CN-121995765-A - Intelligent control system based on multisource data fusion

CN121995765ACN 121995765 ACN121995765 ACN 121995765ACN-121995765-A

Abstract

The invention discloses an intelligent control system based on multi-source data fusion, which belongs to the technical field of intelligent control and comprises a data acquisition preprocessing module, a data fusion module, a device health state predictive diagnosis module, a multi-target self-adaptive control strategy generation module, a cooperative control distribution module and a real-time feedback parameter optimization module, wherein the intelligent control system based on multi-source data fusion is constructed to realize full life cycle management of device states; the system remarkably improves the reliability, energy efficiency optimization capability and maintenance response speed of equipment operation through standardized processing of a data acquisition preprocessing module, causal association modeling of a data fusion module, residual life prediction and fault identification of a health state predictive diagnosis module, dynamic optimization of a multi-target self-adaptive control strategy generation module, distributed execution of a cooperative control distribution module and closed loop correction of a real-time feedback parameter optimization module.

Inventors

XIE YIMIN
Shi Kunkun
WANG XIUSHENG
HONG CHENG
Feng Caoyu

Assignees

卓亿智能装备(江苏)有限公司

Dates

Publication Date: 20260508
Application Date: 20260225

Claims (10)

1. The intelligent control system based on multi-source data fusion is characterized by comprising a data acquisition preprocessing module, a data fusion module, a device health state predictive diagnosis module, a multi-target self-adaptive control strategy generation module, a cooperative control distribution module and a real-time feedback parameter optimization module; the data acquisition preprocessing module is used for acquiring multi-source heterogeneous data for real-time preprocessing; the data fusion module carries out dynamic association feature fusion according to the preprocessed multi-mode data; the equipment health state predictive diagnosis module predicts the residual life and identifies the fault mode according to the fusion characteristics; the multi-target self-adaptive control strategy generation module generates self-adaptive control parameters according to the health diagnosis result and guides cooperative control allocation; The cooperative control distribution module performs instruction execution according to the self-adaptive control parameters; And the real-time feedback parameter optimization module performs parameter self-adaptive updating according to the execution feedback data, optimizes the control strategy and corrects the execution deviation.
2. The intelligent control system based on multi-source data fusion according to claim 1, wherein the data acquisition preprocessing module is used for acquiring vibration signals, temperature field distribution, noise spectrums, surface deformation images and PLC state data in real time through an Internet of things sensor network, an industrial camera, an acoustic array and an equipment log interface, accurately aligning acquisition times of different data sources, eliminating time deviation caused by equipment sampling frequency difference, carrying out noise filtering and outlier processing on the acquired original data, processing missing values through an interpolation method, simultaneously retaining important abnormal event marks, and carrying out structural storage on preprocessed standardized feature data according to equipment ID, time stamp and feature dimension.
3. The intelligent control system based on multi-source data fusion of claim 2, wherein the data fusion module receives the preprocessed data, analyzes time sequence dependency relations among different mode data through Granger causal test, introduces a device mechanism model as priori knowledge, corrects noise interference in pure statistical causal relations, and generates causal relation strength, and the implementation is as follows: , in the formula (i) the formula (ii), Representing the causal strength of the modality i versus the modality j, Represents the Granger causal strength of modality i versus modality j, Representing the physical constraint correction factor between modality i and modality j, The original data sequence representing the i-th modality, The original data sequence representing the j-th modality, n representing the number of total modalities.
4. The intelligent control system based on multi-source data fusion according to claim 3, wherein the data acquisition preprocessing module performs weighted fusion by pearson statistical feature correlation analysis based on causal correlation strength, and is implemented as follows: , in the formula (i) the formula (ii), Representing the dynamically weighted fusion characteristics of modality j, Representing the balance factor of the balance, Representing the pearson correlation coefficient between modality i and modality j, Representing the maximum of the absolute values of pearson correlation coefficients between all modes i and j, Representing the normalized feature vector of modality i.
5. The intelligent control system based on multi-source data fusion of claim 4, wherein the device health status predictive diagnosis module receives the dynamic weighted fusion features, normalizes the dynamic weighted fusion features, square weights the dynamic weighted fusion features, and introduces feature weights to perform weighted average to obtain a health attenuation index, and the method is implemented as follows: , in the formula (i) the formula (ii), Indicating a health decay index (lll) that, Representing the current value of the i-th feature in the dynamically weighted fusion features, Represents the historical maximum of the ith feature, The weight of the i-th feature is represented.
6. The intelligent control system based on multi-source data fusion as set forth in claim 5, wherein said device health status predictive diagnostic module sets a threshold value Comparing the health attenuation index with a threshold value, judging whether the equipment enters an abnormal state, and realizing the fault mode identification index as follows: , in the formula (i) the formula (ii), Which represents the failure mode identification index, Representing the fault sensitivity coefficient.
7. The intelligent control system based on multi-source data fusion as set forth in claim 6, wherein the multi-objective adaptive control strategy generation module receives the health attenuation index and the failure mode identification result output by the equipment health state predictive diagnosis module, maps the diagnosis result to a specific control objective priority, constructs a multi-objective optimization problem including energy consumption efficiency, equipment life, operation precision and safety threshold based on the health diagnosis result, and dynamically maps the health index and the failure identification to weight values of all objectives by setting a weight coefficient matrix.
8. The intelligent control system based on multi-source data fusion of claim 7, wherein the multi-target adaptive control strategy generation module generates a pareto optimal solution set according to the current health state, selects an optimal control strategy from the pareto optimal solution set according to the real-time working condition, uses a health diagnosis result as an input of parameter update through a dynamic weight distribution mechanism to generate specific control parameters, performs stability rapid verification before generating the control parameters, selects a suboptimal solution from the pareto solution set to regenerate if the verification is not passed, classifies the generated control strategy according to priority, and generates a structured control instruction set.
9. The intelligent control system based on multi-source data fusion according to claim 8, wherein the cooperative control allocation module receives the adaptive control parameters and the priority levels output by the multi-target adaptive control strategy generation module, disassembles the adaptive control strategy into executable local tasks, establishes a task resource matching table according to task requirements and an execution unit, inherits the priority levels of the control strategy into the local tasks, monitors the resource occupation state of the execution unit in real time, and if resource conflicts between tasks are found, triggers a fault isolation mechanism, pauses low-priority tasks and reschedules resource allocation paths.
10. The intelligent control system based on multi-source data fusion according to claim 9, wherein the real-time feedback parameter optimization module receives the execution feedback data and the task execution state from the cooperative control distribution module in real time, compares the feedback data with a preset target value, counts deviation indexes, marks abnormal events exceeding a threshold value, analyzes deviation sources in combination with a fault mode identification result of the health state predictive diagnosis module, and performs feedback optimization control strategy on the analysis results.

Description

Intelligent control system based on multisource data fusion Technical Field The invention belongs to the technical field of intelligent control, and particularly relates to an intelligent control system based on multi-source data fusion. Background The existing intelligent control system generally relies on a single type sensor to perform state sensing and decision, and has the problems of limited sensing dimension, poor environmental adaptability, high misjudgment rate and the like. However, the existing intelligent control system for multi-source data fusion has a certain defect that the prior art is based on a pure statistical method, ignores physical causal relation in equipment operation, relies on a single threshold judgment or static model, cannot dynamically quantify equipment health attenuation trend, causes fault early warning hysteresis, cannot dynamically adjust weights according to equipment health states based on fixed rules or single target optimization, adopts a centralized task allocation strategy, lacks a real-time monitoring and dynamic arbitration mechanism for the occupied state of execution unit resources, and causes frequent task conflict. Disclosure of Invention The invention aims to provide an intelligent control system based on multi-source data fusion, which aims to solve the problems in the background technology. In order to achieve the aim, the intelligent control system based on multi-source data fusion comprises a data acquisition preprocessing module, a data fusion module, a device health state predictive diagnosis module, a multi-target self-adaptive control strategy generation module, a cooperative control distribution module and a real-time feedback parameter optimization module; the data acquisition preprocessing module is used for acquiring multi-source heterogeneous data for real-time preprocessing; the data fusion module carries out dynamic association feature fusion according to the preprocessed multi-mode data; the equipment health state predictive diagnosis module predicts the residual life and identifies the fault mode according to the fusion characteristics; the multi-target self-adaptive control strategy generation module generates self-adaptive control parameters according to the health diagnosis result and guides cooperative control allocation; The cooperative control distribution module performs instruction execution according to the self-adaptive control parameters; And the real-time feedback parameter optimization module performs parameter self-adaptive updating according to the execution feedback data, optimizes the control strategy and corrects the execution deviation. Preferably, the data acquisition preprocessing module is in wireless connection with the data fusion module, the data fusion module is in wireless connection with the equipment health state predictive diagnosis module, the equipment health state predictive diagnosis module is in wireless connection with the multi-target self-adaptive control strategy generation module, the multi-target self-adaptive control strategy generation module is in wireless connection with the cooperative control distribution module, and the cooperative control distribution module is in wireless connection with the real-time feedback parameter optimization module. Preferably, the data acquisition preprocessing module acquires vibration signals, temperature field distribution, noise spectrum, surface deformation images and PLC state data in real time through an internet of things sensor network, an industrial camera, an acoustic array and an equipment log interface, accurately aligns acquisition time of different data sources, eliminates time deviation caused by equipment sampling frequency difference, performs noise filtering and outlier processing on acquired original data, processes missing values through an interpolation method, simultaneously reserves important abnormal event marks, and structurally stores preprocessed standardized characteristic data according to equipment ID, time stamp and characteristic dimension. Preferably, the data fusion module receives the preprocessed data, analyzes time sequence dependency relations among different mode data through a Granger causal test, introduces a device mechanism model as priori knowledge, corrects noise interference in pure statistical causal relations, and generates causal relation strength, and the implementation is as follows: , in the formula (i) the formula (ii), Representing the causal strength of the modality i versus the modality j,Represents the Granger causal strength of modality i versus modality j,Representing the physical constraint correction factor between modality i and modality j,The original data sequence representing the i-th modality,The original data sequence representing the j-th modality, n representing the number of total modalities. Preferably, the data acquisition preprocessing module performs weighted fusion through pearson statistical