[go: up one dir, main page]

CN120822124A - A fault prediction method for oil and gas pressure system in hydropower station by integrating perception data and knowledge graph - Google Patents

A fault prediction method for oil and gas pressure system in hydropower station by integrating perception data and knowledge graph

Info

Publication number
CN120822124A
CN120822124A CN202510992266.7A CN202510992266A CN120822124A CN 120822124 A CN120822124 A CN 120822124A CN 202510992266 A CN202510992266 A CN 202510992266A CN 120822124 A CN120822124 A CN 120822124A
Authority
CN
China
Prior art keywords
oil
data
model
fault
gas pressure
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202510992266.7A
Other languages
Chinese (zh)
Inventor
唐佳庆
李明山
毛哲
范峰
余芳
戴伟民
辛博
周少宁
张莉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shiyan Jiulong Electric Power Group Co ltd
Huanglongtan Hydropower Plant of State Grid Hubei Electric Power Co Ltd
Original Assignee
Shiyan Jiulong Electric Power Group Co ltd
Huanglongtan Hydropower Plant of State Grid Hubei Electric Power Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shiyan Jiulong Electric Power Group Co ltd, Huanglongtan Hydropower Plant of State Grid Hubei Electric Power Co Ltd filed Critical Shiyan Jiulong Electric Power Group Co ltd
Priority to CN202510992266.7A priority Critical patent/CN120822124A/en
Publication of CN120822124A publication Critical patent/CN120822124A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/10—Pre-processing; Data cleansing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2131—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on a transform domain processing, e.g. wavelet transform
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/20—Ensemble learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/042—Knowledge-based neural networks; Logical representations of neural networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/02—Knowledge representation; Symbolic representation
    • G06N5/022—Knowledge engineering; Knowledge acquisition
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/04—Inference or reasoning models
    • G06N5/041—Abduction
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/20—Administration of product repair or maintenance
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computing Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Business, Economics & Management (AREA)
  • General Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Pure & Applied Mathematics (AREA)
  • Medical Informatics (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Analysis (AREA)
  • Computational Mathematics (AREA)
  • Algebra (AREA)
  • Probability & Statistics with Applications (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)

Abstract

本发明属于大数据与人工智能领域，公开了一种融合感知数据与知识图谱的水电站油气压力系统故障预测方法，其特征在于：包含有以下步骤：多模态数据采集与预处理；特征提取与跨模态关联分析；知识图谱构建与动态更新；物理机理模型与数据驱动模型融合；故障模式库构建与知识推理；多模态预测模型训练与优化；故障预警与维护策略生成。本发明具有以下主要有益技术效果：提升了故障预测的精准性，提高了故障根因分析能力，增强了综合感知与认知能力，增强了决策支持能力，提升了模型的可解释性与可信度，实现了全流程智能运维闭环。

This invention belongs to the field of big data and artificial intelligence, and discloses a method for predicting faults in the oil and gas pressure system of a hydropower station that integrates sensory data and a knowledge graph. The method is characterized by comprising the following steps: multimodal data acquisition and preprocessing; feature extraction and cross-modal correlation analysis; knowledge graph construction and dynamic updating; fusion of physical mechanism models and data-driven models; fault pattern library construction and knowledge reasoning; multimodal prediction model training and optimization; and fault warning and maintenance strategy generation. The invention has the following major beneficial technical effects: improved fault prediction accuracy, enhanced fault root cause analysis capabilities, enhanced comprehensive perception and cognitive capabilities, enhanced decision support capabilities, improved model interpretability and credibility, and achieved a closed-loop intelligent operation and maintenance process.

Description

Hydropower station oil-gas pressure system fault prediction method integrating perception data and knowledge graph
Technical Field
The invention belongs to the field of big data and artificial intelligence, in particular relates to a hydropower station oil-gas pressure system fault prediction method integrating sensing data and a knowledge map, and aims to realize early fault early warning and predictive maintenance of equipment through multi-sensor data sensing and dynamic knowledge reasoning by combining mechanical performance degradation analysis, multi-source heterogeneous data fusion, knowledge map construction and a multi-mode sensing model.
Background
The pressure oil tank and the pressure air tank of the oil-gas pressure system of the hydropower station are connected in a gas-liquid linkage manner, so that the cooperative adjustment of pressure and air is realized, and the stability and flexibility of the system are improved. Hydropower station oil-gas pressure system as the core auxiliary system of the hydroelectric generating set, the running state directly affects the power generation efficiency and safety. The traditional operation and maintenance mode relies on manual inspection and periodic maintenance. Although most hydropower stations establish a data acquisition platform for acquiring data such as vibration, oil and gas equipment of the hydropower station, the related data are not associated and are difficult to uniformly analyze.
In the prior art, LSTM and transducer models are mostly adopted for predicting related sensing data, and faults of the sensing data are frequently abnormal through sensing data change abnormality, but cross-modal information such as oil liquid state, environmental parameters and the like is ignored. And the single time sequence prediction algorithm is difficult to capture the degradation trend of the complex oil-gas pressure system and the degradation mechanism of fatigue crack and seal failure. In addition, during the operation of hydroelectric power generation, the working condition environment of the oil-gas pressure system is not constant along with environmental change and operation maintenance, the influence of the oil-gas components on the pressure system is frequently changed, and the related operation knowledge graph is required to be continuously updated so as to avoid the failure of the prediction model.
CN117150032A discloses an intelligent maintenance system and method of a hydropower station generator set, and relates to the field of hydropower station equipment management; the system comprises a database construction module, a periodic maintenance module, a precision maintenance module and a component diagnosis list, wherein the database construction module is used for collecting data in a generator set and constructing a database and a database model based on the data, the periodic maintenance module is used for automatically carrying out equipment state evaluation and defect statistical analysis in a preset time period, generating a corresponding report, automatically pushing and executing a maintenance work list according to a periodic work list of equipment, and the precision maintenance module is used for intelligently tracking and comprehensively judging the running state of the equipment, alarming when the equipment is abnormal and automatically generating the component diagnosis list. The device can realize the accurate positioning of abnormal equipment, the automatic triggering of an abnormal processing flow and the accurate maintenance of the equipment, reduce the unnecessary loss of the equipment, and improve the sensing accuracy and the high efficiency of equipment defects and abnormal conditions. However, there is still room for improvement in failure prediction.
In view of the above, the invention integrates heterogeneous data such as vibration, oil, pressure, temperature and the like to construct a unified feature space. And dynamically updating causal relation and fault mode in the knowledge graph based on the degradation characteristics, and eliminating noise interference by combining a physical mechanism model and a deep learning model, thereby improving the accuracy and the robustness of the fault prediction of the oil and gas system.
Disclosure of Invention
In order to solve the problems, the invention aims to disclose a hydropower station oil gas pressure system fault prediction method integrating perception data and a knowledge graph, which is realized by adopting the following technical scheme.
A hydropower station oil-gas pressure system fault prediction method integrating perception data and knowledge maps is characterized by comprising the following steps:
step 1, multi-mode data acquisition and preprocessing;
step 2, feature extraction and cross-modal correlation analysis;
Step 3, knowledge graph construction and dynamic updating;
Step 4, fusing a physical mechanism model and a data driving model;
step 5, constructing a fault mode library and reasoning knowledge;
step 6, training and optimizing the multi-mode prediction model;
and 7, generating fault early warning and maintenance strategies.
The invention improves the dynamic property of generalization of the failure prediction of the hydro-pneumatic pressure system of the hydropower station.
The invention has the following main beneficial technical effects:
1. the accuracy of fault prediction is improved, namely hidden correlations among different sensor data and knowledge graph information are revealed through cross-modal correlation analysis, and early and weak fault characteristics are captured.
2. The fault root analysis capability is improved, namely, when potential faults are predicted, possible fault components, failure modes and propagation paths thereof can be quickly traced through the topological structure of the knowledge graph and rich associated information, so that the problem core can be accurately positioned.
3. The comprehensive perception and cognition capability is enhanced, namely, through fusing multi-mode perception data (such as sensor data of vibration, temperature, pressure and the like) and a knowledge graph (comprising structural knowledge of equipment structure, historical faults, maintenance records, expert rules and the like), the system state is more comprehensively and deeply understood, and the limitation of a single data source or model is broken through.
4. The decision support capability is enhanced, namely the fault early warning and maintenance strategy generation part not only predicts the fault probability, but also fuses the associated information (such as spare part condition, maintenance cost, influence range and best maintenance practice) provided by the knowledge graph to generate more intelligent and operable maintenance proposal and coping strategies, so as to assist operation and maintenance personnel to make optimal decisions and reduce unplanned downtime.
5. The interpretation and the credibility of the model are improved, namely the knowledge graph is used as a background knowledge base, interpretation and support of a semantic layer are provided for the prediction result of the data driving model, and the credibility and the user acceptance of the model result are enhanced.
6. The method forms a complete intelligent analysis and prediction closed loop from data acquisition, processing, analysis and model construction to final early warning and decision support, and remarkably improves the intelligent level and efficiency of operation and maintenance of key equipment (oil gas pressure system) of the hydropower station.
Drawings
FIG. 1 is a schematic flow chart of the method of the present application.
Detailed Description
So that those skilled in the art can better understand and practice the present patent, a detailed description of embodiments of the application will be presented with reference to the drawings.
Please refer to fig. 1, a hydropower station oil gas pressure system fault prediction method integrating sensing data and a knowledge graph is characterized by comprising the following steps:
step 1, multi-mode data acquisition and preprocessing;
step 2, feature extraction and cross-modal correlation analysis;
Step 3, knowledge graph construction and dynamic updating;
Step 4, fusing a physical mechanism model and a data driving model;
step 5, constructing a fault mode library and reasoning knowledge;
step 6, training and optimizing the multi-mode prediction model;
and 7, generating fault early warning and maintenance strategies.
The invention improves the dynamic property of generalization of the failure prediction of the hydro-pneumatic pressure system of the hydropower station.
The embodiment of the invention is described in detail as follows:
step 1: the multi-mode data acquisition and preprocessing, namely sensing state data of the multi-mode data in real time through vibration, liquid level and pressure sensors arranged on an oil gas pressure oil tank, eliminating fluctuation and noise of the sensed data, and smoothing the sensed data by adopting a filtering algorithm, wherein the specific substeps are as follows:
Step 1-1, installing vibration sensors at the joints of a bearing seat of a pressure oil tank, a gear box shell and a sealing element, monitoring the vibration frequency of equipment in real time, adopting a magnetic oil particle sensor to detect the concentration of metal particles, adopting a spectrum analyzer to detect the index and the liquid level of oil oxidation products, acquiring the pressure at the oil pump outlet, the top of the oil tank and the joint of the sealing element, adopting an infrared thermal imager to monitor the temperature distribution of the surface of the oil tank, synchronizing the data of the multiple sensors to a centralized monitoring center through an industrial Ethernet (such as Profinet) or an OPC UA protocol, and sampling each sensor by 1KHz, wherein the numerical record is as the following formula:
,
Wherein the method comprises the steps of Is a time domain sensing signal (which can be a vibration signal, a pressure signal and an oil index),For the amplitude of the i-th sensor,For the sampling frequency of the i-th sensor,For the phase value used for calibration for the ith sensor.
Step 1-2, carrying out N layers of discrete wavelet decomposition on the sensing signal to obtain corresponding detail coefficients and approximation coefficients, wherein the detail coefficients and approximation coefficients are shown in the formula:
,
Wherein the method comprises the steps of Coefficients for the N-th layer wavelet decomposition, i.e., associated low frequency components; and the detail coefficient of the k layer corresponds to high-frequency components with different scales.
The decomposition coefficients are processed according to soft threshold values:
,
,
c is the original coefficient (e.g And),As a result of the threshold value being set,The standard deviation of noise, M is the signal length;
reconstructing a sensing signal, wherein the signal strength after noise elimination is as follows:
,
Step 1-3, performing Z-score standardization (the mean value is 0 and the standard deviation is 1) on sensing data of an oil gas pressure system so as to ensure that the data of different sensors are in the same dimension, wherein the method is shown as follows, and adopting a3 sigma principle to reject abnormal values (such as points exceeding the mean value plus or minus 3 times of the standard deviation) in the data of the pressure sensors:
Wherein the method comprises the steps of AndThe mean and variance, respectively.
The step 2 of feature extraction and cross-modal correlation analysis is to extract time domain features, frequency domain features and operation related features of mechanical abrasion and vibration of an oil gas pressure system reflected by a sensing signal so as to realize cross-modal alignment of a multi-source sensor and better serve early fault detection, and specifically comprises the following sub-steps:
Extracting and calculating time domain characteristics of each sensing signal of the oil gas pressure system, wherein the time domain characteristics comprise root mean square value (RMS), peak-to-peak value (PPV) and Kurtosis (Kurtosis) characteristics, and the time domain characteristics are used for detecting bearing wear or gear cracks;
the root mean square value is calculated as follows to measure the energy intensity and vibration intensity of the signal:
,
The method comprises the steps that (1) the ith sampling point of a signal x is obtained, and N is the total number of sampling points used for calculating the signal;
the peak-to-peak calculation formula is as follows to measure the dynamic range and impact characteristics of the signal, especially abnormal characteristics:
,
The kurtosis characteristic calculation formula is as follows to reflect the signal spike characteristics:
,
as has been shown in the foregoing, AndThe mean and variance of the signal, respectively.
And 2-2, extracting main frequency components by adopting Fast Fourier Transform (FFT), and detecting early faults such as crack resonance and the like by combining envelope analysis, wherein the early faults are shown in the formula:
,
Wherein the method comprises the steps of For the frequency signal, f is the frequency variable, N is the nth frequency cost, and N is the total frequency component.
And 2-3, extracting oil related characteristics including temperature difference between the surface of an oil tank and the ambient temperature, acid value and moisture content characteristics of oil spectrum analysis and metal particle concentration characteristics.
And 2-4, analyzing the relation between the vibration signal and the concentration of oil particles by adopting a Pearson correlation coefficient, and constructing a coupling relation between the concentration of metal particles existing in oil pump abrasion and vibration spectrum deviation for the follow-up knowledge graph, wherein the coupling relation is specifically shown as a formula:
,
r is a correlation coefficient, xi and yi are values of two sensors or modalities respectively, AndRespectively the average of n samples.
Step 3, knowledge graph construction and dynamic updating, wherein the step is to construct a fault-related knowledge graph according to the working mechanism of the oil gas pressure system, provide a more definite entity relationship and an association mapping relationship between performance and faults, and dynamically update in time, and specifically comprises the following sub-steps:
Step 3-1, extracting equipment entities (such as an oil gas pressure tank, an oil pump, a sealing element and a bearing) and fault types (such as seal failure, bearing abrasion and oil pump cavitation) by adopting a named entity method:
,
Wherein the method comprises the steps of As an entity type (e.g. "device"),As an attribute (such as "pressure"),Is an attribute value (e.g., "1.5 MPa").
And 3-2, extracting monitoring relations and causal relations between equipment entity faults and perceived data features according to the working mechanism specification, wherein the monitoring relations (such as { "oil liquid sensor", "monitoring", "oil liquid particle concentration" }) and causal relations (such as { "seal wear", "causal", "oil liquid leakage" } and { "oil liquid leakage", "causal", "pressure fluctuation" }).
Step 3-3, extracting equipment faults, associated indexes and causal relations from a working mechanism specification and technical analysis maintenance manual by adopting a large language model or a natural language processing tool, constructing a related knowledge graph, and carrying out characterization modeling on entity relations by adopting a graph neural network:
,
Wherein the method comprises the steps of A vector is embedded for a layer i node,In order to be a contiguous matrix,In order for the weight matrix to be learnable,To activate the function.
And 3-4, updating a knowledge graph (such as abrasion of a sealing element when the concentration of oil particles is more than 13) according to the monitoring attribute indexes and fault phenomena in the real-time operation process, wherein the method specifically comprises the following steps:
,
Wherein the method comprises the steps of For newly added entities or attributes (such as "oil pump", "particle concentration"),Is a new relationship (such as connection, cause and effect).
And 4, fusing a physical model with a data model, further constructing a physical mechanism model by utilizing the knowledge graph generated in the step 3, fusing a data-driven prediction model, simulating oil flow and a pressure model of an oil-gas pressure system to eliminate noise influence, and simultaneously predicting outlet pressure fluctuation of an oil pump better to accurately predict faults of the oil-gas system, wherein the method specifically comprises the following sub-steps:
Step 4-1, establishing an oil fluid density, speed and pressure correlation model by adopting a fluid dynamics model, and predicting the pressure fluctuation of an oil gas pressure tank port:
,
Wherein the method comprises the steps of In order to achieve a fluid density,Is the fluid velocity,Is pressure (pressure),Is of dynamic viscosity,Is an external force.
Step 4-2, calculating the oil tank surface temperature distribution by using a Fourier heat conduction equation (∂ T/∂ T=alpha (∂ 2T/∂ x 2)), and predicting the heat aging rate of the sealing element by combining the heat boundary conditions (such as the convective heat transfer coefficient h=10W/m 2K).
Step 4-3, partitioning the sensing signal data according to 1024 points, and extracting a time sequence dependency relationship by adopting a transducer model so as to be used for predicting the residual service life of bearing abrasion:
,
, respectively input data, query, key, value matrix.
And 4-4, carrying out weighted fusion on the data obtained by the physical model and the predicted value of the data model so as to reduce the influence of data noise on the predicted result:
,
such as oil and gas port pressure physical model Simulating and calculating the obtained value for the dynamics model obtained in the step 4-1,And (3) obtaining a predicted value obtained in the step (4-3), wherein alpha is a model weight and can be taken as 0.6.
Step 5, constructing a fault mode library and knowledge reasoning, wherein the step extracts a fault mode based on the steps and the historical data, and builds a related network model for fault classification and causal reasoning, and the method specifically comprises the following substeps:
and 5.1, constructing a fault mode library based on historical fault cases, and classifying the fault modes into mechanical faults and oil faults, wherein the mechanical faults comprise bearing wear, gear breakage and seal aging, the oil faults comprise emulsification, oxidization and water exceeding, and the corresponding parameter characteristics of the fault modes are as follows:
,
Wherein the method comprises the steps of Based on the pattern library, the k-th fault characteristic (such as 'oil pressure abnormality') is used for classifying historical fault data through a machine learning method such as random forest and the like, and a new fault pattern (such as 'oil pump cavitation') is automatically supplemented.
And 5.2, carrying out probability prediction reasoning on the oil gas faults according to attribute characteristics by adopting a Monte Carlo simulation method based on a fault mode library:
,
Wherein the method comprises the steps of The posterior probability after the occurrence of the feature, such as carrying out Monte Carlo simulation on the oil tank temperature field for a plurality of times, predicts the probability (such as 95%) that the thermal stress of the sealing element exceeds a critical value (> 150 MPa%), and triggers maintenance early warning.
And 5.3, performing causal reasoning by adopting a Bayesian network, and determining the occurrence probability of related faults according to attribute indexes, such as a formula:
,
If the probability of the oil particle concentration >12 is 0.8 and the vibration spectrum deviation >20% is approximately 0.7, the probability of bearing wear is calculated (p=0.92).
And 6, training and optimizing the multi-mode prediction model, wherein the step is to train and verify the end-to-end model related to the step, design a loss function aiming at the training result of the model so as to balance minimum mean square error and cross entropy, optimize and adjust training learning rate aiming at data characteristics, and verify the result, and specifically comprises the following sub-steps.
Step 6-1, taking a weighted index as a loss function, and balancing a minimum Mean Square Error (MSE) and a cross entropy (CrossEntropy), optimizing the prediction accuracy of a physical model through the MSE, and meeting the probability prediction accuracy of fault type identification through CrossEntropy:
,
Wherein the method comprises the steps of Weights are the penalty functions.
Step 6-2, adjusting the learning rate of a transducer prediction model by using Bayesian optimization, adding an L2 regularization (lambda=0.01) and an early-stop mechanism (patience =10) in model training to prevent overfitting, and simultaneously adopting an Adam optimizer to perform gradient descent optimization training, wherein the following formula is as follows:
,
Wherein the method comprises the steps of AndModel parameters and learning rates, respectively.
And 6-3, adopting 5-fold cross validation to ensure the generalization capability of the model under different working conditions (such as high load and low load), and evaluating the accuracy, F1 fraction and recall rate of the model.
Step 7, generating fault early warning and maintenance strategies, wherein the step adopts a Z-score to calculate a threshold according to historical data, and generates corresponding maintenance actions according to different fault probabilities so as to determine immediate shutdown checking or regular maintenance, and the method specifically comprises the following steps of:
step 7-1, calculating and setting a fault early warning threshold value from historical data by adopting a Z-score:
,
if the short-term trend threshold of the oil particle concentration is EMA+3σ (e.g. 150 μm) using exponential moving average calculation, early warning is started.
And 7-2, generating maintenance suggestions according to the fault probability, such as high-probability immediate stop check (when the probability of the oxidation degree of oil liquid >8 exceeds 0.8, immediate stop maintenance and oil liquid replacement), and low-probability periodic maintenance (such as RUL >200 hours and periodic maintenance every 150 hours):
。
the invention has the following main beneficial technical effects:
1. the accuracy of fault prediction is improved, namely hidden correlations among different sensor data and knowledge graph information are revealed through cross-modal correlation analysis, and early and weak fault characteristics are captured.
2. The fault root analysis capability is improved, namely, when potential faults are predicted, possible fault components, failure modes and propagation paths thereof can be quickly traced through the topological structure of the knowledge graph and rich associated information, so that the problem core can be accurately positioned.
3. The comprehensive perception and cognition capability is enhanced, namely, through fusing multi-mode perception data (such as sensor data of vibration, temperature, pressure and the like) and a knowledge graph (comprising structural knowledge of equipment structure, historical faults, maintenance records, expert rules and the like), the system state is more comprehensively and deeply understood, and the limitation of a single data source or model is broken through.
4. The decision support capability is enhanced, namely the fault early warning and maintenance strategy generation part not only predicts the fault probability, but also fuses the associated information (such as spare part condition, maintenance cost, influence range and best maintenance practice) provided by the knowledge graph to generate more intelligent and operable maintenance proposal and coping strategies, so as to assist operation and maintenance personnel to make optimal decisions and reduce unplanned downtime.
5. The interpretation and the credibility of the model are improved, namely the knowledge graph is used as a background knowledge base, interpretation and support of a semantic layer are provided for the prediction result of the data driving model, and the credibility and the user acceptance of the model result are enhanced.
6. The method forms a complete intelligent analysis and prediction closed loop from data acquisition, processing, analysis and model construction to final early warning and decision support, and remarkably improves the intelligent level and efficiency of operation and maintenance of key equipment (oil gas pressure system) of the hydropower station.
The above-described embodiments are only preferred embodiments of the present invention, and should not be construed as limiting the present invention. The protection scope of the present invention is defined by the claims, and the protection scope includes equivalent alternatives to the technical features of the claims. I.e., equivalent replacement modifications within the scope of this invention are also within the scope of the invention.

Claims (8)

1.一种融合感知数据与知识图谱的水电站油气压力系统故障预测方法，其特征在于：包含有以下步骤：1. A method for predicting faults in the oil and gas pressure system of a hydropower station by integrating sensory data and knowledge graphs, characterized by comprising the following steps: 步骤1：多模态数据采集与预处理；Step 1: Multimodal data acquisition and preprocessing; 步骤2：特征提取与跨模态关联分析；Step 2: Feature extraction and cross-modal correlation analysis; 步骤3：知识图谱构建与动态更新；Step 3: Knowledge graph construction and dynamic update; 步骤4：物理机理模型与数据驱动模型融合；Step 4: Fusion of physical mechanism model and data-driven model; 步骤5：故障模式库构建与知识推理；Step 5: Fault mode library construction and knowledge reasoning; 步骤6：多模态预测模型训练与优化；Step 6: Multimodal prediction model training and optimization; 步骤7：故障预警与维护策略生成。Step 7: Generate fault warning and maintenance strategy. 2.根据权利要求1所述的一种融合感知数据与知识图谱的水电站油气压力系统故障预测方法，其特征在于：步骤1中，通过安装在油气压力油罐的震动、液位和压力传感器实时感知其状态数据，并消除感知数据波动和噪声，采用滤波算法对感知数据进行平滑，具体子步骤如下：2. The method for predicting faults in a hydropower station oil and gas pressure system by integrating sensory data with a knowledge graph according to claim 1 is characterized in that: in step 1, vibration, liquid level, and pressure sensors installed on the oil and gas pressure tank are used to sense its status data in real time, and fluctuations and noise in the sensed data are eliminated. A filtering algorithm is used to smooth the sensed data. The specific sub-steps are as follows: 步骤1-1：在压力油罐的轴承座、齿轮箱外壳和密封件连接处安装振动传感器，实时监测设备振动频率，采用磁性油液颗粒传感器检测金属颗粒浓度、采用光谱分析仪检测油液氧化产物指标和液位，在油泵出口、油罐顶部和密封件接口处压力传感器采集压力，采用红外热成像仪监测油罐表面温度分布，通过工业以太网或OPC UA协议将上述传感器数据同步至集中监控中心，各传感器以1KHz进行采样，其数值记录为以下公式：，其中为时域感知信号，为第i个传感器的幅值，为第i个传感器的采样频率，为第i个传感器用于校准的相位值；Step 1-1: Install vibration sensors on the bearing seat, gearbox housing, and seal connections of the pressure oil tank to monitor the equipment vibration frequency in real time. Use a magnetic oil particle sensor to detect the metal particle concentration, and a spectrum analyzer to detect the oil oxidation product index and liquid level. Pressure sensors are installed at the oil pump outlet, the top of the oil tank, and the seal interface to collect pressure. Use an infrared thermal imager to monitor the surface temperature distribution of the oil tank. Synchronize the above sensor data to the centralized monitoring center via Industrial Ethernet or OPC UA protocol. Each sensor samples at 1 kHz, and its value is recorded as follows: ,in is the time domain perception signal, is the amplitude of the i-th sensor, is the sampling frequency of the i-th sensor, is the phase value of the i-th sensor used for calibration; 步骤1-2：对感知信号进行N层离散小波分解，得到相应的细节系数和近似系数，如公式所示：，其中为第N层小波分解的系数，即相关的低频分量；为第k层的细节系数，对应不同尺度的高频分量；Step 1-2: Perform N-layer discrete wavelet decomposition on the perception signal to obtain the corresponding detail coefficients and approximation coefficients, as shown in the formula: ,in is the coefficient of the Nth layer wavelet decomposition, that is, the relevant low-frequency component; is the detail coefficient of the kth layer, corresponding to high-frequency components of different scales; 对分解系数按软阈值进行处理：，，c为原始系数（如和），为阈值，为噪声标准差，M为信号长度；Process the decomposition coefficients using soft thresholding: , , c is the original coefficient (such as and ), is the threshold, is the noise standard deviation, M is the signal length; 对感知信号进行重构，消除噪声后的信号强度为：；The signal strength after reconstructing the perception signal and eliminating noise is: ; 步骤1-3：对油气压力系统感知数据进行Z-score标准化，标准化方法为：均值为0，标准差为1；以确保不同传感器的数据在同一量纲下，如下式所示：并采用3σ原则剔除压力传感器数据中的异常值：，其中和分别为均值和方差。Step 1-3: Perform Z-score normalization on the oil and gas pressure system sensing data. The normalization method is: mean 0, standard deviation 1; to ensure that the data from different sensors are in the same dimension, as shown in the following formula: And use the 3σ principle to eliminate outliers in the pressure sensor data: ,in and are the mean and variance respectively. 3.根据权利要求1所述的一种融合感知数据与知识图谱的水电站油气压力系统故障预测方法，其特征在于：步骤2中，提取感知信号反映出来的油气压力系统的机械磨损和振动的时域特征、频域特征和运行相关特征，以实现多源传感器的跨模态对齐，更好服务于早期故障检测；具体包括以下子步骤：3. The method for predicting faults in the oil and gas pressure system of a hydropower station by integrating sensory data with a knowledge graph according to claim 1 is characterized by: in step 2, the time domain characteristics, frequency domain characteristics, and operation-related characteristics of the mechanical wear and vibration of the oil and gas pressure system reflected in the sensory signals are extracted to achieve cross-modal alignment of multi-source sensors, thereby better serving early fault detection; the method specifically comprises the following sub-steps: 步骤2-1：提取计算油气压力系统各感知信号的时域特征，具体包括均方根值、峰峰值和峭度特征，以用作检测轴承磨损或齿轮裂纹；Step 2-1: Extract and calculate the time domain features of each sensing signal of the oil and gas pressure system, including the root mean square value, peak-to-peak value, and kurtosis features, to detect bearing wear or gear cracks; 均方根值计算公式如下，以衡量信号的能量强度和振动强度：，为信号x的第i个采样点，N为用于计算信号采样点总数；The RMS value is calculated as follows to measure the energy intensity and vibration intensity of the signal: , is the i-th sampling point of signal x, and N is the total number of sampling points used to calculate the signal; 峰峰值计算公式如下，用以衡量信号的动态范围和冲击特性，尤其是异常特性：；The peak-to-peak value calculation formula is as follows, which is used to measure the dynamic range and impact characteristics of the signal, especially the abnormal characteristics: ; 峭度特征计算公式如下，以反映信号尖峰特性：，如前所示，和分别为信号的均值和方差；The kurtosis characteristic calculation formula is as follows to reflect the signal peak characteristics: , as shown above, and are the mean and variance of the signal respectively; 步骤2-2：采用快速傅里叶变换提取主要频率成分，并结合包络分析检测裂纹共振等早期故障，如公式所示：，其中为频率信号，f为频率变量，n为第n个频率成本，N为总频率成分；Step 2-2: Use fast Fourier transform to extract the main frequency components and combine it with envelope analysis to detect early faults such as crack resonance, as shown in the formula: ,in is the frequency signal, f is the frequency variable, n is the nth frequency cost, and N is the total frequency component; 步骤2-3：提取油液相关的特征，包括油罐表面与环境温度的温差，油液光谱分析的酸值和水分含量特征，及金属颗粒浓度特征；Step 2-3: Extract oil-related features, including the temperature difference between the tank surface and the ambient temperature, the acid value and moisture content characteristics of the oil spectral analysis, and the metal particle concentration characteristics; 步骤2-4：采用皮尔逊相关系数分析振动信号与油液颗粒浓度的关系，为后续知识图谱构造抽取油泵磨损存在的金属颗粒浓度与振动频谱偏移的耦合关系，具体如公式所示：，r为相关系数，xi和yi分别为两种传感器或模态的数值，和分别为n个样本的均值。Step 2-4: Use the Pearson correlation coefficient to analyze the relationship between the vibration signal and the oil particle concentration. This is used to extract the coupling relationship between the metal particle concentration and the vibration spectrum shift caused by oil pump wear for subsequent knowledge graph construction. The specific relationship is shown in the formula: , r is the correlation coefficient, xi and yi are the values of the two sensors or modes respectively, and are the means of n samples respectively. 4.根据权利要求1所述的一种融合感知数据与知识图谱的水电站油气压力系统故障预测方法，其特征在于：步骤3中，根据油气压力系统工作机理构建故障相关的知识图谱，提供更加明确的实体关系及性能表现与故障间的关联映射关系，并及时进行动态更新，具体包括以下子步骤：4. The method for predicting faults in a hydropower station oil and gas pressure system by integrating sensory data and a knowledge graph according to claim 1 is characterized in that: in step 3, a fault-related knowledge graph is constructed based on the working mechanism of the oil and gas pressure system, providing a more clear entity relationship and correlation mapping relationship between performance and faults, and dynamically updating it in a timely manner, specifically comprising the following sub-steps: 步骤3-1：采用命名实体方法提取设备实体和故障类型：，其中为实体类型（如“设备”），为属性（如“压力”），为属性值（如“1.5MPa”）；Step 3-1: Use named entity method to extract equipment entities and fault types: ,in is the entity type (e.g., "device"), is a property (such as "pressure"), is the attribute value (such as "1.5MPa"); 步骤3-2：根据工作机理规范抽取设备实体故障与感知数据特征间的监测关系和因果关系，监测关系和因果关系；Step 3-2: Extract the monitoring relationship and causal relationship between the equipment entity fault and the perception data features according to the working mechanism specification; 步骤3-3：采用大语言模型或自然语言处理工具从工作机理规范和技术解析维护手册抽取设备故障，关联指标及因果关系，构建相关知识图谱，并采用图神经网络对实体关系进行表征建模：，其中为第l层节点嵌入向量，为邻接矩阵，为可学习权重矩阵，为激活函数；Step 3-3: Use a large language model or natural language processing tool to extract equipment failures, related indicators, and causal relationships from the work mechanism specifications and technical analysis maintenance manuals, build relevant knowledge graphs, and use graph neural networks to represent and model entity relationships: ,in is the embedding vector of the l-th layer node, is the adjacency matrix, is the learnable weight matrix, is the activation function; 步骤3-4：根据实时运行过程中的监测属性指标和故障现象，更新知识图谱，具体表示为：，其中为新增实体或属性（如“油泵”、“颗粒浓度”），为新增关系（如连接、因果）。Step 3-4: Update the knowledge graph based on the monitoring attribute indicators and fault phenomena during real-time operation, specifically expressed as follows: ,in To add new entities or attributes (such as "oil pump", "particle concentration"), To add new relationships (such as connection, cause and effect). 5.根据权利要求1所述的一种融合感知数据与知识图谱的水电站油气压力系统故障预测方法，其特征在于：步骤4中，利用步骤3生成的知识图谱进一步构建物理机理模型，融合数据驱动预测模型，模拟油气压力系统的油液流动和压力模型，以消除噪声影响，同时更好预测油泵出口压力波动，准确对油气系统故障进行预测，具体包括如下子步骤：5. The method for predicting oil and gas pressure system faults in a hydropower station by integrating sensory data and a knowledge graph according to claim 1 is characterized in that: in step 4, a physical mechanism model is further constructed using the knowledge graph generated in step 3, and the data-driven prediction model is integrated to simulate the oil flow and pressure model of the oil and gas pressure system to eliminate the influence of noise, while better predicting oil pump outlet pressure fluctuations and accurately predicting oil and gas system faults. The method specifically includes the following sub-steps: 步骤4-1：采用流体动力学模型，建立油液流体密度、速度、压力关联模型，并对油气压力罐口压力波动进行预测：，其中为流体密度，为流体速度、为压力、为动力粘度、为外部力；Step 4-1: Use the fluid dynamics model to establish a correlation model between oil fluid density, velocity, and pressure, and predict the pressure fluctuation at the oil and gas pressure tank port: ,in is the fluid density, is the fluid velocity, For pressure, is the dynamic viscosity, is an external force; 步骤4-2：采用傅里叶热传导方程：∂T/∂t = α(∂²T/∂x²)，计算油罐表面温度分布，结合热边界条件预测密封件热老化速率；Step 4-2: Use the Fourier heat conduction equation: ∂T/∂t = α(∂²T/∂x²) to calculate the surface temperature distribution of the oil tank and predict the thermal aging rate of the seal in combination with the thermal boundary conditions; 步骤4-3：将感知信号数据按1024点进行分块，并采用Transformer模型提取时序依赖关系，以用于预测轴承磨损的剩余使用寿命：，，分别为输入数据，查询、键、值矩阵；Step 4-3: Divide the sensor signal data into 1024-point blocks and use the Transformer model to extract temporal dependencies for predicting the remaining service life of bearing wear: , , are input data, query, key, and value matrices respectively; 步骤4-4：将物理模型获得数据和数据模型预测值进行加权融合，以减少数据噪声对预测结果的影响：，如油气口压力物理模型为步骤4-1所得动力学模型模拟计算所得值，为步骤4-3所得预测值；α为模型权重，取为0.6。Step 4-4: Perform weighted fusion of the data obtained from the physical model and the predicted value of the data model to reduce the impact of data noise on the prediction results: , such as the oil and gas port pressure physical model is the value obtained by kinetic model simulation calculation in step 4-1, is the predicted value obtained in step 4-3; α is the model weight, which is taken as 0.6. 6.根据权利要求1所述的一种融合感知数据与知识图谱的水电站油气压力系统故障预测方法，其特征在于：步骤5中，基于前述步骤和历史数据提取故障模式，为故障分类和因果推理建立相关的网络模型，具体包括以下子步骤：6. The method for predicting oil and gas pressure system faults in a hydropower station by integrating sensory data and knowledge graphs according to claim 1 is characterized in that: in step 5, based on the previous steps and historical data, fault modes are extracted and a relevant network model for fault classification and causal reasoning is established, which specifically includes the following sub-steps: 步骤5.1：基于历史故障案例构建故障模式库，将故障模式分类为机械故障和油液故障，其中机械故障包括轴承磨损、齿轮断裂和密封件老化，油液故障包括乳化、氧化和水分超标，与故障模式对应的为相关的参数特征，如下所示：，其中为第k个故障特征，以此模式库为基础，通过随机森林等机器学习方法对历史故障数据进行分类，并自动补充新故障模式；Step 5.1: Build a failure mode library based on historical failure cases and classify the failure modes into mechanical failures and oil failures. Mechanical failures include bearing wear, gear breakage, and seal aging, while oil failures include emulsification, oxidation, and excessive moisture. The corresponding parameter features for each failure mode are as follows: ,in The kth fault feature is used as the basis for classifying historical fault data through machine learning methods such as random forest and automatically adding new fault patterns. 步骤5.2：采用蒙特卡洛模拟方法基于故障模式库，根据属性特征对油气故障进行概率预测推理：，其中为特征出现后的后验概率，如对油罐温度场进行多次蒙特卡洛模拟，预测密封件热应力超过临界值的概率，触发维护预警；Step 5.2: Use the Monte Carlo simulation method to perform probabilistic prediction reasoning on oil and gas faults based on the fault pattern library and attribute characteristics: ,in The posterior probability after the feature appears, such as performing multiple Monte Carlo simulations on the temperature field of the oil tank to predict the probability that the thermal stress of the seal exceeds the critical value, triggering a maintenance warning; 步骤5.3：采用贝叶斯网络进行因果推理，根据属性指标确定相关故障出现概率，如公式：。Step 5.3: Use Bayesian network for causal reasoning and determine the probability of related faults based on attribute indicators, as shown in the formula: . 7.根据权利要求1所述的一种融合感知数据与知识图谱的水电站油气压力系统故障预测方法，其特征在于：步骤6中，步骤是对上述步骤所涉及的端到端模型进行训练验证，针对上述模型训练结果设计损失函数，以平衡最小均方误差和交叉熵，并针对数据特征优化调整训练学习率，并对结果进行验证，具体包括以下子步骤：7. The method for predicting oil and gas pressure system faults in a hydropower station by integrating sensory data and a knowledge graph according to claim 1 is characterized in that: in step 6, the end-to-end model involved in the above steps is trained and verified, a loss function is designed based on the training results of the above model to balance the minimum mean square error and cross entropy, the training learning rate is optimized and adjusted based on the data characteristics, and the results are verified, which specifically includes the following sub-steps: 步骤6-1：以加权指标作为损失函数，目的是平衡最小均方误差和交叉熵，通过MSE优化物理模型的预测精度，通过CrossEntropy满足故障类型识别的概率预测准确性：，其中为损失函数权重；Step 6-1: Use weighted indicators as the loss function to balance the minimum mean square error and cross entropy, optimize the prediction accuracy of the physical model through MSE, and meet the probability prediction accuracy of fault type identification through CrossEntropy: ,in is the loss function weight; 步骤6-2：使用贝叶斯优化调整Transformer预测模型的学习率，并在模型训练中加入L2正则化和早停机制，防止过拟合，同时采用Adam优化器进行梯度下降优化训练，如下式：，其中和分别为模型参数和学习率；Step 6-2: Use Bayesian optimization to adjust the learning rate of the Transformer prediction model, and add L2 regularization and early stopping mechanism to the model training to prevent overfitting. At the same time, use the Adam optimizer for gradient descent optimization training, as shown below: ,in and are model parameters and learning rate respectively; 步骤6-3：采用5折交叉验证，确保模型在不同工况下的泛化能力，并对模型的准确率、F1分数和召回率进行评估。Step 6-3: Use 5-fold cross-validation to ensure the generalization ability of the model under different working conditions and evaluate the model's accuracy, F1 score, and recall rate. 8.根据权利要求1所述的一种融合感知数据与知识图谱的水电站油气压力系统故障预测方法，其特征在于：步骤7中，根据历史数据采用Z-score计算阈值，并根据不同故障概率生成对应的维护动作，以确定立即停机检查或定期维护，具体包括以下步骤：8. The method for predicting oil and gas pressure system faults in a hydropower station by integrating sensory data and knowledge graphs according to claim 1 is characterized in that: in step 7, a threshold is calculated using a Z-score based on historical data, and corresponding maintenance actions are generated based on different fault probabilities to determine whether to shut down the system for immediate inspection or scheduled maintenance, specifically comprising the following steps: 步骤7-1：采用Z-score从历史数据中计算设定故障预警阈值：，使用指数移动平均计算时，当油液颗粒浓度的短期趋势阈值为EMA + 3σ，启动预警；Step 7-1: Use Z-score to calculate and set the fault warning threshold from historical data: ,When using the exponential moving average calculation, when the short-term trend threshold of oil particle concentration is EMA + 3σ, the warning is activated; 步骤7-2：根据故障概率生成维护建议，高概率立即停机检查，低概率定期维护：。Step 7-2: Generate maintenance recommendations based on the probability of failure. If the probability is high, the machine should be shut down for inspection immediately; if the probability is low, regular maintenance should be performed. .
CN202510992266.7A 2025-07-18 2025-07-18 A fault prediction method for oil and gas pressure system in hydropower station by integrating perception data and knowledge graph Pending CN120822124A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202510992266.7A CN120822124A (en) 2025-07-18 2025-07-18 A fault prediction method for oil and gas pressure system in hydropower station by integrating perception data and knowledge graph

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202510992266.7A CN120822124A (en) 2025-07-18 2025-07-18 A fault prediction method for oil and gas pressure system in hydropower station by integrating perception data and knowledge graph

Publications (1)

Publication Number Publication Date
CN120822124A true CN120822124A (en) 2025-10-21

Family

ID=97372856

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202510992266.7A Pending CN120822124A (en) 2025-07-18 2025-07-18 A fault prediction method for oil and gas pressure system in hydropower station by integrating perception data and knowledge graph

Country Status (1)

Country Link
CN (1) CN120822124A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN120974290A (en) * 2025-10-23 2025-11-18 浪潮鸿启(山东)数字科技有限公司 A Smart Noise Detection Method for Industrial Sensors
CN121052845A (en) * 2025-11-03 2025-12-02 南通兴发玻璃有限公司 Knowledge Graph-Based Glass Product Quality Traceability and Prediction System

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN120974290A (en) * 2025-10-23 2025-11-18 浪潮鸿启(山东)数字科技有限公司 A Smart Noise Detection Method for Industrial Sensors
CN121052845A (en) * 2025-11-03 2025-12-02 南通兴发玻璃有限公司 Knowledge Graph-Based Glass Product Quality Traceability and Prediction System

Similar Documents

Publication Publication Date Title
CN117542169B (en) Automatic equipment temperature abnormality early warning method based on big data analysis
US20250238657A1 (en) Oil pipeline leak detection
CN120822124A (en) A fault prediction method for oil and gas pressure system in hydropower station by integrating perception data and knowledge graph
CN120163069B (en) Electromechanical system fault diagnosis system based on deep learning
CN109033930B (en) Equipment fault diagnosis method based on fault mechanism and statistical model online learning
CN111596604A (en) An intelligent fault diagnosis and self-healing control system and method for engineering equipment based on digital twin
CN109947088A (en) Equipment fault early-warning system based on model lifecycle management
CN118917834A (en) Power equipment fault early warning system
CN118898205B (en) Equipment fault detection method and system based on digital twinning
CN120195983A (en) Pump station operating condition monitoring method, system and storage medium based on digital twin
CN118332442A (en) Deep learning-based wind turbine generator fault diagnosis method and system
CN120408301A (en) A method for intelligent prediction of turbine status
CN119850195B (en) Operation and maintenance management method and system based on digital twin
CN118410279A (en) Deep learning-based turbine unit fault processing method and system
CN118153948B (en) Hydrogen energy storage equipment leakage fault warning method and device based on deep learning
CN119599636A (en) Thermal power plant equipment management method and system based on data fusion
CN116204825A (en) A data-driven fault detection method for production line equipment
CN118517367A (en) Self-diagnosis method of turbine guide vane fault based on deep learning and graph neural network
CN118152981A (en) Heat exchange station fault diagnosis method and system based on multi-mode data fusion
CN119377840A (en) A real-time detection method and device for abnormal pressure data of heat network pipeline
CN119903372A (en) An electrical equipment fault analysis system based on multi-source data fusion
CN119416113A (en) A method and system for monitoring energy equipment health based on big data
CN120100646A (en) A wind turbine monitoring and early warning system based on IoT sensing technology
CN119005753A (en) Hydropower station flow jump alarm threshold prediction method and system based on deep learning self-adaption
KR20240003469A (en) Predictive diagnosis method and system of nuclear power plant equipment

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination