CN-121979159-A - Industrial interconnection system sensor fault tolerance control method based on artificial intelligence

CN121979159ACN 121979159 ACN121979159 ACN 121979159ACN-121979159-A

Abstract

The purpose of the disclosure is to provide an industrial interconnection system sensor fault tolerance control method based on artificial intelligence, which comprises the steps of releasing unnecessary shutdown waiting constraint in the traditional method by constructing a low constraint switching framework with direction sensing capability, reconstructing a real signal through a software algorithm under the sensor fault by utilizing a state observer and a self-adaptive compensation mechanism, realizing preset performance control by utilizing an error pretreatment mechanism, approaching an unknown item of a system by utilizing an artificial intelligence technology based on a fuzzy logic system, and introducing a dynamic surface signal processing technology to greatly reduce the calculation power requirement of a controller. Therefore, the high-efficiency, safe and low-cost operation of the industrial interconnection system under the complex working condition is realized.

Inventors

DUAN KEN
LIU ZHI

Assignees

广东工业大学
人工智能与数字经济广东省实验室(广州)

Dates

Publication Date: 20260505
Application Date: 20260209

Claims (10)

1. The fault tolerance control method for the sensor of the industrial interconnection system based on the artificial intelligence is characterized by comprising the following steps: Converting any industrial interconnection system in the physical world into a digital state equation which can be identified by a controller, and establishing a sensor fault model according to signal attenuation caused by sensor aging; Constructing a non-singular error scalar transformation function, converting limited tracking errors into unconstrained variables, and constructing a barrier Lyapunov function to ensure that the state of the transformation errors is always kept within a constraint boundary; approximating an unknown nonlinear continuous function by adopting an artificial intelligence technology based on a fuzzy logic system, and converting a complex unknown nonlinear function into a product of a known basis function and an unknown weight vector; constructing a state observer and a fault self-healing module according to the system state non-testability and the characteristics of gain failure faults of the sensor; Based on a back-stepping method and a dynamic surface control technology, combining a tangent barrier Lyapunov function, and recursively constructing a controller; and according to the judging result, the switching of various working condition modes is completed, and the fault tolerance is improved.
2. The fault-tolerant control method for sensor fault of industrial interconnection system based on artificial intelligence according to claim 1, wherein the converting any industrial interconnection system of physical world into a digital state equation recognizable by a controller and establishing a sensor fault model according to signal attenuation caused by sensor aging comprises: Aiming at an industrial interconnection system formed by coupling M subsystems, combining the condition that the system has a switching working condition, and establishing a feedback nonlinear mathematical model of an ith subsystem: ; Wherein, the Is a system state; is a switching signal; Is a nonlinear physical characteristic existing inside the system; to characterize the strong interconnection coupling that exists between subsystems, Is an interference term that is used to determine the interference, In order to control the input of the device, ; Representing time; Representing an output vector of the system; Is the first The order of the subsystems; and (3) building a sensor fault model by combining signal attenuation caused by aging of an industrial field sensor: ; Wherein, the As an actual measurement value of the sensor, Is an unknown sensor failure factor; introduction of fault parameters Will output true Expressed as: ; Wherein, the For the sensor to measure the output of the sensor, And Respectively the approximation error and the approximation value of the fault parameters; Representing time; Is the system sensor failure time.
3. The artificial intelligence based industrial interconnect system sensor fault tolerant control method of claim 1, wherein said constructing a non-singular error scalar transformation function converts a limited tracking error to an unconstrained variable, comprising: construction introducing polynomial buffer term Is a non-singular error scalar transformation function : ; Wherein, the And is also provided with , And Is a design parameter; ; Wherein t represents the time, Is a positive design constant; A predetermined convergence time set for the user.
4. The fault-tolerant control method for industrial interconnection system sensor fault based on artificial intelligence according to claim 1, wherein the constructing the obstacle Lyapunov function to ensure that the conversion error state is always kept within the constraint boundary comprises: Constructing a tangent barrier Lyapunov function to achieve full state constraints Inner: ; Wherein, the Is a state error constraint boundary; is an unconstrained variable; Is a user-defined state constraint.
5. The fault-tolerant control method for industrial interconnection system sensor fault based on artificial intelligence according to claim 1, wherein the approximating the unknown nonlinear continuous function by adopting the artificial intelligence technology based on the fuzzy logic system, converting the complex unknown nonlinear function into the product of the known basis function and the unknown weight vector, comprises: Constructing a fuzzy rule by utilizing the universal approximation characteristic of a fuzzy logic system: Rules of If (1) Is that And Is that Then Is that Wherein Is the fuzzy rule number; the unknown nonlinear function is expressed as: ; Wherein, the , Is an ideal fuzzy weight vector; is a fuzzy basis function vector; is a fuzzy approximation error.
6. The fault-tolerant control method for sensor faults of an industrial interconnection system based on artificial intelligence according to claim 1, wherein said constructing a state observer based on system state non-testability and characteristics of sensor having gain failure faults comprises: For the problem of undetectable system state, when When a plurality of switching systems are activated, the following observers are constructed to estimate the system state : ; Wherein, the A derivative vector that is a state estimate; Is a vector of state estimates; is an observer gain matrix; ; ; Is the actual measurement value of the sensor; A correction term for the observer; is an estimate of the sensor fault parameter; Is an estimation error; ; is an ideal estimated value of the fuzzy weight vector; is a fuzzy basis function vector; For controlling input, command And is also provided with So that Is a strict Hurwitz matrix and there is one matrix Satisfy the following requirements Wherein Is a positive definite symmetric matrix.
7. The fault-tolerant control method for sensor faults of an industrial interconnection system based on artificial intelligence according to claim 1, wherein the self-healing module comprises: On-line acquisition of fault parameters The following adaptive update law is designed: ; Wherein, the Is an adaptive gain; ; And (3) with Can be derived from claim 6; is a positive constant; Is the actual measurement value of the sensor; Is a positive constant.
8. The fault-tolerant control method for industrial interconnection system sensor fault based on artificial intelligence according to claim 1, wherein the recursive construction of the controller based on the back-stepping and dynamic surface control technique in combination with the tangent barrier lyapunov function comprises: definition of first level conversion error plane Constructing a Lyapunov function, and designing a virtual control law and an adaptive law by using a backstepping technology and a Young inequality; a first order low pass filter is introduced: Wherein, the method comprises the steps of, Is a positive constant which is used to control the temperature, Is a filtered virtual control signal; as a virtual controller, the following errors at each level: ; Based on the first Stage error plane and barrier lyapunov function, deriving the actual control inputs and adaptive law: ; Wherein, the , And Is a design parameter, and the control law compensates the influence of fuzzy approximation errors, dynamic surface filtering errors and interconnection terms.
9. The method for fault-tolerant control of industrial interconnection system sensor fault based on artificial intelligence according to claim 1, wherein the step of implanting a logic judgment module based on a working condition switching direction into the controller to obtain a judgment result, and completing switching of multiple types of working condition modes according to the judgment result comprises the steps of: Assume that there is a set of handover subsystems A binary map is defined, which is a function of the mapping, , And is also provided with Is provided with For a subsystem family The system switch signal needs to satisfy the following dwell time conditions: ; Wherein, the Is binary dependent average residence time; is the initial time; For the expiration time T Representing the time interval Total run time; is a positive constant; Representing the time interval Total switching times; Assuming the presence of elements First, an operator is defined, and the first element of the switching ordered pair in the group where the specific switching subsystem is located is collected: ; For the first Subsystem defining a quadratic lyapunov function And its derivative satisfies the following dissipation inequality: Wherein, the method comprises the steps of, Is the attenuation rate; Is a constant term; for arbitrary switching pairs The lyapunov function satisfies the following jump condition: Wherein, the method comprises the steps of, ; According to the method of claim 6, Can be calculated; Is that Is the maximum eigenvalue of (2); Is that Is a minimum feature value of (a).
10. The fault-tolerant control method for sensor faults of an industrial interconnection system based on artificial intelligence of claim 9, wherein the fault-tolerant control method comprises the following steps: Attenuation Rate The definition is as follows: ; Constant term Is defined as ; Wherein, the Is that Is a characteristic value of (2); Is calculated according to claim 6; Is a matrix Is the maximum eigenvalue of (2); is a positive constant; are all design parameters.

Description

Industrial interconnection system sensor fault tolerance control method based on artificial intelligence Technical Field The disclosure relates to the technical field of industrial automation control technology and artificial intelligence application, in particular to a low constraint working condition switching control method which is applicable to complex industrial scenes such as multi-joint mechanical arms, interconnected power systems and the like and has sensor fault tolerance control and preset performance capability. Background In the modern intelligent manufacturing and energy industry, large-scale collaborative operation interconnection systems such as multi-joint collaborative mechanical arms, flexible production lines and multi-region interconnection power grids play a core role. Such systems are typically coupled by a plurality of physical components (subsystems) that require frequent switching operations between a plurality of distinct operating modes. For example, the industrial robot switches between a "high-speed handling" mode and a "high-definition assembly" mode, or the micro-grid switches between a "grid-connected power generation" and an "independent power supply" mode. Frequent abrupt changes in this condition are extremely prone to mechanical jitters or energy fluctuations of the device due to the complex kinetic coupling between the various components. To prevent operating mode switching from causing equipment failure, existing industrial control systems typically employ a conservative time-delay switching strategy (e.g., average residence time mechanism). This strategy forces the device to have to pause or maintain the current state for a longer fixed time after each switching condition to wait for the system to stabilize. However, this "forced waiting" mechanism severely restricts the production efficiency, and in the high-beat operation of the pipeline, even if the equipment is switched from "vigorous motion" to "steady operation", the controller still forces the equipment to wait, resulting in an increase in invalid operation time, and the production beat cannot meet the requirement of the high-speed production line. The prior art lacks of intelligent matching of the working time of the working condition switching direction, and can not flexibly adjust the waiting time according to the actual running state, so that the equipment performance can not be fully released. In addition, such industrial interconnect systems typically operate in harsh environments where high temperatures, shock or electromagnetic interference is severe. Core sensors such as position encoders and transformers are extremely prone to signal drift or gain failure after long-term service. A disadvantage of the prior art is that conventional fine control algorithms generally assume that the sensor data is completely reliable. Once the sensor has hidden faults, the controller outputs wrong driving instructions according to distorted feedback signals, so that the mechanical arm collision, the overheat burnout of a motor or the unstable off-grid of a power grid are extremely easy to cause serious industrial safety accidents. In addition to the stability and fault tolerance problems described above, high precision industrial operations place very stringent constraints on the dynamic response process of the system. Although the traditional control method can ensure the final stability of the system, the instantaneous error in the adjusting process cannot be quantitatively restrained, which directly leads to uncontrollable production beats. Therefore, it is necessary to design a control strategy that can strictly limit the system state within the "preset safety envelope", i.e. the predetermined performance control, to ensure that the dynamic error of the system is always constrained within the allowed physical range regardless of the severe switching of the operating conditions. On the other hand, in actual operation, complex industrial interconnection systems often face serious model uncertainties, such as unknown physical characteristics of friction, abrasion and the like. Conventional adaptive control relies on linear parameterization assumptions, which make it difficult to cope with unstructured uncertainties. To solve this problem, artificial intelligence techniques such as neural networks and fuzzy logic systems based on general approximation principles are widely introduced for online learning and approximating these unknown nonlinear dynamics. Meanwhile, for the multivariable strong coupling system, the academic world often adopts a back-step method to design a controller. However, when the engineering is landed, as the number of system modules increases, the differential operand required by the backstepping method expands exponentially, namely, the problem of differential explosion. This requires that the controller must be equipped with an expensive high performance computing chip, which adds significantly