CN-115840919-B - Fault prediction method and device for sensing control system

CN115840919BCN 115840919 BCN115840919 BCN 115840919BCN-115840919-B

Abstract

The invention provides a method and a device for predicting faults of a sensing control system, and relates to the technical field of sensing control. The method comprises the steps of dividing equipment in a sensing control system according to working stages to obtain a plurality of equipment sets, acquiring parameter information of equipment in each equipment set in real time, respectively evaluating by adopting health evaluation models corresponding to the equipment sets according to the parameter information of the equipment in each equipment set to obtain evaluation results of each working stage, determining fault characteristics from the parameter information of the equipment in each equipment set according to the evaluation results of each working stage, and finally carrying out fault prediction on the sensing control system according to the fault characteristics by adopting a fault prediction model constructed on the basis of a kernel principal component analysis algorithm and a random forest algorithm to obtain a system fault prediction result. The fault prediction result is more accurate, the obtained fault characteristics are effective fault characteristics, the accuracy of system fault prediction is further improved, and the reliability of the whole sensing control system is further improved.

Inventors

SUN JINHU
YANG YUN
JI SHAOFENG
LI ZUOYUN
HUANG QIN
Lu Dengbo
LIU YONG

Assignees

老肯医疗科技股份有限公司
西南交通大学

Dates

Publication Date: 20260512
Application Date: 20221207

Claims (8)

1. The fault prediction method of the sensing control system is characterized by comprising the following steps of: Dividing the equipment in the sensing control system according to the working stage to obtain a plurality of equipment sets, wherein the equipment sets comprise one or more pieces of equipment; acquiring parameter information of each device centralized device in real time; Respectively carrying out evaluation by adopting health evaluation models corresponding to all equipment sets according to the parameter information of the equipment in all equipment sets to obtain evaluation results of all working stages; comparing the evaluation results of each working stage with corresponding stage thresholds to obtain a plurality of comparison results, determining an abnormal working stage according to the plurality of comparison results, extracting parameter information of equipment in a corresponding equipment set according to the abnormal working stage, taking the comparison result corresponding to the abnormal working stage as a weight, taking the parameter information of the equipment in the corresponding equipment set as a characteristic parameter, and multiplying the characteristic parameter by the weight to obtain a fault characteristic; And according to the fault characteristics, carrying out fault prediction on the sensing control system by adopting a fault prediction model constructed based on a kernel principal component analysis algorithm and a random forest algorithm to obtain a system fault prediction result.
2. The method for predicting a failure of a sensing system according to claim 1, wherein the construction of the health assessment model comprises the steps of: respectively acquiring historical data samples of each equipment set, wherein the historical data samples of each equipment set comprise historical equipment parameter information and health states of corresponding working stages; And training by adopting an evaluation model based on XGBoost algorithm according to the historical data sample of each equipment set to obtain a health evaluation model corresponding to each equipment set.
3. The method of claim 1, further comprising the steps of: Acquiring a fault training set and a fault testing set; according to the fault training set, performing model construction by adopting a kernel principal component analysis algorithm and a random forest algorithm to obtain an initial prediction model; And optimizing the initial prediction model by adopting the fault test set to obtain a fault prediction model.
4. The method for predicting faults of a sensing control system according to claim 3, wherein the initial prediction model is obtained by carrying out model construction by adopting a kernel principal component analysis algorithm and a random forest algorithm according to the fault training set, and the method comprises the following steps: carrying out normalization processing on the characteristic samples in the fault training set to obtain preprocessed characteristic samples; Adding random noise to the preprocessed feature sample to obtain a new feature sample; Performing kernel principal component extraction on the new feature sample by adopting a kernel principal component analysis algorithm to obtain a new feature sample matrix; And training a preset random forest model according to the new characteristic sample matrix and the fault training set to obtain an initial prediction model.
5. The method of claim 1, further comprising the steps of: and determining fault equipment according to a system fault prediction result, and generating fault reminding information according to the fault equipment.
6.A sensing and control system fault prediction apparatus, comprising: The device dividing module is used for dividing the devices in the sensing control system according to the working stage to obtain a plurality of device sets, wherein the device sets comprise one or more devices; The parameter information acquisition module is used for acquiring the parameter information of each equipment centralized equipment in real time; the evaluation module is used for evaluating by adopting a health evaluation model corresponding to each equipment set according to the parameter information of the equipment in each equipment set respectively to obtain an evaluation result of each working stage; The fault characteristic determining module is used for comparing the evaluation results of each working stage with corresponding stage thresholds respectively to obtain a plurality of comparison results, determining an abnormal working stage according to the plurality of comparison results, extracting parameter information of corresponding equipment centralized equipment according to the abnormal working stage, taking the comparison result corresponding to the abnormal working stage as a weight, taking the parameter information of the corresponding equipment centralized equipment as a characteristic parameter, and multiplying the characteristic parameter by the weight to obtain a fault characteristic; And the fault prediction module is used for predicting the faults of the sensing control system by adopting a fault prediction model constructed based on a kernel principal component analysis algorithm and a random forest algorithm according to the fault characteristics to obtain a system fault prediction result.
7. An electronic device, comprising: A memory for storing one or more programs; A processor; The sensing system fault prediction method of any one of claims 1-5 when the one or more programs are executed by the processor.
8. A computer readable storage medium having stored thereon a computer program, which when executed by a processor implements the sensing system fault prediction method according to any one of claims 1-5.

Description

Fault prediction method and device for sensing control system Technical Field The invention relates to the technical field of sensing control, in particular to a sensing control system fault prediction method and device. Background In the field of sensing control, faults or potential fault risks of equipment can bring adverse effects such as work failure, progress stagnation delay and the like to links such as cleaning, drying, disinfection, sterilization and the like, so that stable and timely supply of disinfection medical equipment is influenced, and even medical accidents such as operation suspension and the like are caused. In fact, this instability, unreliability, is the most prominent problem in the field of sensing. In order to solve the problems of instability and unreliability, besides the conventional approach for improving the reliability of a single device, the real-time monitoring of the state of each device and the prediction and diagnosis of faults of each device are taken as a hopeful effective approach, so that the device can take appropriate active measures such as inspection, maintenance, repair and the like in good time. At present, fault prediction and diagnosis technologies have been applied in the industrial field, along with the high-speed development of industrial internet of things and intelligent sensing technologies, industrial production equipment increasingly starts to adopt a factory equipment fault prediction and diagnosis system so as to timely and accurately monitor and predict the production process or equipment state in real time, and the fault prediction method is also applicable in the sensing and control field, however, the existing fault prediction method has the problem of inaccuracy, so that the reliability of a sensing and control system is lower. Disclosure of Invention The invention aims to provide a fault prediction method and device for a sensing control system, which are used for solving the problem that the reliability of the sensing control system is low due to inaccurate fault prediction in the prior art. In a first aspect, an embodiment of the present application provides a method for predicting a failure of a sensing control system, including the following steps: Dividing the equipment in the sensing control system according to the working stage to obtain a plurality of equipment sets, wherein the equipment sets comprise one or more pieces of equipment; acquiring parameter information of each device centralized device in real time; Respectively carrying out evaluation by adopting health evaluation models corresponding to all equipment sets according to the parameter information of the equipment in all equipment sets to obtain evaluation results of all working stages; determining fault characteristics from the parameter information of the equipment in each equipment set according to the evaluation results of each working stage; And according to the fault characteristics, carrying out fault prediction on the sensing control system by adopting a fault prediction model constructed based on a kernel principal component analysis algorithm and a random forest algorithm to obtain a system fault prediction result. Based on the first aspect, in some embodiments of the invention, the building of the health assessment model includes the following step 5: respectively acquiring historical data samples of each equipment set, wherein the historical data samples of each equipment set comprise historical equipment parameter information and health states of corresponding working stages; And training by adopting an evaluation model based on XGBoost algorithm according to the historical data sample of each equipment set to obtain a health evaluation model corresponding to each equipment set. 0 Is based on the first aspect, in some embodiments of the invention, based on the evaluation results of the respective working phases, from Determining fault characteristics in parameter information of each device centralized device, including the following steps: Comparing the evaluation results of each working stage with corresponding stage thresholds respectively to obtain a plurality of comparison results; Determining an abnormal working stage according to the comparison results; extracting parameter information of the centralized equipment of the corresponding equipment according to the abnormal working stage; and determining fault characteristics according to the parameter information of the corresponding equipment centralized equipment and the corresponding comparison result. Based on the first aspect, in some embodiments of the present invention, determining the fault feature according to the parameter information of the corresponding device centralized device and the corresponding comparison result includes the following steps: Taking a comparison result corresponding to the abnormal working stage as a weight; 0, taking parameter information of the corresponding equipment cent