CN-121168724-B - Pump station equipment risk intelligent prediction method based on big data analysis

CN121168724BCN 121168724 BCN121168724 BCN 121168724BCN-121168724-B

Abstract

The invention relates to the technical field of fault prediction and health management, and discloses an intelligent pump station equipment risk prediction method based on big data analysis, which comprises the steps of collecting equipment operation and environment data; the method comprises the steps of constructing a mixed intelligent causal analysis model for automatic screening to obtain association characteristics among equipment parameters, constructing a risk analysis model which combines topological analysis and time sequence characteristics of a graph and is calibrated by environmental data, outputting a systematic running risk index and a dynamic critical value, constructing a long time sequence prediction model, carrying out long-term trend prediction on the risk index, calling the causal model to carry out root cause tracing when the predicted risk reaches the critical value, and generating a prevention management scheme comprising clear early warning grade and specific resource requirements. The invention realizes the full-flow intellectualization from the unified risk assessment, the accurate prediction and the closed-loop decision support.

Inventors

FANG XIANWEN
LU KE
FANG NA
ZHANG RUOYUAN
LIU HAIHAO

Assignees

安徽理工大学

Dates

Publication Date: 20260512
Application Date: 20250901

Claims (6)

1. The intelligent pump station equipment risk prediction method based on big data analysis is characterized by comprising the following steps of: collecting operation parameters and environment data of pump station equipment; Constructing a causal analysis model, constructing a causal relation constraint set through a gradient lifting decision tree based on the operation parameters, and identifying the causal relation constraint set through a Bayesian network to obtain parameter association characteristics; Constructing a risk analysis model, performing equipment association topology analysis and time sequence feature analysis on the operation parameters and parameter association features to obtain first operation risk data and a first risk critical value, checking based on environment data to obtain an operation risk data check value and a risk critical check value, and generating operation risk data and a risk critical value; the method comprises the steps of identifying signal periodic frequency of operation risk data, inputting a time delay aggregation model to obtain periodic characteristics, carrying out frequency domain analysis on a time sequence of the operation risk data by adopting a Long Bu-Skoger periodic chart algorithm to obtain a maximum power value, taking the frequency corresponding to the maximum power value as the signal periodic frequency, carrying out decoupling treatment on the signal periodic frequency to obtain phase information, determining a periodic starting point based on the phase information to obtain a fundamental frequency period and a harmonic frequency, inputting the fundamental frequency period and the harmonic frequency into the time delay aggregation model, sampling the operation risk data, calculating a mean value and a variance to obtain periodic characteristics, and carrying out autoregressive generation on the signal periodic frequency based on the periodic characteristics to obtain a prediction risk index; and when the predicted risk index reaches the risk critical value, generating a risk early warning grade and a prevention management scheme through the causal analysis model.
2. The intelligent pump station equipment risk prediction method based on big data analysis according to claim 1, wherein the operation parameters comprise a triaxial vibration value at a water pump bearing seat, a bearing temperature, inlet and outlet pressures of a pump and three-phase current and winding temperature of a driving motor, and the environmental data comprise environmental temperature, air humidity and air quality index of a pump room.
3. The intelligent pump station equipment risk prediction method based on big data analysis according to claim 1, wherein the step of constructing a causal analysis model to obtain parameter association features comprises the following steps: Analyzing the operation parameters by utilizing a gradient lifting decision tree, and sorting according to the feature importance of each parameter to construct a causal relation constraint set; And taking the causal relation constraint set as priori knowledge and search constraint, inputting the causal relation constraint set into a Bayesian network, generating a network topology structure of a directional dependency relation among all operation parameters, and obtaining the parameter association characteristics according to the network topology structure.
4. The intelligent prediction method for pump station equipment risk based on big data analysis according to claim 1, wherein the process of performing equipment association topology analysis and time sequence feature analysis to obtain first operation risk data and a first risk critical value comprises the following steps: The operating parameters are used as network nodes, a topological network is formed according to the parameter association characteristics, and the score of each node is calculated in an iterative mode to obtain a topological risk influence factor; and obtaining first operation risk data and a first risk critical value through self-adaptive weighted fusion based on the topological risk influence factor and the deviation degree.
5. The intelligent pump station equipment risk prediction method based on big data analysis according to claim 1, wherein the step of verifying based on environmental data to obtain a risk data verification value and a risk critical verification value comprises the following steps: Performing causal inference on the operation parameters and the environment data to form an environment-parameter map, and calculating a difference value output by the current environment and the reference environment under the map to obtain an environment influence deviation; and converting the environmental impact deviation into an operation risk data check value and a risk critical check value according to a preset mapping rule based on the interval to which the environmental impact deviation belongs.
6. The intelligent pump station equipment risk prediction method based on big data analysis according to claim 1, wherein the specific process of generating the risk early warning level and the prevention management scheme comprises the following steps: When the predicted risk index reaches the risk critical value, the causal analysis model is applied to match risk tracing with equipment criticality, a risk early warning grade is generated, and a prevention management scheme is provided.

Description

Pump station equipment risk intelligent prediction method based on big data analysis Technical Field The invention relates to the technical field of fault prediction and health management, in particular to an intelligent pump station equipment risk prediction method based on big data analysis. Background Under the background of modern enterprise resource planning and operation management optimization, how to formulate a set of fault prediction and health management schemes for pump station equipment becomes a key management challenge for optimizing enterprise resource allocation, reducing operation cost and guaranteeing production continuity. Aiming at the formulation of the asset health management strategy, the method generally adopted by the industry is characterized in that the history maintenance record and cost data in the normal working state of equipment are taken as decision basis, and the equipment shutdown risk and the potential economic loss are taken as management targets to form a set of management mechanism for supporting operation and maintenance resource scheduling. However, the existing operation and maintenance management mechanism has fundamental shortages in the design of resource allocation models and decision rules. It is generally assumed that the health assessment criteria of the device is constant and maintenance decisions are made by comparing whether the current operating status data deviates from a preset fixed operating cost threshold. Such management frameworks treat the device as an isolated asset that is separated from its actual commercial operating environment, and existing mechanisms fail to fully account for fluctuations in external commercial operating conditions, and lack quantitative assessment of how these conditions dynamically affect device maintenance requirements, and thus overall operating budget and resource utilization efficiency. Therefore, the method is difficult to grasp the interaction relation between equipment and the running business environment of the equipment from the aspects of enterprise overall operation performance and cost effectiveness, and cannot provide reliable decision support for efficiently maintaining budget allocation and operation risk control. In real commercial operation practice, multidimensional operating environment factors are core variables that directly affect enterprise maintenance expenditures and asset availability. These factors dynamically define a reasonable cost boundary for the normal operating state of the device, which boundary varies in real time with the external market environment and physical operating conditions. For example, during high temperature production seasons or service peaks, the upper acceptable operating cost limits for the motor winding temperature should be dynamically adjusted. However, the existing maintenance decision rule does not take such adaptability change caused by the operation environment transition into consideration, still relies on a static cost baseline set based on past experience as a criterion for triggering maintenance action, which is extremely easy to cause mismatch of maintenance resources, increases operation cost due to excessive maintenance, or causes unplanned shutdown due to insufficient maintenance to cause revenue loss, and finally damages the operation benefit and market competitiveness of enterprises. Therefore, the intelligent pump station equipment risk prediction method based on big data analysis is provided. Disclosure of Invention The invention aims to provide an intelligent pump station equipment risk prediction method based on big data analysis, which is used for fault prediction and health management of pump station equipment. Aiming at the problems of low data acquisition efficiency, insufficient environment interference suppression capability, risk prediction hysteresis and the like in the prior art, the invention realizes real-time evaluation and active maintenance of the health state of equipment through dynamic data acquisition frequency regulation and control, multi-source feature fusion modeling and self-adaptive decision mechanism. The technical scheme of the invention is as follows: collecting operation parameters and environment data of pump station equipment; Constructing a causal analysis model, constructing a causal relation constraint set through a gradient lifting decision tree based on the operation parameters, and identifying the causal relation constraint set through a Bayesian network to obtain parameter association characteristics; Constructing a risk analysis model, performing equipment association topology analysis and time sequence feature analysis on the operation parameters and parameter association features to obtain first operation risk data and first risk critical values, checking based on environment data to obtain operation risk data check values and risk critical check values, and generating operation risk data and risk critical va