Intelligent controller based on artificial intelligence and control method thereof

Abstract

The invention discloses an intelligent controller based on artificial intelligence and a control method thereof, and belongs to the technical field of artificial intelligence. The intelligent controller comprises a data acquisition module, a data preprocessing module, an AI processing module, a decision execution module, a feedback adjustment module and a fault prediction and diagnosis module; according to the invention, the feedback adjustment module continuously monitors the deviation between the actual operation state of the system and the set target through a closed-loop control mechanism, and adjusts the model parameters of the AI processing module, so that the controller can adapt to the change of the system characteristics, maintain good control performance, and improve the reliability of the system. The fault prediction and diagnosis module can recognize an abnormal mode in advance and trigger early warning, the fault handling capacity of the system is enhanced, the AI processing module conducts feature extraction, mode recognition and trend prediction on complex data through an advanced machine learning algorithm, more complex control tasks can be processed, higher-level intelligent decision making is achieved, and the system reliability is improved. And the stability and the reliability of the whole system are improved.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology, and specifically refers to an intelligent controller based on artificial intelligence and a control method thereof. Background Art

[0002] In recent years, artificial intelligence technology has made significant breakthroughs in machine learning, knowledge graphs, natural language processing, computer vision and other fields, providing new ideas for solving the limitations of traditional controllers. AI can learn and analyze large amounts of data, extract the laws hidden in complex systems, adaptively adjust control strategies, and even predict the future state of the system, thereby significantly improving the intelligence level of the control system;

[0003] However, the existing intelligent controller technology still has certain defects. The existing intelligent controller is limited by the data acquisition efficiency and the complexity of processing capabilities. The response to system state changes is not fast enough. The generation of control instructions may be delayed or the accuracy is insufficient, resulting in poor system dynamic performance. There is a lack of effective online adjustment mechanism, and its control performance may gradually deviate from the optimal state. It relies on predefined control rules or relatively simple mathematical models to handle control tasks. For complex, nonlinear or highly coupled system control problems, it is difficult to achieve refined and intelligent control decisions. It is relatively weak in fault prediction and diagnosis, and can often only respond passively after a fault occurs. There is a lack of effective early warning mechanism. Therefore, an intelligent controller based on artificial intelligence and a control method thereof are proposed. Summary of the invention

[0004] The object of the present invention is to provide an intelligent controller based on artificial intelligence and a control method thereof to solve the problems raised in the above background technology.

[0005] To achieve the above-mentioned object, the present invention provides the following technical solutions: an intelligent controller based on artificial intelligence and a control method thereof, comprising a data acquisition module, a data preprocessing module, an AI processing module, a decision execution module, a feedback adjustment module and a fault prediction and diagnosis module;

[0006] The data acquisition module is responsible for acquiring various physical quantities, state parameters, and environmental condition data in real time from sensors and other monitoring devices of the controlled system;

[0007] Wherein, the data preprocessing module is responsible for receiving the data transmitted by the data acquisition module for data preprocessing;

[0008] The AI ​​processing module processes the data by extracting features from the data, and performs advanced processing such as feature extraction, pattern recognition, and trend prediction on the collected data through a preset control strategy to generate control instructions;

[0009] The decision execution module receives the control instruction output by the AI ​​processing module and converts it into a specific control signal;

[0010] The feedback adjustment module compares the actual system response with the expected target by monitoring the control effect, and dynamically adjusts and optimizes the AI ​​processing module through a closed-loop control system;

[0011] Among them, the fault prediction and diagnosis module uses AI to conduct in-depth analysis of system operation data, identify abnormal patterns, achieve early fault warning, and locate the cause of the fault through reasoning mechanism.

[0012] The data acquisition module obtains system status information in real time by connecting to various sensors, sets the sampling frequency, and reads data from the sensors at preset time intervals.

[0013] Among them, the data preprocessing module performs preliminary verification, format conversion, and noise filtering on the collected raw data, encapsulates the preprocessed data into a format recognizable by the controller AI processing module, and transmits it to the AI ​​processing module through an internal interface for further analysis.

[0014] The AI ​​processing module constructs an initial model structure according to the control task and system characteristics, performs feature extraction engineering processing on the input data, and generates feature vectors required by the model, including numerical feature standardization, normalization, discrete feature encoding, and time series feature extraction;

[0015] Through the processed data, the model training divides the data set into training set, validation set and test set, sets the initial parameters of the model, uses the training set data to iteratively update the model parameters in the gradient descent until the preset number of training rounds is reached, evaluates the model performance on the validation set, calculates the accuracy and loss function value indicators, adjusts the model structure and parameter training strategy according to the evaluation results, encapsulates the trained model into a callable module inside the controller, and interfaces with the data acquisition module and the decision execution module;

[0016] The logistic regression model construction formula is as follows:

[0017]

[0018] Among them, the gradient descent model training formula is as follows:

[0019] θ t+1 =θ t -η▽ θ J(θ t ),

[0020] In the formula, η is the learning rate, J(θ t ) is expressed as the loss function, ▽θ J(θ t ), expressed as the gradient of the loss function parameter θ.

[0021] Among them, the decision execution module receives the control instructions processed by AI, calls the corresponding control algorithm or driver according to the analysis results, converts the instructions into specific physical control actions, interacts with the actuator of the controlled system in real time through the communication protocol, sends control signals, receives feedback information from the actuator, and confirms whether the control instructions have been executed correctly.

[0022] Among them, the feedback adjustment module sets the expected target state according to the control task, continuously monitors the actual operating state of the controlled system through the data acquisition module, compares it with the set target, calculates the deviation between the actual state and the target state, and generates a feedback signal based on the control error and the control algorithm of the controller. The feedback signal is sent back to the AI ​​processing module to trigger the model parameter adjustment and update the control strategy of the controller.

[0023] Among them, the fault prediction and diagnosis module trains an anomaly detection model based on historical fault data to identify abnormal patterns in system operation in real time. During operation, it continuously receives system status data and performs real-time analysis through the anomaly detection model. Once an anomaly is found, it immediately triggers a warning signal, conducts in-depth analysis of the abnormal data, extracts feature information related to the fault, and uses a pre-trained fault diagnosis model in combination with fault characteristics and a knowledge base to predict the type and cause of the fault.

[0024] Among them, an intelligent controller based on artificial intelligence and its control method, the specific implementation steps are as follows:

[0025] S1, real-time collection of various sensor data of the controlled system;

[0026] S2, perform pre-processing operations such as cleaning, standardization, and noise reduction on the collected raw data;

[0027] S3, automatically extract key features that reflect the system status through AI algorithms;

[0028] S4. Select a suitable AI model based on the characteristics of the system and the control task, and train it using simulated data;

[0029] S5. During operation, the AI ​​processing module receives newly collected data in real time, generates optimal control instructions through model calculation, and acts on the controlled system through the decision execution module;

[0030] S6, continuously update model parameters based on system operation feedback, and the controller can quickly adjust the control strategy when system characteristics change;

[0031] S7. Use AI models to monitor system operation data in real time, identify potential failure modes through anomaly detection algorithms, and predict failure types and causes based on the failure knowledge base.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] 1. The present invention acquires a large amount of system status information in real time through the data acquisition module. The AI ​​processing module can quickly process this data and generate accurate control instructions, so that the controller can respond immediately to system changes and improve the dynamic response speed and accuracy of the control system;

[0034] 2. The present invention continuously monitors the deviation between the actual operating state of the system and the set target through a closed-loop control mechanism through the feedback adjustment module, and adjusts the model parameters of the AI ​​processing module accordingly, so that the controller can adapt to changes in system characteristics and maintain good control performance. The fault prediction and diagnosis module can identify abnormal modes in advance and trigger early warnings, thereby enhancing the system's fault response capabilities and improving the stability and reliability of the overall system;

[0035] 3. The present invention uses advanced machine learning algorithms to extract features, identify patterns and predict trends of complex data through the AI ​​processing module, which goes beyond the control method of traditional controllers based on fixed rules or simple mathematical models, and can handle more complex control tasks and achieve higher-level intelligent decision-making;

[0036] 4. The present invention trains an anomaly detection model based on historical fault data through a fault prediction and diagnosis module, which can monitor the system operation status in real time and discover potential faults in a timely manner. It can not only provide early warning and reduce the impact of sudden faults on the system, but also locate the cause of the fault through a fault diagnosis model, guide maintenance personnel to quickly troubleshoot and repair problems, and shorten downtime and maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a structural schematic diagram of an intelligent controller based on artificial intelligence and a control method thereof of the present invention;

[0038] Figure 2 The present invention is an operation flow chart of an intelligent controller based on artificial intelligence and a control method thereof. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] Example

[0041] See also Figure 1-2 As shown, the present invention provides a technical solution: including a data acquisition module, a data preprocessing module, an AI processing module, a decision execution module, a feedback adjustment module and a fault prediction and diagnosis module;

[0042] The data acquisition module is responsible for acquiring various physical quantities, state parameters, and environmental condition data in real time from sensors and other monitoring devices of the controlled system;

[0043] Wherein, the data preprocessing module is responsible for receiving the data transmitted by the data acquisition module for data preprocessing;

[0044] The AI ​​processing module processes the data by extracting features from the data, and performs advanced processing such as feature extraction, pattern recognition, and trend prediction on the collected data through a preset control strategy to generate control instructions;

[0045] The decision execution module receives the control instruction output by the AI ​​processing module and converts it into a specific control signal;

[0046] The feedback adjustment module compares the actual system response with the expected target by monitoring the control effect, and dynamically adjusts and optimizes the AI ​​processing module through a closed-loop control system;

[0047] Among them, the fault prediction and diagnosis module uses AI to conduct in-depth analysis of system operation data, identify abnormal patterns, achieve early fault warning, and locate the cause of the fault through reasoning mechanism.

[0048] The data acquisition module obtains system status information in real time by connecting to various sensors, sets the sampling frequency, and reads data from the sensors at preset time intervals.

[0049] Among them, the data preprocessing module performs preliminary verification, format conversion, and noise filtering on the collected raw data, encapsulates the preprocessed data into a format recognizable by the controller AI processing module, and transmits it to the AI ​​processing module through an internal interface for further analysis.

[0050] The AI ​​processing module constructs an initial model structure according to the control task and system characteristics, performs feature extraction engineering processing on the input data, and generates feature vectors required by the model, including numerical feature standardization, normalization, discrete feature encoding, and time series feature extraction;

[0051] Through the processed data, the model training divides the data set into training set, validation set and test set, sets the initial parameters of the model, uses the training set data to iteratively update the model parameters in the gradient descent until the preset number of training rounds is reached, evaluates the model performance on the validation set, calculates the accuracy and loss function value indicators, adjusts the model structure and parameter training strategy according to the evaluation results, encapsulates the trained model into a callable module inside the controller, and interfaces with the data acquisition module and the decision execution module;

[0052] The logistic regression model construction formula is as follows:

[0053]

[0054] Among them, the gradient descent model training formula is as follows:

[0055] θ t+1 =θ t -η▽ θ J(θ t ),

[0056] In the formula, η is the learning rate, J(θ t ) is expressed as the loss function, ▽ θ J(θ t ), expressed as the gradient of the loss function parameter θ.

[0057] Among them, the decision execution module receives the control instructions processed by AI, calls the corresponding control algorithm or driver according to the analysis results, converts the instructions into specific physical control actions, interacts with the actuator of the controlled system in real time through the communication protocol, sends control signals, receives feedback information from the actuator, and confirms whether the control instructions have been executed correctly.

[0058] Among them, the feedback adjustment module sets the expected target state according to the control task, continuously monitors the actual operating state of the controlled system through the data acquisition module, compares it with the set target, calculates the deviation between the actual state and the target state, and generates a feedback signal based on the control error and the control algorithm of the controller. The feedback signal is sent back to the AI ​​processing module to trigger the model parameter adjustment and update the control strategy of the controller.

[0059] Among them, the fault prediction and diagnosis module trains an anomaly detection model based on historical fault data to identify abnormal patterns in system operation in real time. During operation, it continuously receives system status data and performs real-time analysis through the anomaly detection model. Once an anomaly is found, it immediately triggers a warning signal, conducts in-depth analysis of the abnormal data, extracts feature information related to the fault, and uses a pre-trained fault diagnosis model in combination with fault characteristics and a knowledge base to predict the type and cause of the fault.

[0060] Among them, an intelligent controller based on artificial intelligence and its control method, the specific implementation steps are as follows:

[0061] S1, real-time collection of various sensor data of the controlled system;

[0062] S2, perform pre-processing operations such as cleaning, standardization, and noise reduction on the collected raw data;

[0063] S3, automatically extract key features that reflect the system status through AI algorithms;

[0064] S4. Select a suitable AI model based on the characteristics of the system and the control task, and train it using simulated data;

[0065] S5. During operation, the AI ​​processing module receives newly collected data in real time, generates optimal control instructions through model calculation, and acts on the controlled system through the decision execution module;

[0066] S6, continuously update model parameters based on system operation feedback, and the controller can quickly adjust the control strategy when system characteristics change;

[0067] S7. Use AI models to monitor system operation data in real time, identify potential failure modes through anomaly detection algorithms, and predict failure types and causes based on the failure knowledge base.

[0068] In this example, specifically: the AI ​​processing module constructs an initial model structure according to the control task and system characteristics, performs feature extraction engineering processing on the input data, and generates feature vectors required by the model, including numerical feature standardization, normalization, discrete feature encoding, and time series feature extraction;

[0069] Through the processed data, the model training divides the data set into training set, validation set and test set, sets the initial parameters of the model, uses the training set data to iteratively update the model parameters in the gradient descent until the preset number of training rounds is reached, evaluates the model performance on the validation set, calculates the accuracy and loss function value indicators, adjusts the model structure and parameter training strategy according to the evaluation results, encapsulates the trained model into a callable module inside the controller, and interfaces with the data acquisition module and the decision execution module;

[0070] The logistic regression model construction formula is as follows:

[0071]

[0072] Among them, the gradient descent model training formula is as follows:

[0073] θ t+1 =θ t -η▽ θJ(θ t ),

[0074] In the formula, η is the learning rate, J(θ t ) is expressed as the loss function, ▽ θ J(θ t ), expressed as the gradient of the loss function parameter θ.

[0075] In this example, specifically: the decision execution module receives the control instructions processed by AI, calls the corresponding control algorithm or driver according to the analysis results, converts the instructions into specific physical control actions, interacts with the actuator of the controlled system in real time through the communication protocol, sends control signals, receives feedback information from the actuator, and confirms whether the control instructions have been executed correctly.

[0076] In this example, specifically: the fault prediction and diagnosis module trains an anomaly detection model based on historical fault data to identify abnormal patterns in system operation in real time. During operation, the system status data is continuously received and real-time analysis is performed through the anomaly detection model. Once an anomaly is found, an early warning signal is immediately triggered, and an in-depth analysis of the abnormal data is performed to extract feature information related to the fault. The pre-trained fault diagnosis model is used in combination with the fault characteristics and knowledge base to predict the fault type and cause.

[0077] Working principle: By connecting various sensors and other monitoring devices, the physical quantity, state parameters and environmental condition data of the controlled system are obtained in real time, the appropriate sampling frequency is set, data is read from the sensor at preset time intervals, the raw data transmitted by the data acquisition module is received, and preliminary verification is performed to ensure data integrity, the pre-processed data is encapsulated into a format that can be recognized by the controller AI processing module, and is passed to the AI ​​processing module through the internal interface;

[0078] Build the initial model structure according to the control task and system characteristics, extract features and process the input data, use the processed data set to train the model through optimization algorithms such as gradient descent, encapsulate the trained model into a callable module inside the controller, interface with the data acquisition module and the decision execution module, receive the control instructions output by the AI ​​processing module, parse their meaning and execution requirements, call the corresponding control algorithm or driver according to the parsing results, convert the abstract instructions into specific physical control actions, interact with the actuator of the controlled system in real time through the communication protocol, send control signals, and receive feedback from the actuator to confirm whether the control instructions are correct. It has been correctly executed. The actual operating status of the controlled system is continuously monitored through the data acquisition module, compared with the set target, and the deviation between the actual status and the target status is calculated. According to the control error and the control algorithm built into the controller, a feedback signal is generated, and the signal is sent back to the AI ​​processing module. The anomaly detection model is trained using historical fault data. During operation, the system status data is continuously received and analyzed in real time through the anomaly detection model. Once an abnormal pattern is found, the early warning signal is immediately triggered to conduct in-depth analysis of the abnormal data, extract feature information related to the fault, and use the pre-trained fault diagnosis model, combined with the fault characteristics and knowledge base, to predict the type and cause of the fault.

[0079] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

[0080] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. An intelligent controller based on artificial intelligence and a control method thereof, characterized in that: It includes data acquisition module, data preprocessing module, AI processing module, decision execution module, feedback adjustment module and fault prediction and diagnosis module; The data acquisition module is responsible for acquiring various physical quantities, state parameters, and environmental condition data in real time from sensors and other monitoring devices of the controlled system; Wherein, the data preprocessing module is responsible for receiving the data transmitted by the data acquisition module for data preprocessing; The AI ​​processing module processes the data by extracting features from the data, and performs advanced processing such as feature extraction, pattern recognition, and trend prediction on the collected data through a preset control strategy to generate control instructions; The decision execution module receives the control instruction output by the AI ​​processing module and converts it into a specific control signal; The feedback adjustment module compares the actual system response with the expected target by monitoring the control effect, and dynamically adjusts and optimizes the AI ​​processing module through a closed-loop control system; Among them, the fault prediction and diagnosis module uses AI to conduct in-depth analysis of system operation data, identify abnormal patterns, achieve early fault warning, and locate the cause of the fault through reasoning mechanism.

2. The intelligent controller based on artificial intelligence and the control method thereof according to claim 1, characterized in that: The data acquisition module obtains system status information in real time by connecting to various sensors, sets the sampling frequency, and reads data from the sensors at preset time intervals.

3. The intelligent controller based on artificial intelligence and the control method thereof according to claim 1, characterized in that: The data preprocessing module performs preliminary verification, format conversion, and noise filtering on the collected raw data, encapsulates the preprocessed data into a format recognizable by the controller AI processing module, and transmits it to the AI ​​processing module through an internal interface for further analysis.

4. The intelligent controller based on artificial intelligence and the control method thereof according to claim 1, characterized in that: The AI ​​processing module constructs an initial model structure according to the control task and system characteristics, performs feature extraction engineering processing on the input data, and generates feature vectors required by the model, including numerical feature standardization, normalization, discrete feature encoding, and time series feature extraction; Through the processed data, the model training divides the data set into training set, validation set and test set, sets the initial parameters of the model, uses the training set data to iteratively update the model parameters in the gradient descent until the preset number of training rounds is reached, evaluates the model performance on the validation set, calculates the accuracy and loss function value indicators, adjusts the model structure and parameter training strategy according to the evaluation results, encapsulates the trained model into a callable module inside the controller, and interfaces with the data acquisition module and the decision execution module; The logistic regression model construction formula is as follows: Among them, the gradient descent model training formula is as follows: In the formula, η is the learning rate, J(θ t ) is expressed as the loss function, It is expressed as the gradient of the loss function parameter θ.

5. The intelligent controller based on artificial intelligence and the control method thereof according to claim 1, characterized in that: The decision-making execution module receives the control instructions processed by AI, calls the corresponding control algorithm or driver according to the analysis results, converts the instructions into specific physical control actions, interacts with the actuator of the controlled system in real time through the communication protocol, sends control signals, receives feedback information from the actuator, and confirms whether the control instructions have been executed correctly.

6. The intelligent controller based on artificial intelligence and the control method thereof according to claim 1, characterized in that: The feedback adjustment module sets the expected target state according to the control task, continuously monitors the actual operating state of the controlled system through the data acquisition module, compares it with the set target, calculates the deviation between the actual state and the target state, generates a feedback signal based on the control error and the control algorithm of the controller, and sends the feedback signal back to the AI ​​processing module to trigger the model parameter adjustment and update the control strategy of the controller.

7. The intelligent controller based on artificial intelligence and the control method thereof according to claim 1, characterized in that: The fault prediction and diagnosis module trains an anomaly detection model based on historical fault data to identify abnormal patterns in system operation in real time. During operation, the system status data is continuously received and real-time analysis is performed through the anomaly detection model. Once an anomaly is found, an early warning signal is immediately triggered, and an in-depth analysis of the abnormal data is performed to extract feature information related to the fault. The pre-trained fault diagnosis model is used in combination with the fault characteristics and knowledge base to predict the fault type and cause.

8. An intelligent controller based on artificial intelligence and a control method thereof, the specific implementation steps are as follows: S1, real-time collection of various sensor data of the controlled system; S2. Perform pre-processing operations such as cleaning, standardization, and noise reduction on the collected raw data; S3, automatically extract key features that reflect the system status through AI algorithms; S4. Select a suitable AI model based on the characteristics of the system and the control task, and train it using simulated data; S5. During operation, the AI ​​processing module receives newly collected data in real time, generates optimal control instructions through model calculation, and acts on the controlled system through the decision execution module; S6, continuously update model parameters based on system operation feedback, and the controller can quickly adjust the control strategy when system characteristics change; S7. Use AI models to monitor system operation data in real time, identify potential failure modes through anomaly detection algorithms, and predict failure types and causes based on the failure knowledge base.

Images