CN-121980415-A - Multi-index decision-based real-time monitoring method and system for equipment health degree

CN121980415ACN 121980415 ACN121980415 ACN 121980415ACN-121980415-A

Abstract

The invention relates to the technical field of equipment state monitoring and fault diagnosis, in particular to a method and a system for monitoring equipment health degree in real time based on multi-index decision. The equipment health degree real-time monitoring method based on multi-index decision-making comprises the following steps of establishing a multi-dimensional index treatment model according to equipment measuring points, extracting blood edge relations among all levels of treatment indexes in the multi-dimensional index treatment model based on a time sequence map engine, constructing a blood edge map, calculating initialization weights of all levels of treatment indexes based on the blood edge map, calculating weight dynamic adjustment coefficients of all levels of treatment indexes based on equipment real-time working conditions, obtaining dynamic weights of all levels of treatment indexes adapting to the equipment real-time working conditions based on the initialization weights of all levels of treatment indexes, constructing a multi-dimensional decision-making equipment health degree assessment model based on all levels of treatment indexes and all levels of treatment indexes of the equipment real-time working conditions, and calculating equipment real-time health degree comprehensive scores.

Inventors

DENG KAIWEN
LIU CONGRUI
WANG YI
Pang Wuhua
Hao Weipeng

Assignees

西安西热电站信息技术有限公司
西安热工研究院有限公司

Dates

Publication Date: 20260505
Application Date: 20260112

Claims (10)

1.‌ A method for monitoring the health degree of equipment in real time based on multi-index decision-making is characterized by comprising the following steps: establishing a multidimensional index treatment model according to the equipment measuring points; extracting blood margin relations among all levels of treatment indexes in the multi-dimensional index treatment model based on a time sequence map engine, and constructing a blood margin map; calculating the initialization weight of each level of treatment indexes based on the blood margin map; Calculating a dynamic weight adjustment coefficient of each level of treatment index based on the real-time working condition of the equipment, and obtaining the dynamic weight of each level of treatment index adapting to the real-time working condition of the equipment based on the dynamic weight adjustment coefficient of each level of treatment index and the initialization weight of each level of treatment index; And constructing a multi-dimensional decision-making equipment health degree evaluation model based on the dynamic weights of the treatment indexes of each level and the real-time working conditions of the equipment, calculating the real-time health degree comprehensive score of the equipment, and acquiring the key influence treatment indexes.
2. The method for real-time monitoring of equipment health based on multi-index decision according to claim 1, wherein the establishing a multi-dimensional index treatment model according to the equipment measuring points comprises: acquiring measurement point data related to the whole life cycle of equipment; Dividing index dimensions according to the measurement point data; And screening the core monitoring index from the index dimension, and carrying out standardization treatment to construct a multi-dimension index treatment model.
3. The method for monitoring the health of equipment based on multi-index decision according to claim 2, wherein the index dimension comprises at least one of a running state dimension, a performance parameter dimension, an environmental impact dimension and an operation and maintenance history dimension, and the normalization processing comprises data cleaning, outlier rejection and normalization conversion.
4. The method for monitoring the health of a device based on multi-index decision according to claim 1, wherein the time series pattern engine extracts blood margin relations among the treatment indexes of each level in the multi-dimensional index treatment model and constructs a blood margin pattern, and the method comprises the following steps: Analyzing the data sources, the calculation logic and the association dependency relationship of the treatment indexes of each level by using a time sequence map engine, and extracting the direct blood relationship and the indirect blood relationship between the treatment indexes of each level; and constructing a visual blood-margin map by taking each level of treatment indexes as nodes and taking a direct blood-margin relation and an indirect blood-margin relation as edges, so as to obtain a conduction path and influence weight association basis between each level of treatment indexes.
5. The method for monitoring the health degree of a device based on multi-index decision according to claim 1, wherein the calculating the initialization weight of each level of treatment index based on the blood-margin map comprises: determining node weights of the level governance indexes based on the level relation of the level governance indexes in the blood margin map, and further obtaining the initialization weights of the level governance indexes; the calculation formula of the node weight of each level of treatment index is shown in formula 1: 1 (1) Wherein d is the depth of the level in the blood-related map, Attenuation coefficient, and 。
6. The method for real-time monitoring of equipment health based on multi-index decision according to claim 5, wherein determining the node weight of each level of treatment index based on the level relation of each level of treatment index in the blood-margin map, and further obtaining the initialization weight of each level of treatment index, comprises: Determining the node weight of each level of treatment index based on the level relation of each level of treatment index in the blood margin map, and determining the initialization weight of the treatment index of each node in the same level based on the node weight of each level of treatment index, so as to obtain the initialization weight of each level of treatment index; The step of determining the initialization weight of the treatment index of each node in the same level based on the node weight of the treatment index of each level, thereby obtaining the initialization weight of the treatment index of each level comprises the following steps: First, for the treatment index of each node in the same hierarchy, the calculation frequency is divided into 7 grades of seconds, minutes, hours, days, weeks, months and years according to the treatment index, and then the frequency weights are respectively 、、、、、 And ; Then, counting the duty ratio s of the normal state code in the latest 10 value state codes of the treatment index, and initializing the weight of the treatment index corresponding to the j-th node of the i-th layer As shown in formula 2: 2, 2 Wherein, the Is the frequency weight of the governance index corresponding to the j-th node of the i-th layer, The duty ratio of the normal state code in the latest 10 value state codes of the treatment index corresponding to the jth node of the ith layer; Node weight of the treatment index of the ith level; is the total number of all nodes of the i-th layer.
7. The method for real-time monitoring of equipment health based on multi-index decision according to claim 1, wherein calculating the dynamic adjustment coefficient of the weight of each level of the treatment index based on the real-time working condition of the equipment, and obtaining the dynamic weight of each level of the treatment index adapted to the real-time working condition of the equipment based on the dynamic adjustment coefficient of the weight of each level of the treatment index and the initialization weight of each level of the treatment index, comprises: Collecting working condition parameters of equipment in real time, and establishing a working condition-weight influence model; Analyzing the difference of the influence degree of different real-time working condition parameters on the treatment indexes of each level according to the working condition-weight influence model, and calculating the weight dynamic adjustment coefficient of the treatment indexes of each level; Correcting the initialized weight of each level treatment index by using the dynamic weight adjustment coefficient of each level treatment index to obtain the dynamic weight of each level treatment index adapting to the real-time working condition of the equipment; the dynamic weight set W of each level of treatment indexes of the real-time working condition of the adapting equipment is obtained by the following formula 3: 3 Wherein, the Controlling index sets for equipment Corresponding initial threshold set =[ ]; And scoring matrix for the equipment working condition parameters.
8. The method for real-time monitoring of equipment health based on multi-index decisions according to claim 7, wherein the constructing the multi-dimensional decision-making equipment health assessment model based on dynamic weights of each level of treatment indexes and each level of treatment indexes of real-time equipment working conditions, calculating equipment real-time health comprehensive scores and obtaining key impact treatment indexes comprises: Based on the dynamic weight of each level of treatment index of the real-time working condition of the equipment, constructing an equipment health evaluation model; Inputting real-time data of each level of treatment indexes in the multi-dimensional index treatment model into the equipment health evaluation model to obtain the equipment real-time health comprehensive score and key influence treatment indexes.
9. The method for monitoring the health degree of a device in real time based on multi-index decision according to claim 8, wherein the calculation of the real-time health degree composite score of the device is as shown in formula 4: 4. The method is to Wherein, the To treat the index Is used for the dynamic weighting of the (c) in the (c), To treat the index Is a target function score of (2); The said Obtained from formula 5: 5. The method is to Wherein, the In order to treat the index of the plant, To treat the index The corresponding index anomaly threshold value is set, To treat the index Corresponding index health threshold value, and 。
10. A system for performing a multi-index decision-based real-time device health monitoring method, comprising: the multi-dimensional index treatment model building module is used for building a multi-dimensional index treatment model according to the equipment measuring points; The blood margin map construction module is used for extracting blood margin relations among all levels of treatment indexes in the multi-dimensional index treatment model based on the time sequence map engine and constructing a blood margin map; the initialization weight acquisition module is used for calculating the initialization weight of each level of treatment indexes based on the blood margin map; The dynamic weight acquisition module is used for calculating the dynamic weight adjustment coefficient of each level of treatment index based on the real-time working condition of the equipment and obtaining the dynamic weight of each level of treatment index adapting to the real-time working condition of the equipment based on the dynamic weight adjustment coefficient of each level of treatment index and the initialization weight of each level of treatment index; And the equipment real-time health comprehensive score acquisition module is used for constructing a multi-dimensional decision-making equipment health evaluation model based on the dynamic weights of the treatment indexes of each level and the treatment indexes of each level of the real-time working condition of the equipment, calculating the equipment real-time health comprehensive score and acquiring key influence treatment indexes.

Description

Multi-index decision-based real-time monitoring method and system for equipment health degree Technical Field The invention relates to the technical field of equipment state monitoring and fault diagnosis, in particular to a method and a system for monitoring equipment health degree in real time based on multi-index decision. Background The existing equipment health degree monitoring method is mainly divided into two types of single-index monitoring and multi-index monitoring. The single index monitoring method only judges the state of the equipment through single physical quantities such as vibration, temperature, pressure and the like, and has the problems of single monitoring dimension and high misjudgment rate, namely, for example, the equipment fault is judged only according to the increase of the vibration amplitude, and the instantaneous fluctuation under the normal working condition can be misjudged as a fault signal or the hidden fault caused by multi-factor coupling is omitted. The multi-index monitoring method introduces a plurality of state parameters, but has obvious defects in an index fusion and decision mechanism, namely, on one hand, the multi-index monitoring method adopts a simple weighted summation mode to fuse indexes, the weight distribution depends on manual experience, the subjectivity is strong, the method cannot adapt to dynamic changes of index importance under different equipment and different working conditions, and on the other hand, the method lacks the capability of real-time preprocessing and abnormality identification of index data, is difficult to cope with noise interference under complex working conditions, and causes lag and insufficient accuracy of health assessment. In addition, the health degree grade of the existing method is rough, the gradual change process from 'health' to 'fault' of equipment cannot be accurately reflected, and the establishment of a fine maintenance strategy is not facilitated. Therefore, a real-time monitoring method for equipment health degree, which can realize multi-index dynamic weighted fusion, has strong anti-interference capability and accurate evaluation, is needed to solve the problems of strong subjectivity, poor adaptability, monitoring lag and the like in the prior art. Disclosure of Invention The invention aims to provide a real-time equipment health monitoring method based on multi-index decision, which aims to solve the problems of strong subjectivity, poor adaptability and monitoring lag of ‌ multi-index monitoring methods in the prior art. In order to solve the problems, the invention provides a multi-index decision-based real-time monitoring method for the health degree of equipment, which adopts the following technical scheme: a real-time monitoring method for equipment health based on multi-index decision-making comprises the following steps: establishing a multidimensional index treatment model according to the equipment measuring points; extracting blood margin relations among all levels of treatment indexes in the multi-dimensional index treatment model based on a time sequence map engine, and constructing a blood margin map; calculating the initialization weight of each level of treatment indexes based on the blood margin map; Calculating a dynamic weight adjustment coefficient of each level of treatment index based on the real-time working condition of the equipment, and obtaining the dynamic weight of each level of treatment index adapting to the real-time working condition of the equipment based on the dynamic weight adjustment coefficient of each level of treatment index and the initialization weight of each level of treatment index; And constructing a multi-dimensional decision-making equipment health degree evaluation model based on the dynamic weights of the treatment indexes of each level and the real-time working conditions of the equipment, calculating the real-time health degree comprehensive score of the equipment, and acquiring the key influence treatment indexes. Further, the establishing a multidimensional index treatment model according to the equipment measuring points comprises the following steps: acquiring measurement point data related to the whole life cycle of equipment; Dividing index dimensions according to the measurement point data; And screening the core monitoring index from the index dimension, and carrying out standardization treatment to construct a multi-dimension index treatment model. Further, the index dimension comprises at least one of a running state dimension, a performance parameter dimension, an environment influence dimension and an operation and maintenance history dimension, and the standardized processing comprises data cleaning, outlier rejection and normalized conversion. Further, the extracting the blood margin relation between the treatment indexes of each level in the multi-dimension index treatment model based on the time sequence map engine and constructing the blood