PLC control cabinet, comprising the following steps:

[0010] Step 1: Obtain the real-time data and historical data of the device operation and preprocess them using an edge computing device, and transmit the preprocessed real-time data and historical data to the PLC system in real time;

[0011] Step 2: Based on the real-time data and historical data, perform adaptive control based on the reward function, state space, and action space of deep reinforcement learning to optimize the control strategy;

[0012] Step 3: Perform fuzzification processing on the input real-time data based on the intelligent adjustment mechanism of fuzzy control, make output control decisions according to the fuzzy control rules, and fuse the control strategy to generate the final control instruction;

[0013] Step 4: Based on the real-time data and historical data of the device operation, perform fault prediction and intelligent diagnosis on the key components and overall operation status of the device based on convolutional neural network and long short-term memory network to obtain the fault diagnosis result;

[0014] Step 5: Build an edge computing architecture, use a priority scheduling algorithm to schedule real-time tasks, and apply dynamic load balancing technology to monitor the load of each node in real time;

[0015] Step 6: Based on the control strategy, control decision, and fault diagnosis result, dynamically adjust the operating parameters of the PLC control cabinet according to the current operating state, load condition, and energy price information of the device.

[0016] Preferably, as a further optimization of this technical solution, in Step 3, the design of the fuzzy control rules is based on load changes and temperature fluctuations to adjust the output;

[0017] The membership function of the fuzzy rule is calculated by the following formula:

[0018]

[0019] where, μ i (t) is the membership function of the i-th fuzzy rule; x i is the current input value; α i is the slope control parameter of the membership function, which determines the change speed of the membership degree; is the fuzzy center value of the input;

[0020] The final output is obtained by weighted average:

[0021]

[0022] where, u(t) is the output signal of the controller; c i is the output value corresponding to the fuzzy control rule i.

[0023] Preferably, as a further optimization of this technical solution, in step two, the state space includes the temperature, pressure, rotational speed, load current, load voltage of the device during real-time operation, as well as the average load within the past 24 hours and the cumulative operating duration of the device; the action space is to control the start / stop of the motor, the level of adjusting the valve opening, and the gear of adjusting the output voltage.

[0024] Preferably, as a further optimization of this technical solution, in step one, the real-time data of the device operation is obtained through real-time acquisition by a temperature sensor, a humidity sensor, a load current sensor, and a load voltage sensor, the historical data is read from the local database, and the preprocessing includes data filtering and normalization processing.

[0025] Preferably, as a further optimization of this technical solution, in step four, the input layer of the convolutional neural network receives the real-time data and historical data of the device operation, and performs feature extraction and fault mode recognition through the convolutional layer, pooling layer, and fully connected layer; the long short-term memory network receives the time series of the device operation data and performs time series analysis and abnormal trend prediction through the forget gate, input gate, and output gate.

[0026] Preferably, as a further optimization of this technical solution, in step five, the priority scheduling algorithm adopts the EarliestDeadlineFirs algorithm to determine the priority of tasks according to the deadline of the tasks.

[0027] Preferably, as a further optimization of this technical solution, in step six, the dynamic adjustment of the operating parameters of the PLC control cabinet includes adjusting the rotational speed of the motor, adjusting the opening of the valve, and changing the magnitude of the output voltage.

[0028] To solve the above technical problems, another technical solution adopted by this application is: an adaptive intelligent control system for a PLC control cabinet, the system includes: a data acquisition and preprocessing module, a deep reinforcement learning control module, a fuzzy control adjustment module, a fault prediction and diagnosis module, an edge computing and task scheduling module, and a parameter adjustment module;

[0029] The data acquisition and preprocessing module is configured to acquire the real-time data and historical data of the device operation, and perform preprocessing using edge computing devices, and transmit the preprocessed real-time data to the PLC system in real time based on the IoT communication protocol;

[0030] The deep reinforcement learning control module is configured to perform adaptive control based on the real-time data and historical data, the reward function of deep reinforcement learning, as well as the state space and action space, and optimize the control strategy;

[0031] The fuzzy control adjustment module is configured to perform fuzzification processing on the input real-time data based on an intelligent adjustment mechanism of fuzzy control, and make an output control decision according to fuzzy control rules;

[0032] The fault prediction and diagnosis module is configured to perform fault prediction and intelligent diagnosis on the key components and overall operating status of the device based on the convolutional neural network and long short-term memory network according to the real-time data and historical data of the device operation, and obtain a fault diagnosis result;

[0033] The edge computing and task scheduling module is configured to build an edge computing architecture, schedule real-time tasks using a priority scheduling algorithm, and apply dynamic load balancing technology to monitor the load of each node in real time;

[0034] The parameter adjustment module is configured to dynamically adjust the operating parameters of the PLC control cabinet based on the control strategy, control decision, and fault diagnosis result, according to the current operating status, load condition, and energy price information of the device.

[0035] Preferably, as a further aspect of the present technical solution, the data acquisition and preprocessing module is further configured with a data encryption and verification mechanism.

[0036] Preferably, as a further aspect of the present technical solution, the deep reinforcement learning control module adopts a distributed training architecture and uses multiple edge computing devices for parallel computing to accelerate the training process.

[0037] Due to the above technical solutions adopted in the embodiments of the present invention, the following advantages are achieved:

[0038] 1. The present invention reduces the computing burden of the PLC main control unit through the edge computing architecture, enabling the system to quickly process real-time data, optimize control decisions, respond quickly to load changes and external disturbances, and enhance the adaptability to complex working environments;

[0039] 2. The present invention adopts an adaptive control strategy of deep reinforcement learning and dynamically adjusts the control strategy according to real-time data and historical data, improving the adaptability of the system to load fluctuations and environmental changes, and ensuring control accuracy and system stability;

[0040] 3. The present invention can identify potential fault modes in advance and give early warnings through the method combining LSTM and CNN, reduce the impact of faults, and improve the reliability of the system;

[0041] 4. The present invention adopts a task management method based on priority scheduling and load balancing to efficiently allocate resources, avoid computing bottlenecks and task conflicts, and improve work efficiency and system stability;

[0042] 5. By breaking through the technical bottlenecks of traditional PLC systems in terms of response speed, control accuracy, fault warning ability, etc., and using cutting-edge technologies to enhance the system's intelligence and adaptability, the present invention can operate stably in complex production environments, has significant technological innovation and practicality, effectively improves the efficiency and reliability of the PLC control cabinet system, and has broad market application prospects.

[0043] The above summary is only for the purpose of the specification and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the present invention will become readily apparent by reference to the drawings and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application or in the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0045] Figure 1 It is a schematic flowchart of an adaptive intelligent control method for a PLC control cabinet of the present invention;

[0046] Figure 2 It is a schematic diagram of the functional modules of an adaptive intelligent control system for a PLC control cabinet of the present invention;

[0047] Figure 3 It is a structural diagram of Embodiment 3 of the present invention;

[0048] Figure 4 It is a structural diagram of the interior of the cabinet body of Embodiment 3 of the present invention;

[0049] Figure 5 It is a structural diagram of the transmission box of Embodiment 3 of the present invention;

[0050] Figure 6 It is a structural diagram of the ventilation opening of Embodiment 3 of the present invention.

[0051] Reference numerals: 10, main body assembly; 11, cabinet body; 12, cabinet door; 13, control panel; 14, first temperature sensor; 15, light sensor; 16, rain and snow sensor; 17, second temperature sensor; 20, ventilation assembly; 21, ventilation opening; 22, transmission box; 23, drive motor; 24, transmission wheel; 25, transmission chain; 26, fixed shaft; 27, shutter; 30, heat dissipation assembly; 31, mounting seat; 32, fan; 33, electric guide rail; 34, baffle. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] The following describes the embodiments of the present disclosure in detail with reference to the accompanying drawings.

[0053] It should be clear that the following illustrates the implementation manners of the present disclosure through specific specific examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific implementation manners, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present disclosure without making creative efforts belong to the scope of protection of the present disclosure.

[0054] It should also be noted that the following describes various aspects of the embodiments within the scope of the appended claims. It should be obvious that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is illustrative only. Based on the present disclosure, those skilled in the art should understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects described herein can be used to implement the device and / or practice the method. Additionally, this device and / or this method can be implemented using other structures and / or functions in addition to one or more of the aspects described herein.

[0055] It should also be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present disclosure in a schematic manner. The diagrams only show the components related to the present disclosure, rather than being drawn according to the number, shape, and size of the components in actual implementation. The type, quantity, and proportion of each component in actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0056] In addition, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0057] Embodiment 1:

[0058] Figure 1 It is a schematic flowchart of an adaptive intelligent control method for a PLC control cabinet according to an embodiment of the present invention. It should be noted that if there are substantially the same results, the method of this application is not limited to Figure 1 the shown process sequence. As Figure 1Shown: An adaptive intelligent control method for a PLC control cabinet, comprising the following steps:

[0059] Step 1: Obtain the real-time data and historical data of equipment operation and preprocess them using edge computing devices, and transmit the preprocessed real-time data and historical data to the PLC system in real time;

[0060] Specifically, first, use various sensors such as temperature sensors, humidity sensors, load current sensors, and load voltage sensors to collect equipment operation data in real time. These sensors are accurately deployed at key parts of the equipment to ensure that key data reflecting the equipment operation status can be obtained. For example, temperature sensors are installed near easily heated components such as the motor housing and inside the control cabinet to monitor the equipment operation temperature; load current and voltage sensors are connected in series in the power supply line of the equipment to collect current and voltage data in real time;

[0061] Then, read historical data from the local database. The local database has pre-stored the past operation data of the equipment, including operation parameters and fault records under different working conditions. When reading historical data, the system will screen relevant data according to the current requirements, such as data in the past week, month, or under specific working conditions, to ensure the relevance and effectiveness of the data;

[0062] Next, use edge computing devices to perform data filtering and normalization processing on the obtained real-time data and historical data. For some parameters with large amounts of data, such as vibration data or image data of equipment operation, data compression algorithms are used for processing to reduce the data transmission volume without losing key information and improve the transmission efficiency. For example, the wavelet transform compression algorithm is used to compress vibration data, which can not only retain the main features of the data but also greatly reduce the data volume;

[0063] Finally, based on IoT communication protocols such as MQTT, CoAP, etc., transmit the preprocessed real-time data and historical data to the PLC system in real time. These protocols have characteristics such as lightweight, low power consumption, and high reliability, and are suitable for data transmission in the industrial Internet of Things environment. For example, the MQTT protocol uses the publish / subscribe mode to efficiently transmit data from edge computing devices to the PLC system to ensure the real-time and stability of the data; to ensure the reliability of data transmission, the system also uses data verification and retransmission mechanisms. During data transmission, CRC (Cyclic Redundancy Check) or parity check calculations are performed on the data. The receiving end judges whether the data is complete according to the verification result. If the data verification fails, the receiving end will send a retransmission request to the sending end to ensure the accuracy and integrity of the data;

[0064] By performing data preprocessing on edge computing devices, the amount and complexity of data that the PLC system needs to process can be reduced. The PLC system no longer needs to process the original data that may contain noise and dimensional differences, and can allocate more computing resources to logic control and decision execution, improving the operation efficiency and response speed of the PLC system; through preprocessing operations such as data filtering, normalization, and compression, the accuracy, comparability, and transmission efficiency of the data are improved. The high-quality data provides a reliable basis for subsequent operations such as deep reinforcement learning, fuzzy control, fault prediction and diagnosis, helping to improve the performance of the entire system and the accuracy of decision-making; by using IoT communication protocols to transmit data in real time, it is ensured that the PLC system can timely obtain the latest operating status information of the device. Combining real-time data and historical data, the system can respond more quickly, timely adjust the control strategy, adapt to changes in the device operating status, and improve the real-time performance and self-adaptability of the system.

[0065] Step 2: Based on real-time data and historical data, perform adaptive control based on the reward function, state space, and action space of deep reinforcement learning to optimize the control strategy;

[0066] Among them, deep reinforcement learning is a technology that combines deep learning and reinforcement learning. In this step, it interacts with the environment through an agent, continuously tries different actions, and learns the optimal control strategy according to the reward signal feedback from the environment;

[0067] Among them, the state space is all the input features of the control system, such as temperature, load, voltage, etc. These features comprehensively reflect the operating conditions of the device;

[0068] The action space is the corresponding control output, such as switch status, adjustment of output voltage or current, etc. These outputs directly act on the device and affect its operating state;

[0069] The reward function is defined as:

[0070] R(s t , a t ) = -(C error (t)|;

[0071] To encourage the system to approach the target state;

[0072] Among them, C error (t) is the error between the target state and the current state. Taking voltage as an example:

[0073] C error (t) = |V target - V actual |;

[0074] Among them, Vtarget is the target voltage, V actual is the actual voltage;

[0075] The setting of this reward function is such that the closer the system is to the target state, the higher the reward value obtained, thereby guiding the agent to learn a strategy that makes the system tend towards the target state;

[0076] The specific implementation process is as follows:

[0077] First, use a deep learning framework (such as TensorFlow or PyTorch) to construct a deep neural network model for approximating the optimal control strategy. The input of the network is the information of the state space, and the output is the value estimation of each action in the action space. Commonly used structures such as multi-layer perceptron (MLP), convolutional neural network (CNN), or recurrent neural network (RNN) are adopted. Select a suitable network structure according to the characteristics of the state space. For example, for state variables containing time series information (such as the average load in the past 24 hours), RNN or its variants (such as LSTM, GRU) can be used to process to capture the long-term dependencies in the time series;

[0078] Then, update the Q-value function through the Q-learning algorithm to calculate the optimal control strategy for the next action. The calculation formula is:

[0079]

[0080] where, Q(s t , a t ) is the Q-value corresponding to the current state s t and the action a t ; α is the learning rate, which determines the degree of integration of new knowledge and old knowledge;

[0081] γ is the discount factor, which is used to consider the influence of future rewards;

[0082] is the Q-value of the optimal action under the next state s t+1 ;

[0083] When training and optimizing a deep neural network model, the agent continuously conducts experiments in the environment, selects actions based on the current state, and observes the rewards and new states feedback by the environment. Using these empirical data (state, action, reward, new state), the parameters of the deep neural network are updated through optimization algorithms (such as stochastic gradient descent, Adam, etc.), enabling the network to better predict the optimal action. During the training process, the experience replay technique is usually adopted, storing the agent's empirical data in an experience pool and then randomly sampling a batch of data for training to break the correlation between the data and improve the stability and efficiency of training. In addition, a strategy for balancing exploration and exploitation (such as the ε-greedy strategy) can be adopted. At the beginning of training, the agent is encouraged to conduct more exploration and try different actions to discover more potential optimal strategies. As training progresses, the proportion of using the learned optimal strategies is gradually increased to improve the convergence speed of the control strategy.

[0084] As training progresses, the deep neural network model is continuously optimized, and the control strategy is updated accordingly. The updated control strategy is evaluated regularly. For example, the system is run in a test environment to observe whether the operating performance indicators of the device (such as production output, quality, energy consumption, etc.) have improved. According to the evaluation results, the hyperparameters of the deep reinforcement learning algorithm (such as the learning rate, discount factor, etc.) are adjusted to further optimize the control strategy. Through continuous training and evaluation, the deep reinforcement learning algorithm can gradually learn the optimal control strategy adapted to the device operating environment and achieve adaptive control of the PLC control cabinet.

[0085] Step 3: The intelligent adjustment mechanism based on fuzzy control performs fuzzy processing on the input real-time data, makes output control decisions according to the fuzzy control rules, and integrates the control strategy to generate the final control instruction.

[0086] Specifically, first, the system takes the real-time data collected by the sensor, such as load, temperature, etc., as the input of the fuzzy controller. These precise values are not convenient for directly judging by the fuzzy rules and need to be converted into fuzzy quantities. Through a specific fuzzy processing method, based on the preset fuzzy subsets and membership functions, the input data is mapped into the corresponding fuzzy sets to determine its membership degree to each fuzzy set. For example, the load value is mapped into fuzzy sets such as "light load", "medium load", "heavy load", etc. according to the membership function to obtain the membership degree of each fuzzy set.

[0087] Then, the system pre - designs fuzzy control rules according to factors such as load changes and temperature fluctuations; for example, "IF the load change is large AND the temperature is high THEN adjust the control significantly". When the input data is fuzzified, the system will match these rules. During the reasoning process, according to the degree of satisfaction of the input data with the antecedents of each rule, the activation strength of each rule, that is, the weight, is determined. Then, based on these weights, the consequents of each rule are synthesized to obtain a fuzzy output; for example, if there are two rules, the weight of rule one is 0.6 and the output is "increase the control amount"; the weight of rule two is 0.4 and the output is "decrease the control amount". After synthesis, a fuzzy control decision result is obtained.

[0088] Finally, the output result of fuzzy control is fused with the control strategy obtained from deep reinforcement learning. The control strategy of deep reinforcement learning provides a control direction from the perspective of overall system optimization and long - term learning, while fuzzy control quickly adjusts real - time, uncertain and non - linear situations. By weighted average or other fusion algorithms, the two are combined to generate the final control instruction; for example, if the deep reinforcement learning strategy tends to increase the operating speed of the device, fuzzy control fine - tunes the speed according to the current temperature and load conditions. After fusion, an accurate speed adjustment value is obtained, such as "increase the motor speed by X revolutions per minute", and then the PLC control cabinet is controlled to drive the device to perform corresponding actions to ensure the stable and efficient operation of the system.

[0089] Step 4: Based on the real - time data and historical data of the device operation, perform fault prediction and intelligent diagnosis on the key components and overall operation status of the device based on convolutional neural network and long - short - term memory network to obtain the fault diagnosis result.

[0090] Specifically, the specific process of fault mode recognition based on convolutional neural network (CNN) is as follows:

[0091] First, the pre - processed device operation data is input into the input layer of the convolutional neural network. These data can be multi - dimensional, for example, simultaneously including multiple features such as temperature, current, and voltage. The CNN performs convolution operations by sliding the convolution kernels in the convolution layer on the input data to extract local features in the data; each convolution kernel can detect specific types of features, such as edges, textures, etc.; the pooling layer downsamples the output of the convolution layer to reduce the data dimension while retaining important features; after being processed by multiple convolution layers and pooling layers, the CNN can automatically learn the complex feature patterns in the device operation data.

[0092] Then, the fully - connected layer integrates the features extracted by the convolution layer and the pooling layer and maps them to different fault mode categories through activation functions.

[0093] Finally, the output layer gives the predicted results of the possible fault modes of the key components of the device, such as motor faults, sensor faults, etc.

[0094] The specific process for time series prediction based on the Long Short-Term Memory Network (LSTM) is as follows:

[0095] First, the LSTM network receives the time series of device operation data as input. Since the operating state of the device often has temporal correlation, for example, the temperature at the current moment may be affected by the temperature at the previous moment and the device operation history, the LSTM can capture this time series information;

[0096] Then, the LSTM unit controls the flow and memory of information through the forget gate, input gate, and output gate. The forget gate determines whether to discard some information from the previous moment, the input gate determines whether to add the new information at the current moment to the memory unit, and the output gate determines which information in the memory unit to output. This mechanism enables the LSTM to effectively process long sequence data and predict the future trend of the device operating state;

[0097] Finally, through the time series analysis of the device operation data, the LSTM can detect abnormal trends in the device operating state. For example, if the temperature of the device continues to rise and exceeds the normal range within a period of time, the LSTM can predict this abnormal trend in advance and issue a warning signal;

[0098] The specific process for comprehensive fault diagnosis is as follows:

[0099] First, comprehensively analyze the fault mode recognition result obtained by the CNN and the abnormal trend prediction result obtained by the LSTM. For example, if the CNN identifies that there may be a fault in the motor, and at the same time the LSTM predicts that the current value of the motor will continue to rise abnormally in the future, then it can be more certain that the possibility of the motor having a fault is relatively high;

[0100] Then, based on the result of the comprehensive analysis, give the fault diagnosis result of the key components and the overall operating state of the device, including information such as whether there is a fault, the type of the fault, the severity of the fault, etc. If a fault is detected, the system can immediately take corresponding measures, such as alarm, adjusting the device operation parameters, or starting the standby device, etc.;

[0101] The mathematical formulas and algorithms involved are as follows:

[0102] 1. Convolutional Neural Network (CNN)

[0103] Convolution operation: Let the input data be X, the convolution kernel be W, and the output Y of the convolutional layer can be expressed as:

[0104]

[0105] Among them, i and j are the indices of the output feature map, m and n are the indices of the convolutional kernel, and b is the bias term;

[0106] Pooling operation (taking max pooling as an example): Let the input feature map be Y, the pooling window size be k×k, and the output Z of the pooling layer can be expressed as:

[0107]

[0108] 2. Long Short-Term Memory Network (LSTM)

[0109] Forget gate:

[0110] f t = σ(W f · [h t-1 , x t + b f );

[0111] Among them, f t is the output of the forget gate, σ is the sigmoid activation function, W f is the weight matrix of the forget gate, h t-1 is the hidden state at the previous moment, x t is the input at the current moment, and b f is the bias term of the forget gate;

[0112] Input gate:

[0113] i t = σ(W i · [h t-1 , x t + b i );

[0114] Among them, i t is the output of the input gate, W i is the weight matrix of the input gate, and b i is the bias term of the input gate;

[0115] Cell state update:

[0116]

[0117] Among them, is the candidate cell state, W C is the weight matrix of the cell state update, b C is the bias term of the cell state update, and ⊙ represents element-wise multiplication;

[0118] Output gate:

[0119] o t = σ(W o · [h t-1, x t +b p );

[0120] h t = o t ⊙tanh(C t );

[0121] Among them, o t is the output of the output gate, W o is the weight matrix of the output gate, b o is the bias term of the output gate, h t is the hidden state at the current moment.

[0122] Step 5: Build an edge computing architecture, use a priority scheduling algorithm to schedule real-time tasks, and apply dynamic load balancing technology to monitor the load of each node in real time;

[0123] Specifically, first, deploy edge computing nodes at the network edge close to the device or data source. These nodes have certain computing, storage, and network communication capabilities. Connect the edge computing nodes to the PLC system through a high-speed network to build an edge computing architecture; for example, in a factory workshop, install edge computing devices in the control cabinets near the production line and establish a high-speed communication connection with the PLC control cabinet responsible for controlling the production line equipment to ensure that data can be quickly transmitted and processed;

[0124] Then, use a priority scheduling algorithm (such as EarliestDeadlineFirst, EDF) to schedule real-time tasks. First, assign a priority to each real-time task. The determination of the priority is based on factors such as the deadline of the task and the importance of the task. For tasks with a tight deadline and a significant impact on the system operation, assign a higher priority; then, the scheduling algorithm schedules high-priority tasks for execution in the order of priority; in practical applications, when multiple tasks request resources, the scheduler will, according to the priority of the tasks, preferentially allocate computing resources to high-priority tasks to ensure that these tasks can be completed in a timely manner;

[0125] Finally, use a load monitoring tool to monitor the load conditions of each edge computing node in real time. The monitoring indicators include CPU usage, memory usage, network bandwidth occupancy, etc. When the load of a certain node is too high, the dynamic load balancing mechanism will be activated. It will, according to the preset algorithm, migrate some tasks to nodes with lower loads for execution; for example, adopt a dynamic load balancing algorithm based on task priority and node load. For tasks with a lower priority and a high load on the current execution node, migrate them to nodes with lower loads to achieve an even distribution of the load of each node and avoid the situation where the performance of the system is affected due to the overloading of some nodes.

[0126] Step 6: Based on the control strategy, control decision, and fault diagnosis results, dynamically adjust the operating parameters of the PLC control cabinet according to the operating status, load condition, and energy price information of the current equipment.

[0127] Specifically, the system first collects and integrates information such as control strategies, control decisions, and fault diagnosis results, and at the same time obtains the operating status, load condition, and energy price information of the current equipment. For example, it obtains the control strategy from the deep reinforcement learning module, the control decision from the fuzzy control adjustment module, and the fault diagnosis result from the fault prediction and diagnosis module; it monitors the equipment operating status and load condition in real time through sensors, and obtains the energy price information from the energy supply system or relevant platforms, and comprehensively analyzes this information to clarify the current actual situation and requirements of the equipment. If the fault diagnosis result shows that a certain component of the equipment has a potential fault risk, combined with the equipment operating status and load condition, it judges whether it is necessary to reduce the equipment operating power to avoid faults; according to the energy price information, it evaluates whether adjusting the equipment operating mode under the current load can achieve energy conservation and efficiency improvement.

[0128] Then, based on the above analysis results, it makes targeted adjustments to the operating parameters of the PLC control cabinet. If the control strategy and decision indicate that it is necessary to improve the equipment production efficiency, and the equipment operating status is good, the load is within the tolerable range, and the energy price is in the low valley period, the system will appropriately increase the motor speed, increase the valve opening, or raise the output voltage level to increase the production capacity of the equipment; on the contrary, if the equipment has potential faults, or the load is too high and may affect the equipment life, or the energy price is high, the system will reduce the motor speed, decrease the valve opening, or lower the output voltage level to ensure the safe operation of the equipment and reduce the energy consumption cost; when adjusting the motor speed, the system will accurately calculate the appropriate speed value according to the motor characteristic curve and the current load condition; when adjusting the valve opening, it will consider factors such as pipeline pressure, flow demand, and equipment operation stability; when changing the output voltage, it is necessary to ensure that all components of the equipment can work normally under the adjusted voltage and will not have a negative impact on the equipment performance.

[0129] After the operating parameters are adjusted, the system continuously monitors the operating status and performance indicators of the equipment, collects the feedback information after adjustment, observes whether the production efficiency of the equipment reaches the expectation, whether the potential faults are alleviated, whether the energy consumption is reduced, etc. According to the feedback information, it optimizes and improves the adjustment strategy. If it is found that although the adjusted motor speed reduces the energy consumption, it causes a significant drop in production efficiency, the system will re-evaluate and fine-tune the speed value to find a better balance between energy conservation and production efficiency. Through continuous feedback and optimization, the operating parameters of the PLC control cabinet are always in the best state, ensuring that the equipment can operate stably, efficiently, and economically under different working conditions and environments.

[0130] In one embodiment, specifically, in step three, the design of the fuzzy control rules adjusts the output based on the load change and temperature fluctuation;

[0131] The membership function of the fuzzy rules is calculated by the following formula:

[0132]

[0133] where, μ i (t) is the membership function of the i-th fuzzy rule; x i is the current input value; α i is the slope control parameter of the membership function, which determines the change speed of the membership degree; is the fuzzy center value of the input;

[0134] The final output is obtained by weighted average:

[0135]

[0136] where, u(t) is the output signal of the controller; c i is the output value corresponding to the fuzzy control rule i;

[0137] Fuse the above fuzzy control output result with the control strategy obtained in step two to generate the final control instruction;

[0138] The detailed steps are as follows:

[0139] 1. Fuzzy control rule design

[0140] Determine the input variables: Select the load change and temperature fluctuation as the input variables of the fuzzy controller. The load change reflects the increase or decrease of the current load of the device compared to the normal load, and the temperature fluctuation reflects the change range of the temperature at the key parts of the device. By monitoring and analyzing these two variables, the operating state and working environment of the device can be understood;

[0141] Divide the fuzzy subsets: Divide multiple fuzzy subsets for each input variable. For example, for the load change, it can be divided into fuzzy subsets such as "negative large", "negative small", "zero", "positive small", "positive large", etc.; for the temperature fluctuation, it can be divided into fuzzy subsets such as "substantial decrease", "slight decrease", "stable", "slight increase", "substantial increase", etc. Each fuzzy subset represents the value range of the input variable within a certain range;

[0142] Formulate the fuzzy rules: According to the operating characteristics and control experience of the device, formulate the fuzzy control rules based on the load change and temperature fluctuation; for example:

[0143] IF the load change is "positive large" AND the temperature fluctuation is "substantial increase" THEN the output is "substantially reduce the control amount";

[0144] IF the load change is "small negative" AND the temperature fluctuation is "stable", THEN the output is "slightly increase the control amount";

[0145] These rules describe the output actions that the controller should take under different input combinations;

[0146] 2. Calculation of membership functions of fuzzy rules

[0147] For each fuzzy rule, use the given formula:

[0148]

[0149] Calculate its membership function;

[0150] where x i is the current input value, such as the current load change value or temperature fluctuation value; α i is the slope control parameter, and different fuzzy subsets can set different α i values to adjust the change speed of the membership degree; is the fuzzy center value of the input, representing the center position of this fuzzy subset;

[0151] For example, for the "small positive" fuzzy subset of load change, assume its fuzzy center value is 5 (indicating a 5% increase in load), and the slope control parameter α i is 0.5; when the current load change value x i is 6, substituting into the formula can calculate the membership function value μ i (t) of this input value for the "small positive" fuzzy subset;

[0152] 3. Calculation of the final output

[0153] Calculate the output value c i corresponding to each fuzzy rule. These output values are predetermined according to the consequents of the fuzzy rules. For example, "substantially reduce the control amount", "slightly increase the control amount", etc. can correspond to specific numerical values.

[0154] Use the weighted average formula:

[0155]

[0156] Calculate the final controller output signal u(t);

[0157] where N is the total number of fuzzy rules; by multiplying the membership function value μ i (t) of each fuzzy rule with the corresponding output value c i and summing them up, the output result considering all rules is obtained;

[0158] 4. Control Strategy Fusion

[0159] The output result obtained from fuzzy control is fused with the control strategy obtained from deep reinforcement learning. Methods such as weighted average and priority selection can be used for fusion. For example, according to the actual application scenario, different weights are set for the fuzzy control output and the deep reinforcement learning control strategy respectively, and the results of both are weighted and averaged according to the weights to obtain the final control instruction;

[0160] The mathematical formulas and algorithm explanations are as follows:

[0161] Membership function formula:

[0162]

[0163] is an S-shaped membership function, also known as the sigmoid function. When , the membership function value is 0.5; when x i deviates from , the membership function value will gradually approach 0 or 1, specifically depending on the size of x i and the positive or negative of α i ; the larger the absolute value of α i , the steeper the change of the membership function, which means that when the input value is close to the fuzzy center value, the change speed of the membership degree is faster;

[0164] Final output calculation formula:

[0165]

[0166] The influence of all fuzzy rules is integrated through weighted average; the membership function value μ i (t) represents the activation degree of this rule under the current input situation, and the output value c i represents the control action corresponding to this rule. The product of the two and summation obtain the final controller output signal, which reflects the result of the joint action of all rules.

[0167] In one embodiment, specifically, in step two, the state space includes the temperature, pressure, rotational speed, load current, load voltage of the device during real-time operation, as well as the average load within the past 24 hours and the cumulative operation duration of the device; the action space is to control the start / stop of the motor, the level of adjusting the valve opening, and the gear of adjusting the output voltage;

[0168] Among them, the state space covers various aspects of information such as the temperature, pressure, rotational speed, load current, load voltage of the device during real-time operation, as well as the average load within the past 24 hours and the cumulative operation duration of the device. These parameters comprehensively and accurately reflect the current operating condition and historical operating situation of the device. Real-time operating parameters can directly show the current working state of the device. For example, too high a temperature may mean potential faults or excessive load in the device; changes in pressure, rotational speed, load current, and voltage can also reflect the operating stability and working efficiency of the device; while the average load within the past 24 hours and the cumulative operation duration of the device provide historical trend information on the device's operation. The average load can help determine whether the recent working intensity of the device is normal, and the cumulative operation duration helps evaluate the wear and aging conditions of the device. By incorporating this information into the state space, the deep reinforcement learning algorithm can comprehensively consider various state factors of the device, make more scientific and reasonable decisions, and improve the accuracy and adaptability of the control strategy.

[0169] The action space includes operations such as starting / stopping the control motor, adjusting the valve opening level, and adjusting the output voltage gear. These actions are key control means that directly affect the operating state of the device. Controlling the start and stop of the motor can reasonably allocate device resources according to production requirements, avoiding energy waste and device wear caused by the device running when unnecessary; adjusting the valve opening level can accurately control the flow rate and pressure of the fluid, meet the requirements of different production processes for flow rate and pressure, and ensure the stability of the production process and product quality; adjusting the output voltage gear can provide appropriate power supply for the device according to the change of the device load, which can not only ensure the normal operation of the device but also improve energy utilization efficiency and avoid damage to the device caused by too high or too low voltage. Incorporating these operations into the action space enables the deep reinforcement learning algorithm to select the most appropriate action according to the information in the state space, achieve effective control of the device, and optimize the operating performance of the device.

[0170] The state space and the action space are interrelated and interact with each other. The information in the state space provides a basis for the decision-making in the action space. The deep reinforcement learning algorithm analyzes and learns the information in the state space and selects the optimal action to adjust the operating state of the device, enabling the device to develop towards the desired target state. For example, when parameters such as temperature and load current in the state space indicate that the device load is too high, the algorithm may select actions such as reducing the motor speed, decreasing the valve opening, or lowering the output voltage gear from the action space to reduce the device load and ensure the stable operation of the device; after the actions in the action space are executed, they will change the operating state of the device, thus updating the information in the state space, forming a closed-loop adaptive control process. This collaborative relationship enables the system to continuously adapt to various changes during the device operation, continuously optimize the control strategy, and improve the overall performance and stability of the system.

[0171] In one embodiment, specifically, in step one, the real-time data of the device operation is obtained through real-time acquisition by a temperature sensor, a humidity sensor, a load current sensor, and a load voltage sensor. The historical data is read from a local database. The preprocessing includes data filtering and normalization. The real-time data is collected by the temperature sensor, the humidity sensor, the load current sensor, and the load voltage sensor. These sensors are like the "antennae" of the system and can accurately capture the key information of the device operation state. The temperature sensor monitors the temperature of the key parts of the device, and can timely detect potential overheating hazards of the device to prevent device failures caused by too high temperature. The humidity sensor is crucial for some devices sensitive to environmental humidity and can ensure the stable operation of the device in a suitable humidity environment. The load current and voltage sensors directly reflect the power consumption of the device. Abnormal fluctuations in current and voltage indicate that the device may have electrical faults or abnormal load changes, providing a basis for the system to adjust the control strategy and enabling the system to quickly respond to changes in the device operation state.

[0172] The historical data is read from the local database. This part of the data is the "experience library" of the long-term operation of the device. The database stores the operation data of the device under different working conditions and time periods, including data under normal operation and fault states. When performing fault diagnosis and prediction, the historical data plays a significant role. By comparing the current real-time data with the historical normal data, subtle changes in the device operation trend can be found, and potential faults can be detected in advance. When optimizing the control strategy, analyzing the control effects under different working conditions in the historical data provides a reference for the system to select the optimal control strategy under similar working conditions, improving the adaptive ability of the system.

[0173] The collected real-time data and the read historical data are subjected to data filtering and normalization processing. Data filtering removes noise interference to ensure data authenticity. In an industrial environment, the data collected by sensors is easily affected by electromagnetic interference, mechanical vibration, etc. and generates noise. Without filtering, the noise will interfere with subsequent analysis and lead to decision-making errors. Normalization processing unifies data with different ranges and dimensions into the same interval, facilitating model processing and comparison. The data collected by different sensors, such as temperature and voltage, have large differences in numerical range and dimension. Normalization processing eliminates this difference, making the data comparable in model training and analysis, improving the algorithm efficiency and accuracy, and laying a good data foundation for subsequent operations such as deep reinforcement learning, fuzzy control, and fault diagnosis.

[0174] In one embodiment, specifically, in step four, the input layer of the convolutional neural network receives the real-time data and historical data of the device operation, and performs feature extraction and fault mode recognition through the convolutional layer, pooling layer, and fully connected layer. The long short-term memory network receives the time series of the device operation data and performs time series analysis and abnormal trend prediction through the forget gate, input gate, and output gate.

[0175] Specifically, the input layer of the convolutional neural network receives the real-time operation data and historical data of the device after preprocessing such as filtering and normalization. The data undergoes a convolution operation through a convolution kernel in the convolutional layer to extract local features. The pooling layer downsamples the output of the convolutional layer to reduce the data dimension and retain important features. The fully connected layer integrates the features and maps them to different fault mode categories through an activation function to achieve the recognition of the fault modes of the key components of the device;

[0176] The long short-term memory network receives the time series of the device operation data. It determines whether to discard the information of the previous moment through the forget gate, decides whether to add new information of the current moment through the input gate, and determines whether to output the information in the memory unit through the output gate, so as to perform time series analysis and predict the abnormal trend of the device operation status;

[0177] Based on the fault mode recognition result obtained by the convolutional neural network and the abnormal trend prediction result obtained by the long short-term memory network, it is judged whether the device has a fault, the type of the fault, and the severity of the fault, and the final fault diagnosis result is output;

[0178] Among them, CNN is good at extracting local features of data and can effectively identify the fault modes of devices; LSTM can capture the time series information of data and predict the future trend of the device operation status. Combining the two can give full play to their advantages and improve the accuracy of fault prediction and diagnosis.

[0179] In one embodiment, specifically, in step five, the priority scheduling algorithm adopts the EarliestDeadlineFirs algorithm to determine the priority of tasks according to the deadline of the tasks;

[0180] The EarliestDeadlineFirst (EDF) algorithm is a real-time scheduling algorithm based on the deadline of tasks. Its core idea is to give priority to scheduling the task with the earliest deadline for execution. Let the task set be T = {T 1 , T 2 , …, T n}, and each task T i has a deadline d i . At each scheduling moment t, the algorithm selects the task T j with the earliest deadline for execution, that is:

[0181]

[0182] Among them, "ready" means that the task has arrived and has not been completed.

[0183] In one embodiment, specifically, in step six, dynamically adjusting the operation parameters of the PLC control cabinet includes adjusting the speed of the motor, regulating the opening of the valve, and changing the magnitude of the output voltage;

[0184] Among them, adjust the speed of the motor: Adjusting the motor speed requires considering multiple factors. If the equipment load is light, the production efficiency requirement is not high, and the energy price is at a high level, in order to reduce energy consumption, the motor speed can be appropriately reduced; through precise calculation, combined with the characteristic curve of the motor and the current load situation, determine the appropriate speed value; for example, for a motor driving a water pump, according to the flow-head curve of the water pump and the current water supply demand, calculate the motor speed that can not only meet the water supply requirements but also reduce energy consumption.

[0185] Adjust the opening of the valve: The adjustment of the valve opening should be combined with the pipeline pressure, flow demand, and equipment operation stability. When it is necessary to increase the pipeline flow, appropriately increase the valve opening; if the pipeline pressure is too high, in order to ensure the safe and stable operation of the equipment, reduce the valve opening; for example, in chemical production, according to the progress of the chemical reaction and the requirement for the raw material flow, precisely adjust the valve opening to ensure the smooth progress of the reaction.

[0186] Change the magnitude of the output voltage: The change of the output voltage is determined based on the power demand, operating characteristics of the equipment, and energy price. For some voltage-sensitive equipment, a stable and appropriate voltage needs to be provided. When the energy price is low and the equipment load is large, the output voltage can be appropriately increased to ensure the normal operation of the equipment; on the contrary, when the energy price is high and the load is small, reduce the output voltage to save energy; for example, for some high-precision electronic equipment, the output voltage needs to be precisely adjusted according to its rated voltage and actual working conditions.

[0187] Embodiment 2:

[0188] Figure 2 is a schematic diagram of the functional modules of an adaptive intelligent control system of a PLC control cabinet according to an embodiment of the present application. As Figure 2 shown, an adaptive intelligent control system of a PLC control cabinet, the system includes: a data acquisition and preprocessing module, a deep reinforcement learning control module, a fuzzy control adjustment module, a fault prediction and diagnosis module, an edge computing and task scheduling module, and a parameter adjustment module;

[0189] The data acquisition and preprocessing module is configured to acquire real-time data and historical data of equipment operation, and perform preprocessing using edge computing devices, and transmit the preprocessed real-time data to the PLC system in real time based on the IoT communication protocol;

[0190] The deep reinforcement learning control module is configured to perform adaptive control based on the real-time data and historical data, optimize the control strategy based on the reward function, state space, and action space of deep reinforcement learning;

[0191] The fuzzy control adjustment module is configured to perform fuzzification processing on the input real-time data based on the intelligent adjustment mechanism of fuzzy control, and make output control decisions according to the fuzzy control rules;

[0192] The fault prediction and diagnosis module is configured to perform fault prediction and intelligent diagnosis on the key components and overall operating status of the device based on the convolutional neural network and long short-term memory network according to the real-time data and historical data of the device operation, and obtain the fault diagnosis result;

[0193] The edge computing and task scheduling module is configured to build an edge computing architecture, schedule real-time tasks using a priority scheduling algorithm, and apply dynamic load balancing technology to monitor the load of each node in real time;

[0194] The parameter adjustment module is configured to dynamically adjust the operating parameters of the PLC control cabinet based on the control strategy, control decision, and fault diagnosis result, according to the current operating status, load condition, and energy price information of the device.

[0195] In one embodiment, specifically, the data acquisition and preprocessing module is also configured with a data encryption and verification mechanism.

[0196] In one embodiment, specifically, the deep reinforcement learning control module adopts a distributed training architecture and uses multiple edge computing devices for parallel computing to accelerate the training process.

[0197] For other details of the implementation technical solutions of each module in the adaptive intelligent control system of a PLC control cabinet in the above embodiments, reference can be made to the description in the adaptive intelligent control method and system of a PLC control cabinet in the above embodiments, which will not be elaborated here.

[0198] It should be noted that each embodiment in this specification is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other. For system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiments.

[0199] Embodiment 3:

[0200] Problems in the prior art: PLC control cabinets are usually installed in various industrial environments, including outdoors, workshops, or factory buildings. In an environment with high temperature or strong light, the opening degree of the ventilation openings of the PLC control cabinet is large, and the temperature inside the control cabinet will rise rapidly, causing the PLC and other electronic components to overheat, thus affecting their performance and lifespan. In rainy or snowy weather, the high environmental humidity will have an adverse impact on the internal electrical components.

[0201] Such as Figures 3 - 6As shown in the figure, an adaptive intelligent control device for a PLC control cabinet is provided in an embodiment of the present invention, including a main body component 10, and the main body component 10 includes a cabinet body 11, a cabinet door 12, a control panel 13, a first temperature sensor 14, a light sensor 15, a rain and snow sensor 16, and a second temperature sensor 17;

[0202] The cabinet door 12 is hinged to the front surface of the cabinet body 11, a PLC control system is installed inside the cabinet body 11, the control panel 13 is installed on the front surface of the cabinet door 12, the first temperature sensor 14, the light sensor 15, and the rain and snow sensor 16 are installed on the top of the cabinet body 11, the second temperature sensor 17 is installed on the top of the inner side wall of the cabinet body 11, and the signal output ends of the first temperature sensor 14, the light sensor 15, the rain and snow sensor 16, and the second temperature sensor 17 are signal-connected to the signal input end of the control panel 13. The internal and external environmental temperatures of the device are monitored by the first temperature sensor 14 and the second temperature sensor 17 respectively, the environmental light is monitored by the light sensor 15, the environmental rain and snow weather is monitored by the rain and snow sensor 16, and the signals are fed back to the control panel 13, and the control panel 13 controls the device to complete intelligent adjustment;

[0203] Ventilation components 20 and heat dissipation components 30 are arranged on both sides of the main body component 10.

[0204] In this embodiment, specifically: the ventilation component 20 includes a ventilation opening 21 and a transmission box 22;

[0205] Ventilation openings 21 are symmetrically arranged on both sides of the cabinet body 11, and a transmission box 22 is fixedly connected to one side of the ventilation opening 21, and the transmission box 22 is used to protect the internal transmission mechanism.

[0206] In this embodiment, specifically: a driving motor 23 is installed on one side of the transmission box 22.

[0207] In this embodiment, specifically: transmission wheels 24 are uniformly and automatically connected to the inner side wall of the transmission box 22, and the output shaft of the driving motor 23 is fixedly connected to one end of the transmission wheel 24. When the driving motor 23 works, it drives the transmission wheel 24 to rotate.

[0208] In this embodiment, specifically: a transmission chain 25 is sleeved between adjacent transmission wheels 24, and multiple transmission wheels 24 are driven by the transmission chain 25 to rotate synchronously.

[0209] In this embodiment, specifically: a fixed shaft 26 is fixedly connected to one side of the transmission wheel 24, the fixed shaft 26 is rotatably connected to the inner side wall of the ventilation opening 21, and a shutter 27 is fixedly connected to the outer side wall of the fixed shaft 26. Multiple fixed shafts 26 rotate synchronously, driving the shutter 27 to rotate, so as to adjust the opening degree of the ventilation opening 21.

[0210] In this embodiment, specifically: the heat dissipation component 30 includes a mounting base 31 and a fan 32;

[0211] The mounting bases 31 are symmetrically and fixedly connected to both sides of the cabinet body 11, and the fans 32 are evenly mounted on the mounting bases 31. By the operation of the fans 32, external air is introduced into the device to accelerate the heat dissipation efficiency of the device.

[0212] In this embodiment, specifically: electric guide rails 33 are symmetrically mounted on one side of the mounting base 31, and a baffle 34 is slidably connected to one side of the electric guide rail 33. The baffle 34 is slidably connected to the inner side wall of the mounting base 31. By driving the baffle 34 to move through the electric guide rail 33, the air intake channel of the fan 32 is closed.

[0213] When the present invention is working: the internal and external environmental temperatures of the device are respectively monitored by the first temperature sensor 14 and the second temperature sensor 17, and the signals are fed back to the control panel 13. When the temperature inside the device is relatively high, the control cabinet is cooled by the operation of the fan 32. The ambient light is monitored by the light sensor 15. When the external environment is high temperature or strong light, the motor 23 is driven to work, driving the transmission wheel 24 to rotate. The transmission wheels 24 are driven by a transmission chain 25 to rotate synchronously, and the plurality of fixed shafts 26 rotate synchronously, driving the louvers 27 to rotate, automatically reducing the area of the ventilation holes. The ambient rain and snow weather is monitored by the rain and snow sensor 16. In rainy and snowy weather, the ventilation opening 21 is closed, and at the same time, the electric guide rail 33 drives the baffle 34 to move, and the air intake channel of the fan 32, thereby reducing the adverse effects of the external environment on the PLC and other electronic components. The ventilation component 20 and the heat dissipation component 30 are controlled by the control panel 13 to be adjusted to meet the adaptive intelligent control of the PLC control cabinet, and improve the performance and service life of the PLC and other electronic components.

[0214] Due to the adoption of the above technical solutions in the embodiment of the present invention, it has the following advantages:

[0215] First, the present invention monitors the internal and external environmental temperatures of the device through temperature sensors, monitors the ambient light through a light sensor, monitors the ambient rain and snow weather through a rain and snow sensor, and uses the control panel to control the ventilation component and the heat dissipation component to be adjusted to meet the adaptive intelligent control of the PLC control cabinet, and improve the performance and service life of the PLC and other electronic components.

[0216] Second, when the temperature inside the device is relatively high, the control cabinet is cooled by the operation of the fan. When the external environment is high temperature or strong light, the motor is driven to work to drive the louvers to rotate, automatically reducing the area of the ventilation holes. In rainy and snowy weather, the ventilation opening and the air intake channel of the fan are closed, thereby reducing the adverse effects of the external environment on the PLC and other electronic components.

[0217] The basic principles of the present disclosure have been described in conjunction with specific embodiments. However, it should be noted that the advantages, benefits, effects, etc. mentioned in the present disclosure are only examples and not limitations. It cannot be considered that these advantages, benefits, effects, etc. are essential for each embodiment of the present disclosure. Additionally, the specific details disclosed above are only for illustrative purposes and for ease of understanding, and are not limitations. The above details do not limit the present disclosure to necessarily implementing with the above specific details.

[0218] In the present disclosure, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The block diagrams of the devices, apparatuses, equipment, and systems involved in the present disclosure are only illustrative examples and do not intend to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open-ended words, meaning "including but not limited to", and can be used interchangeably with each other. The word "or" and "and" used herein refer to the word "and / or", and can be used interchangeably with each other, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to", and can be used interchangeably with each other.

[0219] In addition, as used herein, the "or" used in the listing of items starting with "at least one" indicates a separate listing, so that for example, the listing of "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the term "exemplary" does not mean that the described examples are preferred or better than other examples.

[0220] It should also be noted that in the systems and methods of the present disclosure, each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent solutions of the present disclosure.

[0221] Various changes, substitutions, and alterations to the technologies described herein can be made without departing from the teachings of the technology defined by the appended claims. Additionally, the scope of the claims of the present disclosure is not limited to the specific aspects of the processes, machines, manufactures, compositions of events, means, methods, and acts described above. Current or later-developed processes, machines, manufactures, compositions of events, means, methods, or acts that perform substantially the same function or achieve substantially the same result as the corresponding aspects described herein can be utilized. Thus, the appended claims include such processes, machines, manufactures, compositions of events, means, methods, or acts within their scope.

[0222] The foregoing description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of the present disclosure. Thus, the present disclosure is not intended to be limited to the aspects shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein. The foregoing description has been presented for purposes of illustration and description. In addition, the description is not intended to limit embodiments of the present disclosure to the forms disclosed herein. Although numerous example aspects and embodiments have been discussed above, those skilled in the art will recognize some variations, modifications, alterations, additions, and subcombinations thereof.

Claims

1. An adaptive intelligent control method for a PLC control cabinet, characterized in that: The following steps are involved: Obtain the real-time and historical data of equipment operation and pre-process them using edge computing devices, and transmit the pre-processed real-time and historical data to the PLC system in real time; According to real-time data and historical data, adaptive control is performed based on the reward function, state space and action space of deep reinforcement learning to optimize the control strategy; The intelligent adjustment mechanism based on fuzzy control performs fuzzy processing on the input real-time data, makes output control decisions according to fuzzy control rules, and integrates control strategies to generate final control instructions; According to the real-time and historical data of equipment operation, the fault prediction and intelligent diagnosis of key components and overall operation status of the equipment are carried out based on convolutional neural networks and long short-term memory networks to obtain fault diagnosis results; Build an edge computing architecture, use a priority scheduling algorithm to schedule real-time tasks, and use dynamic load balancing technology to monitor the load of each node in real time; Based on the control strategy, control decision and fault diagnosis results, the operating parameters of the PLC control cabinet are dynamically adjusted according to the current equipment operating status, load conditions and energy price information.

2. The adaptive intelligent control method of a PLC control cabinet according to claim 1, characterized in that: The intelligent adjustment mechanism based on fuzzy control performs fuzzy processing on the input real-time data, makes output control decisions according to fuzzy control rules, and integrates control strategies to generate final control instructions, including the fuzzy control rules, to adjust the output based on load changes and temperature fluctuations; The membership function of the fuzzy rule is calculated by the following formula: Among them, μ i (t) is the membership function of the i-th fuzzy rule; x i is the current input value; α i is the slope control parameter of the membership function, which determines the changing speed of the membership; is the fuzzy center value of the input; The final output is obtained by weighted averaging: Where, u(t) is the controller output signal; c i is the output value corresponding to fuzzy control rule i.

3. The adaptive intelligent control method of a PLC control cabinet according to claim 1, characterized in that: The control strategy is optimized by adaptive control based on the reward function of deep reinforcement learning as well as the state space and action space according to real-time data and historical data. The state space includes the real-time operating temperature, pressure, speed, load current and load voltage of the equipment, as well as the average load in the past 24 hours and the cumulative operating time of the equipment; the action space is for controlling the start / stop of the motor, adjusting the level of valve opening and adjusting the gear of the output voltage.

4. The adaptive intelligent control method of a PLC control cabinet according to claim 1 is characterized in that: The real-time data and historical data of the equipment operation are acquired and preprocessed by edge computing equipment, and the preprocessed real-time data and historical data are transmitted to the PLC system in real time, including real-time acquisition of the real-time data of the equipment operation through temperature sensors, humidity sensors, load current sensors and load voltage sensors, and the historical data are read from a local database. The preprocessing includes data filtering and normalization processing.

5. The adaptive intelligent control method of a PLC control cabinet according to claim 1, characterized in that: According to the real-time data and historical data of equipment operation, fault prediction and intelligent diagnosis of key components of the equipment and the overall operation status are performed based on convolutional neural networks and long short-term memory networks to obtain fault diagnosis results, including the input layer of the convolutional neural network receiving the real-time data and historical data of equipment operation, and performing feature extraction and fault pattern recognition through convolutional layers, pooling layers and fully connected layers; the long short-term memory network receives the time series of equipment operation data, and performs time series analysis and abnormal trend prediction through forgetting gates, input gates and output gates.

6. The adaptive intelligent control method of a PLC control cabinet according to claim 1, characterized in that: The edge computing architecture is built, a priority scheduling algorithm is used to schedule real-time tasks, and dynamic load balancing technology is used to monitor the load of each node in real time, including the priority scheduling algorithm using the EarliestDeadlineFirs algorithm to determine the priority of the task according to the deadline of the task.

7. The adaptive intelligent control method of a PLC control cabinet according to claim 1, characterized in that: Based on the control strategy, control decision and fault diagnosis results, the operating parameters of the PLC control cabinet are dynamically adjusted according to the current operating status of the equipment, load conditions and energy price information, wherein the dynamic adjustment of the operating parameters of the PLC control cabinet includes adjusting the speed of the motor, adjusting the opening of the valve and changing the output voltage.

8. An adaptive intelligent control system for a PLC control cabinet, applied to an adaptive intelligent control method for a PLC control cabinet according to any one of claims 1 to 7, characterized in that: The system includes: a data acquisition and preprocessing module, a deep reinforcement learning control module, a fuzzy control adjustment module, a fault prediction and diagnosis module, an edge computing and task scheduling module and a parameter adjustment module; The data acquisition and preprocessing module is configured to acquire real-time data and historical data of the device operation, and preprocess the data using the edge computing device, and transmit the preprocessed real-time data to the PLC system in real time based on the IoT communication protocol; The deep reinforcement learning control module is configured to perform adaptive control and optimize the control strategy based on the reward function, state space and action space of deep reinforcement learning according to real-time data and historical data; The fuzzy control adjustment module is configured to perform fuzzy processing on input real-time data based on the intelligent adjustment mechanism of fuzzy control and make output control decisions according to fuzzy control rules; The fault prediction and diagnosis module is configured to perform fault prediction and intelligent diagnosis on key components and overall operation status of the equipment based on real-time data and historical data of equipment operation, and obtain fault diagnosis results based on convolutional neural networks and long short-term memory networks; The edge computing and task scheduling module is configured to build an edge computing architecture, use a priority scheduling algorithm to schedule real-time tasks, and use dynamic load balancing technology to monitor the load of each node in real time; The parameter adjustment module is configured to dynamically adjust the operating parameters of the PLC control cabinet based on the control strategy, control decision and fault diagnosis results, according to the current equipment operating status, load conditions and energy price information.

9. The adaptive intelligent control system of a PLC control cabinet according to claim 8, characterized in that: The data acquisition and preprocessing module is also configured with a data encryption and verification mechanism.

10. The adaptive intelligent control system of a PLC control cabinet according to claim 8, characterized in that: The deep reinforcement learning control module adopts a distributed training architecture and uses parallel computing of multiple edge computing devices to accelerate the training process.

Images