[go: up one dir, main page]

CN119226985B - Fault detection method and system based on large model - Google Patents

Fault detection method and system based on large model Download PDF

Info

Publication number
CN119226985B
CN119226985B CN202411732254.2A CN202411732254A CN119226985B CN 119226985 B CN119226985 B CN 119226985B CN 202411732254 A CN202411732254 A CN 202411732254A CN 119226985 B CN119226985 B CN 119226985B
Authority
CN
China
Prior art keywords
fault
model
fault detection
time
data
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.)
Active
Application number
CN202411732254.2A
Other languages
Chinese (zh)
Other versions
CN119226985A (en
Inventor
黄礼成
张蓉
刘杰
姚飞
张伟
孙道广
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.)
Nanjing Halu Information Technology Co ltd
Original Assignee
Nanjing Halu Information Technology 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 Nanjing Halu Information Technology Co ltd filed Critical Nanjing Halu Information Technology Co ltd
Priority to CN202411732254.2A priority Critical patent/CN119226985B/en
Publication of CN119226985A publication Critical patent/CN119226985A/en
Application granted granted Critical
Publication of CN119226985B publication Critical patent/CN119226985B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
    • G06F16/367—Ontology
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/10—Pre-processing; Data cleansing
    • G06F18/15—Statistical pre-processing, e.g. techniques for normalisation or restoring missing data
    • 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
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • 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
    • G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
    • G06F40/295—Named entity recognition
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis
    • 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
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/092—Reinforcement 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
    • G06N3/08—Learning methods
    • G06N3/0985—Hyperparameter optimisation; Meta-learning; Learning-to-learn
    • 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
    • G06N5/025—Extracting rules from data
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computational Linguistics (AREA)
  • Computing Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Databases & Information Systems (AREA)
  • Animal Behavior & Ethology (AREA)
  • Probability & Statistics with Applications (AREA)
  • Test And Diagnosis Of Digital Computers (AREA)

Abstract

The invention discloses a fault detection method and system based on a large model, which belongs to the technical field of fault detection and specifically comprises the steps of collecting multi-source fault data from an industrial system, preprocessing, extracting fault characteristics to form a data set, loading a pre-trained large language model, carrying out fine adjustment according to the fault characteristic data set, constructing a fault detection model by utilizing the fine-adjusted model and the fault characteristic data set, carrying out training and evaluation by combining an adaptive optimization strategy to obtain a trained model, inputting the fault data collected in real time into the model, outputting a preliminary fault detection result, further obtaining fault type, position and probability information through a multi-level fault analysis method, and feeding back the result to optimize the model.

Description

Fault detection method and system based on large model
Technical Field
The invention belongs to the technical field of fault detection, and particularly relates to a fault detection method and system based on a large model.
Background
The fault detection method is used for judging whether the system has abnormality or fault by collecting and analyzing various data in the running process of the system and utilizing a specific algorithm or model, and aims to discover and locate the fault in time so as to take corresponding measures for repairing, thereby avoiding or reducing the influence and loss of the fault on the system. Traditional fault detection methods, such as rule-based methods, statistical-based methods, machine learning-based methods and the like, can meet the fault detection requirements of specific fields to a certain extent, but often have the problems of insufficient generalization capability, higher model complexity, limited processing capability for large-scale data and the like.
The invention discloses a fault detection method, a system and electronic equipment, which comprise the steps of obtaining target data corresponding to target equipment, wherein the target equipment is equipment to be detected, the target data comprises a plurality of sub-data, each sub-data is data for fault identification in different data sources corresponding to the target equipment, the target data is input into a fault analysis model, so that the fault analysis model obtains a target fault detection result according to the target data, the target fault detection result comprises a fault type and a fault reason of the target equipment, and the fault analysis model is a model trained based on historical fault data. By using the method of the large model, the labor input of fault detection is effectively reduced, the target data which corresponds to the target equipment and comprises a plurality of sub-data for fault identification is analyzed through the fault analysis model, and compared with the method of identifying faults only from a single data bus or diagnostic data, the accuracy and the comprehensiveness of the fault detection result are improved.
The Chinese patent with the authority bulletin number of CN115629930B discloses a fault detection method, a device, equipment and a storage medium based on a DSP system, which comprises the steps of obtaining a fault signal, inputting the fault signal into a preset fault detection model, classifying the fault signal based on the fault detection model, and diagnosing the classified fault signal to obtain fault information, wherein the fault detection model is obtained by performing iterative training on a model to be trained based on a fault signal sample, class weight information of the fault signal sample and a fault information label of the fault signal sample. According to the technical scheme, based on the pre-trained fault detection model, fault signals generated in the signal processing system of the DSP processor are classified and fault information is diagnosed, the fault information is rapidly and accurately output, a developer or a user is not required to manually debug a large number of problem positioning processes, and the fault detection efficiency of the DSP system is improved.
The above prior art has the following problems that in CN118069400a, although inputting target data into a fault analysis model to obtain a fault detection result is mentioned, real-time requirements and feedback mechanisms are not mentioned, only fault types and fault reasons are mentioned, but fault positions and fault probability information are not mentioned, in CN115629930B, fault signals inside a DSP system are mainly focused and classified and diagnosed based on the signals, the details and accuracy of the fault detection result are limited, and although the fault information is mentioned to be output quickly and accurately, real-time requirements and feedback mechanisms may not be discussed in detail.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a fault detection method and a fault detection system based on a large model, which are used for acquiring multi-source fault data from an industrial system, preprocessing the multi-source fault data, extracting fault characteristics to form a data set, loading a pre-trained large language model, carrying out fine adjustment according to the fault characteristic data set, constructing a fault detection model by utilizing the fine-adjusted model and the fault characteristic data set, carrying out training and evaluation by combining an adaptive optimization strategy to obtain a trained model, inputting the fault data acquired in real time into the model, outputting a preliminary fault detection result, further obtaining fault type, position and probability information by a multi-level fault analysis method, and feeding back the result to optimize the model.
In order to achieve the above purpose, the present invention provides the following technical solutions:
The fault detection method based on the large model comprises the following steps:
Step S1, multi-source fault data are collected from an industrial system and preprocessed, and meanwhile, fault feature extraction is carried out on the preprocessed multi-source fault data to obtain a fault feature data set;
s2, loading a pre-trained large language model, and performing fine adjustment according to a fault characteristic data set;
S3, constructing a fault detection model by utilizing the trimmed large language model and the fault characteristic data set, training and evaluating the fault detection model by combining with a self-adaptive optimization strategy, and optimizing the fault detection model according to an evaluation result to obtain a trained fault detection model;
S4, inputting the fault data acquired in real time into a trained fault detection model, outputting a preliminary fault detection result by the fault detection model according to the input data, carrying out multi-level analysis on the preliminary fault detection result by combining a multi-level fault analysis method to obtain fault type, fault position and fault probability information, and feeding back the real-time fault detection result and the multi-level analysis result to the fault detection model for optimization;
the specific steps of the step S4 include:
s4.1, acquiring real-time fault data, and preprocessing to obtain preprocessed real-time fault data , wherein,Representing the M-th preprocessed real-time fault data, M representing the number of preprocessed real-time fault data;
S4.2 will Inputting the trained fault detection model, and outputting a preliminary fault detection result by the fault detection model according to the input data;
S4.3, carrying out deep analysis on a preliminary fault detection result by combining a multi-level fault analysis method, and judging fault type, fault position and fault probability information by combining characteristic information and historical fault cases in real-time fault data through a knowledge graph and an expert system;
S4.4, feeding back a real-time fault detection result and a multi-level analysis result to a fault detection model, and carrying out online updating according to the feedback result by utilizing a reinforcement learning algorithm;
the specific steps of the S4.3 comprise:
S4.31 acquisition of And a preliminary fault detection result;
S4.32, extracting fault knowledge from historical fault cases, expert experiences and equipment manuals, and representing the extracted fault knowledge into a graphical knowledge structure by utilizing a knowledge graph method, wherein fault type, fault position and fault probability information are represented as nodes, and the relationship between the fault type, the fault position and the fault probability information is represented as edges;
S4.33, inputting the preliminary fault detection result into a preloaded expert system, and matching and reasoning the fault detection result by the expert system according to the fault knowledge in the knowledge graph to obtain fault type, fault position and fault probability information;
s4.34, combining according to the reasoning result of the expert system Judging the specific type and position of the fault;
S4.35, analyzing the historical fault cases by using a machine learning algorithm to obtain probability distribution of fault occurrence, and evaluating the probability of fault occurrence according to the running state of the current equipment and the specific type and position of the fault;
The specific step of S4.4 includes:
S4.41, integrating the real-time fault detection result and the multi-level analysis result to generate a fault detection report;
S4.42, loading a pre-built reinforcement learning model, inputting a fault detection report as a feedback result into the reinforcement learning model, designing a reward function and a state transition rule, and training the reinforcement learning model by using a reinforcement learning strategy to enable the reinforcement learning model to learn and adapt to a new fault mode, thereby obtaining a trained reinforcement learning model, wherein the formula is as follows:
;
Wherein, At time t+1, the quality function in state z, action d, fault signature f and context information g,At time t, the quality function in state z, action d, fault signature f and context information g,The learning rate is indicated as being indicative of the learning rate,Indicating a prize to be earned immediately after action d is taken,Representing the discount factor(s),Representing the maximum value of the quality function at the next state, action, fault signature and context information at time t,、、、Representing the next state, action, fault signature and context information respectively,The characteristic weight is represented by a characteristic weight,The characteristic function is represented by a function of the feature,The constraint weight is represented as a function of the constraint weight,Representing a constraint function;
S4.43, adjusting parameters of the fault detection model according to the training result of the reinforcement learning model, evaluating the accuracy of the adjusted fault detection model, and simultaneously, further optimizing and adjusting the fault detection model according to the evaluation result;
if the evaluation result is not ideal, returning to the step S4.42, retraining the reinforcement learning model, and updating the fault detection model again;
if the evaluation result meets the requirement, outputting the result and ending the flow.
Specifically, the specific formula of the adaptive optimization strategy in step S3 is as follows:
;
Wherein t represents the current time of day, AndRepresenting the first moment estimate and the second moment estimate of the fault detection model at time t respectively,Representing the fault detection model parameters at time t +1,Representing the fault detection model parameters at time t,The step size of the control parameter update is indicated,AndThe exponential decay rate of moment estimation at time t is shown, and b is a constant.
The fault detection system based on the large model comprises a data processing module, a model fine-tuning module, a fault detection module and a fault analysis module;
the data processing module is used for collecting multi-source fault data from the industrial system, preprocessing and extracting features;
The model fine tuning module is used for fine tuning the pre-trained large language model by using the fault characteristic data set;
The fault detection module is used for constructing a fault detection model by utilizing the trimmed large language model and the fault characteristic data set, training and evaluating the fault detection model, and optimizing the model according to the evaluation result;
the fault analysis module is used for inputting the fault data acquired in real time into a trained fault detection model, outputting a preliminary fault detection result and carrying out multi-level analysis by combining a multi-level fault analysis method.
The fault analysis module comprises a real-time detection unit, a multi-level analysis unit and a feedback optimization unit;
the real-time detection unit is used for inputting fault data acquired in real time into a trained fault detection model and outputting a preliminary fault detection result;
the multi-level analysis unit is used for carrying out multi-level analysis on the preliminary fault detection result;
And the feedback optimization unit is used for feeding back the real-time detection result and the multi-level analysis result to the fault detection model for optimization.
An electronic device comprising a memory storing a computer program and a processor implementing the steps of a large model based fault detection method when the computer program is executed.
A computer readable storage medium having stored thereon computer instructions which, when executed, perform the steps of a large model based fault detection method.
Compared with the prior art, the invention has the beneficial effects that:
1. the invention provides a fault detection system based on a large model, which is optimized and improved in terms of architecture, operation steps and flow, and has the advantages of simple flow, low investment and operation cost and low production and working costs.
2. The invention provides a fault detection method based on a large model, which is characterized in that multi-source fault data are collected from an industrial system, preprocessing and feature extraction are carried out, and the pre-trained large language model is combined for fine adjustment, so that a fault detection model is constructed and optimized.
3. The fault detection method based on the large model is provided, fault data acquired in real time are input into a trained fault detection model, a preliminary fault detection result can be rapidly output, detailed fault type, position and probability information can be obtained through further analysis of the result by a multi-level fault analysis method, powerful support is provided for rapid positioning and repairing of faults, meanwhile, the real-time detection result and the analysis result are fed back to the model for optimization, the detection capability and accuracy of the model can be continuously improved, and the stability of an industrial system is improved.
Drawings
FIG. 1 is a schematic diagram of a large model-based fault detection method of the present invention;
FIG. 2 is a schematic flow chart of a fault detection method based on a large model in the invention;
FIG. 3 is a flow chart of the implementation of real-time fault detection results in the fault detection method based on the large model of the present invention;
FIG. 4 is a schematic diagram of a large model-based fault detection system according to the present invention.
Detailed Description
Example 1
Referring to fig. 1-3, the fault detection method based on the large model according to the embodiment of the invention includes the following steps:
Step S1, multi-source fault data are collected from an industrial system and preprocessed, and meanwhile, fault feature extraction is carried out on the preprocessed multi-source fault data to obtain a fault feature data set;
further, the specific steps of step S1 include:
S1.1, determining the type of fault data to be acquired, such as vibration signals, temperature signals and pressure signals, and installing sensors at each key position of an industrial system to acquire the fault data in real time, and simultaneously, ensuring the accuracy and reliability of the sensors and the real-time property and continuity of data acquisition;
S1.2, cleaning the collected original data, removing noise, abnormal values and repeated data, formatting the data, and converting the data into a data format suitable for processing by a data mining algorithm;
S1.3, mining the preprocessed multi-source fault data by using a decision tree classification algorithm, and extracting feature information related to faults, such as fault frequency, fault amplitude and fault type, wherein the decision tree classification algorithm is the prior art content in the field and is not an inventive scheme of the application, and is not repeated herein;
s1.4, integrating the extracted fault characteristic information into a fault characteristic data set for subsequent fault detection and diagnosis.
S2, loading a pre-trained large language model, and performing fine adjustment according to a fault characteristic data set;
further, the specific steps of step S2 include:
S2.1, selecting a pre-training model, namely selecting a model suitable for the current task from the existing pre-training large language models, wherein the models are usually trained on large-scale text data and have certain language understanding and generating capabilities, and selecting ERNIE a large model according to the invention, wherein the ERNIE large model is a hundred-degree developed pre-training model, and the representation capability of the model is enhanced by introducing entity and semantic information.
S2.2, loading a pre-trained large language model, namely loading the selected pre-trained large language model into a memory or a computing device so as to carry out a subsequent fine tuning task;
s2.3, preparing a fault characteristic data set;
and S2.4, fine tuning the model, namely fine tuning the pre-trained large language model by using a fault characteristic data set to enable the pre-trained large language model to be better suitable for the current fault detection task, wherein in the fine tuning process, super parameters such as parameters, learning rate and the like of the large language model can be adjusted to optimize the performance of the large language model on the fault detection task, and a large language model parameter optimization formula is as follows:
;
Wherein, The loss function is represented by a function of the loss,Representing large language models at given input featuresSum parametersDown predictive labelIs a function of the probability of (1),The tag representing the ith fault signature data, i.e. the fault type or category,Indicating the i-th fault signature data,Representing large language model parameters, N represents the number of samples in the fault signature dataset.
It should be noted that the goal of the fine tuning process is to adjust the parametersTo minimize the loss functionThereby improving the accuracy of the large language model in fault detection tasks.
S3, constructing a fault detection model by utilizing the trimmed large language model and the fault characteristic data set, training and evaluating the fault detection model by combining with a self-adaptive optimization strategy, and optimizing the fault detection model according to an evaluation result to obtain a trained fault detection model;
Further, the specific steps of step S3 include:
S3.1, constructing a fault detection model frame and integrating a self-adaptive optimization strategy;
S3.2, dividing a training set and a testing set according to the ratio of 7:3 by using the fault characteristic data set, training a fault detection model by using the training set in the fault characteristic data set, calculating loss through forward propagation, and updating fault detection model parameters through backward propagation;
S3.3, judging whether the performance of the fault detection model meets the requirement according to the evaluation result on the verification set, such as accuracy, and if the performance of the fault detection model is poor, performing fault detection model optimization, including adjusting model architecture, adding training data, adjusting optimization algorithm parameters, and repeating the processes of training, evaluating and optimizing until the performance of the fault detection model reaches a preset level;
And S3.4, when the fault detection model shows good performance on the verification set, the fault detection model is considered to be trained, and parameters and structures of the fault detection model are saved for subsequent deployment and use in an actual environment.
And S4, inputting the fault data acquired in real time into a trained fault detection model, outputting a preliminary fault detection result by the fault detection model according to the input data, carrying out multi-level analysis on the preliminary fault detection result by combining a multi-level fault analysis method to obtain fault type, fault position and fault probability information, and feeding back the real-time fault detection result and the multi-level analysis result to the fault detection model for optimization.
The specific formula of the self-adaptive optimization strategy in the S3 is as follows:
;
Wherein t represents the current time of day, AndRepresenting the first moment estimate and the second moment estimate of the fault detection model at time t respectively,Representing the fault detection model parameters at time t +1,Representing the fault detection model parameters at time t,The step size of the control parameter update is indicated,AndThe exponential decay rate of moment estimation at time t is shown, and b is a constant.
The specific steps of S4 include:
s4.1, acquiring real-time fault data, and preprocessing to obtain preprocessed real-time fault data , wherein,Representing the M-th preprocessed real-time fault data, M representing the number of preprocessed real-time fault data;
S4.2 will Inputting the trained fault detection model, and outputting a preliminary fault detection result by the fault detection model according to the input data;
Further, the specific step of S4.2 includes:
S4.21 acquisition of Loading a trained fault detection model, ensuring that a model file is complete and undamaged, and optimizing model parameters to an optimal state;
S4.22 using normalization method Converting the format into an input format of a fault detection model, and converting the converted input format into a fault detection modelInputting the fault detection model;
the specific process of data conversion comprises the following steps:
(1) Normalizing or normalizing the continuous numerical value characteristics to ensure that all fault characteristic data are on the same scale and eliminate dimension influence;
(2) Filling the missing values with an average, median or mode;
(3) Data type conversion is accomplished, illustratively, converting an integer type to a floating point type.
S4.23 conversion of the fault detection model to the inputPerforming calculation, extracting features and applying learned rules to identify potential faults;
and S4.24, outputting a preliminary fault detection result, wherein the fault detection result is usually presented in a numerical value, a label or a text form, and depends on the design and the requirement of the system.
S4.3, carrying out deep analysis on a preliminary fault detection result by combining a multi-level fault analysis method, and judging fault type, fault position and fault probability information by combining characteristic information and historical fault cases in real-time fault data through a knowledge graph and an expert system;
And S4.4, feeding back a real-time fault detection result and a multi-level analysis result to the fault detection model, and carrying out online updating according to the feedback result by utilizing a reinforcement learning algorithm.
The specific steps of S4.3 include:
S4.31 acquisition of And a preliminary fault detection result;
S4.32, extracting fault knowledge from historical fault cases, expert experiences and equipment manuals, and representing the extracted fault knowledge into a graphical knowledge structure by utilizing a knowledge graph method, wherein the fault type, the fault position and the fault probability information are represented as nodes, the relationship among the nodes is represented as edges, and meanwhile, the knowledge graph method is the prior art content in the field and is not an inventive scheme of the application and is not repeated herein;
S4.33, inputting the preliminary fault detection result into a preloaded expert system, and matching and reasoning the fault detection result by the expert system according to the fault knowledge in the knowledge graph to obtain fault type, fault position and fault probability information;
further, the specific step of S4.33 includes:
(1) Obtaining preliminary fault detection results, wherein the results are usually represented as a series of fault signals, and preprocessing the collected fault signals, including filtering, denoising, normalization and other operations, so as to improve the accuracy and reliability of the signals;
(2) Inputting the preprocessed fault signals into a preloaded expert system, wherein the expert system is a computer program based on an artificial intelligence technology and comprises a large amount of fault knowledge and reasoning rules;
(3) The expert system matches the input fault signals by utilizing fault knowledge in a knowledge graph, wherein the knowledge graph is a graph structure comprising a large number of entities, relations and attributes, the entities can be equipment, parts or fault types, the relations can be connection relations, causal relations and the like, and the attributes can be the model number and parameters of the equipment;
(4) According to the matching result, the expert system performs fault reasoning by using a built-in reasoning mechanism, the reasoning process adopts a case-based reasoning method which is the prior art in the field and is not an inventive scheme of the application, and details are not repeated here;
(5) After reasoning analysis, the expert system outputs fault type, fault position and fault probability information which can help a user to quickly locate and solve faults.
S4.34, combining according to the reasoning result of the expert systemJudging the specific type and position of the fault;
And S4.35, analyzing the historical fault cases by using a machine learning algorithm to obtain probability distribution of fault occurrence, and evaluating the probability of fault occurrence according to the running state of the current equipment and the specific type and position of the fault.
Further, the specific step of S4.35 includes:
(1) Acquisition of And historical fault case data, and preprocessing to generate original data;
(2) Extracting features of the original data by adopting a statistical analysis method to obtain fault features;
(3) Loading a pre-constructed machine learning model based on a decision tree, and training and evaluating the machine learning model based on the decision tree by using the obtained fault characteristics, wherein the training and evaluating of the machine learning model are the prior art content in the field and are not the inventive scheme of the application, and are not repeated herein;
(4) The real-time fault data is input into a trained machine learning model based on a decision tree, the machine learning model based on the decision tree outputs probability distribution of fault occurrence, and the probability of the fault occurrence of the current equipment can be estimated according to the probability distribution.
The specific steps of S4.4 include:
S4.41, integrating the real-time fault detection result and the multi-level analysis result to generate a fault detection report;
S4.42, loading a pre-built reinforcement learning model, inputting a fault detection report as a feedback result into the reinforcement learning model, designing a reward function and a state transition rule, and training the reinforcement learning model by using a reinforcement learning strategy to enable the reinforcement learning model to learn and adapt to a new fault mode, thereby obtaining a trained reinforcement learning model, wherein the formula is as follows:
;
Wherein, At time t+1, the quality function in state z, action d, fault signature f and context information g,At time t, the quality function in state z, action d, fault signature f and context information g,The learning rate is indicated as being indicative of the learning rate,Indicating a prize to be earned immediately after action d is taken,Representing the discount factor(s),Representing the maximum value of the quality function at the next state, action, fault signature and context information at time t,、、、Representing the next state, action, fault signature and context information respectively,The characteristic weight is represented by a characteristic weight,Representing a feature function, for representing the importance of the fault feature,The constraint weight is represented as a function of the constraint weight,Representing constraint functions for representing resource constraints and security requirements;
In the present invention, a characteristic function is introduced For emphasizing the importance of certain specific features. For example, in fault detection, certain fault characteristics may be more important than others, possibly byTo reflect this, feature weightsControlling the influence of the characteristic function. Introducing constraint functionsFor ensuring that certain constraints are met when taking actions, which may be constrained by resource limitations or security requirements, in an industrial environment, for example, may be achievedTo represent these constraints, constraint weightsControlling the influence of the constraint function.
S4.43, adjusting parameters of the fault detection model according to a training result of the reinforcement learning model, and performing accuracy assessment on the adjusted fault detection model, and simultaneously, performing further optimization and adjustment on the fault detection model according to an assessment result, wherein the accuracy assessment is the prior art content in the field and is not an inventive scheme of the application, and is not repeated herein;
if the evaluation result is not ideal, returning to the step S4.42, retraining the reinforcement learning model, and updating the fault detection model again;
if the evaluation result meets the requirement, outputting the result and ending the flow.
Example 2
Referring to FIG. 4, another embodiment of the present invention provides a large model-based fault detection system, comprising:
the system comprises a data processing module, a model fine adjustment module, a fault detection module and a fault analysis module;
The data processing module is used for collecting multi-source fault data from an industrial system, preprocessing and extracting features, wherein the data processing module comprises the steps of collecting the fault data from each sensor and a log file source, and preprocessing comprises cleaning, formatting and normalizing so as to improve the data quality;
the model fine tuning module is used for fine tuning the pre-trained large language model by using the fault characteristic data set so as to be more suitable for a fault detection task;
The fault detection module is used for constructing a fault detection model by utilizing the trimmed large language model and the fault characteristic data set, training and evaluating the fault detection model, and optimizing the model according to the evaluation result;
The fault analysis module is used for inputting the fault data acquired in real time into the trained fault detection model, outputting a preliminary fault detection result and carrying out multi-level analysis by combining a multi-level fault analysis method.
The fault detection module comprises a model construction unit, an evaluation unit and a model optimization unit;
the model construction unit is used for constructing a fault detection model by combining the trimmed large language model and the fault characteristic data set to form a preliminary model capable of detecting faults;
the training and evaluating unit is used for training the fault detection model and evaluating the performance, such as accuracy and recall, of the fault detection model so as to know the performance of the fault detection model on a fault detection task;
and the model optimization unit is used for optimizing the fault detection model according to the evaluation result, such as adjusting the model structure and parameters, and improving the detection performance and generalization capability of the fault detection model.
The fault analysis module comprises a real-time detection unit, a multi-level analysis unit and a feedback optimization unit;
The real-time detection unit is used for inputting the fault data acquired in real time into a trained fault detection model, outputting a preliminary fault detection result and realizing real-time detection of faults;
the multi-level analysis unit is used for carrying out multi-level analysis on the preliminary fault detection result, such as analysis based on fault type, fault position and fault probability, so as to provide more detailed and accurate fault information and facilitate rapid positioning and solving of faults;
And the feedback optimization unit is used for feeding back the real-time detection result and the multi-level analysis result to the fault detection model for optimization, realizing continuous learning and optimization of the model, and improving the adaptability and accuracy of the model.
Example 3
An electronic device includes a memory and a processor, where the memory stores a computer program, and the processor executes the computer program to implement steps of a large model-based fault detection method, and the specific embodiments of the method may be referred to above and will not be described herein.
A computer readable storage medium having stored thereon computer instructions which when executed perform the steps of a large model based fault detection method, wherein the storage medium may be a volatile or non-volatile computer readable storage medium.
The embodiments of the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to the above-described embodiments, which are merely illustrative and not restrictive, and variations, modifications, substitutions and alterations can be made to the above-described embodiments by those having ordinary skill in the art without departing from the spirit and scope of the present invention, and these are all within the protection of the present invention.

Claims (6)

1.基于大模型的故障检测方法，其特征在于，包括：1. A fault detection method based on a large model, characterized by comprising: 步骤S1：从工业系统中采集多源故障数据，并进行预处理，同时，对预处理后的多源故障数据进行故障特征提取，获得故障特征数据集；Step S1: Collect multi-source fault data from the industrial system and perform preprocessing. At the same time, extract fault features from the preprocessed multi-source fault data to obtain a fault feature data set; 步骤S2：加载预训练好的大语言模型，并根据故障特征数据集进行微调；Step S2: Load the pre-trained large language model and fine-tune it according to the fault feature dataset; 步骤S3：利用微调后的大语言模型和故障特征数据集构建故障检测模型，结合自适应优化策略，对故障检测模型进行训练和评估，根据评估结果，对故障检测模型进行优化，获得训练好的故障检测模型；Step S3: construct a fault detection model using the fine-tuned large language model and the fault feature data set, train and evaluate the fault detection model in combination with the adaptive optimization strategy, and optimize the fault detection model according to the evaluation results to obtain a trained fault detection model; 步骤S4：将实时采集的故障数据输入训练好的故障检测模型中，故障检测模型根据输入数据输出初步的故障检测结果，结合多层次故障分析方法，对初步的故障检测结果进行多层次分析，获得故障类型、故障位置和故障概率信息，并将实时的故障检测结果和多层次分析的结果反馈至故障检测模型进行优化；Step S4: input the real-time collected fault data into the trained fault detection model, and the fault detection model outputs preliminary fault detection results according to the input data. Combined with the multi-level fault analysis method, the preliminary fault detection results are subjected to multi-level analysis to obtain the fault type, fault location and fault probability information, and the real-time fault detection results and the results of the multi-level analysis are fed back to the fault detection model for optimization; 所述步骤S4的具体步骤包括：The specific steps of step S4 include: S4.1：获取实时故障数据，并进行预处理，获得预处理后的实时故障数据，其中，表示第M个预处理后的实时故障数据，M表示预处理后的实时故障数据的数量；S4.1: Obtain real-time fault data and perform preprocessing to obtain preprocessed real-time fault data ,in, represents the Mth preprocessed real-time fault data, and M represents the number of preprocessed real-time fault data; S4.2：将输入训练好的故障检测模型中，故障检测模型根据输入数据输出初步的故障检测结果；S4.2: The trained fault detection model is input, and the fault detection model outputs preliminary fault detection results based on the input data; S4.3：结合多层次故障分析方法，对初步的故障检测结果进行深入分析，并通过知识图谱和专家系统，结合实时故障数据中的特征信息和历史故障案例，判断故障类型、故障位置和故障概率信息；S4.3: Combine multi-level fault analysis methods to conduct in-depth analysis of preliminary fault detection results, and use knowledge graphs and expert systems to combine feature information in real-time fault data and historical fault cases to determine fault type, fault location, and fault probability information; S4.4：将实时故障检测结果和多层次分析结果反馈至故障检测模型，利用强化学习算法，根据反馈结果进行在线更新；S4.4: Feedback the real-time fault detection results and multi-level analysis results to the fault detection model, and use the reinforcement learning algorithm to perform online updates based on the feedback results; 所述S4.3的具体步骤包括：The specific steps of S4.3 include: S4.31：获取和初步的故障检测结果；S4.31: Acquisition and preliminary fault detection results; S4.32：从历史故障案例、专家经验、设备手册中提取故障知识，并利用知识图谱方法，将提取的故障知识表示为图形化的知识结构，其中，故障类型、故障位置、故障概率信息表示为节点，它们之间的关系被表示为边；S4.32: Extract fault knowledge from historical fault cases, expert experience, and equipment manuals, and use the knowledge graph method to represent the extracted fault knowledge as a graphical knowledge structure, where the fault type, fault location, and fault probability information are represented as nodes, and the relationship between them is represented as edges; S4.33：将初步的故障检测结果输入到预加载的专家系统中，专家系统根据知识图谱中的故障知识，对故障检测结果进行匹配和推理，获得故障类型、故障位置和故障概率信息；S4.33: Input the preliminary fault detection results into the preloaded expert system. The expert system matches and infers the fault detection results according to the fault knowledge in the knowledge graph to obtain the fault type, fault location and fault probability information; S4.34：根据专家系统的推理结果，结合，判断故障的具体类型和位置；S4.34: Based on the reasoning results of the expert system, combined with , determine the specific type and location of the fault; S4.35：利用机器学习算法，对历史故障案例进行分析，得到故障发生的概率分布，并根据当前设备的运行状态、故障的具体类型和位置，评估故障发生的概率；S4.35: Use machine learning algorithms to analyze historical failure cases to obtain the probability distribution of failures, and evaluate the probability of failures based on the current operating status of the equipment, the specific type and location of the failure; 所述S4.4的具体步骤包括：The specific steps of S4.4 include: S4.41：将实时故障检测结果和多层次分析结果进行整合，生成故障检测报告；S4.41: Integrate the real-time fault detection results and multi-level analysis results to generate a fault detection report; S4.42：加载预构建好的强化学习模型，将故障检测报告作为反馈结果，输入到强化学习模型，并设计奖励函数和状态转移规则，利用强化学习策略对强化学习模型进行训练，使强化学习模型能够学习并适应新的故障模式，获得训练后的强化学习模型，公式为：S4.42: Load the pre-built reinforcement learning model, use the fault detection report as the feedback result, input it into the reinforcement learning model, design the reward function and state transition rules, use the reinforcement learning strategy to train the reinforcement learning model, so that the reinforcement learning model can learn and adapt to the new fault mode, and obtain the trained reinforcement learning model. The formula is: ； ; 其中，表示t+1时刻时，在状态z、动作d、故障特征f和上下文信息g下的质量函数，表示t时刻时，在状态z、动作d、故障特征f和上下文信息g下的质量函数，表示学习率，表示采取动作d后立即获得的奖励，表示折扣因子，表示t时刻时，在下一个状态、动作、故障特征和上下文信息下的质量函数的最大值，、、、分别表示下一个状态、动作、故障特征和上下文信息，表示特征权重，表示特征函数，表示约束权重，表示约束函数；in, represents the quality function under state z, action d, fault feature f and context information g at time t+1, represents the quality function under state z, action d, fault feature f and context information g at time t, represents the learning rate, represents the reward obtained immediately after taking action d, represents the discount factor, represents the maximum value of the quality function under the next state, action, fault characteristics and context information at time t, , , , Respectively represent the next state, action, fault characteristics and context information, represents the feature weight, represents the characteristic function, represents the constraint weight, represents the constraint function; S4.43：根据强化学习模型的训练结果，调整故障检测模型的参数，并对调整后的故障检测模型进行准确性评估，同时，根据评估结果，对故障检测模型进行进一步的优化和调整；S4.43: According to the training results of the reinforcement learning model, adjust the parameters of the fault detection model, and evaluate the accuracy of the adjusted fault detection model. At the same time, according to the evaluation results, further optimize and adjust the fault detection model; 若评估结果不理想，则返回步骤S4.42，重新训练强化学习模型，并再次更新故障检测模型；If the evaluation result is not satisfactory, return to step S4.42, retrain the reinforcement learning model, and update the fault detection model again; 若评估结果满足要求，则输出结果，结束流程。If the evaluation result meets the requirements, the result is output and the process ends. 2.如权利要求1所述的基于大模型的故障检测方法，其特征在于，所述步骤S3中自适应优化策略的具体公式为：2. The large model-based fault detection method according to claim 1, characterized in that the specific formula of the adaptive optimization strategy in step S3 is: ； ; 其中，t表示当前时刻，和分别表示故障检测模型在t时刻的一阶矩估计和二阶矩估计，表示在t+1时刻的故障检测模型参数，表示在t时刻的故障检测模型参数，表示控制参数更新的步长，和表示在t时刻矩估计的指数衰减率，b表示常数。Among them, t represents the current time, and They represent the first-order moment estimation and second-order moment estimation of the fault detection model at time t, respectively. represents the fault detection model parameters at time t+1, represents the fault detection model parameters at time t, represents the step size of the control parameter update, and represents the exponential decay rate of the moment estimate at time t, and b represents a constant. 3.基于大模型的故障检测系统，其用于实现权利要求1-2中任一项所述的基于大模型的故障检测方法，其特征在于，包括：数据处理模块、模型微调模块、故障检测模块、故障分析模块；3. A fault detection system based on a large model, which is used to implement the fault detection method based on a large model as described in any one of claims 1-2, characterized in that it comprises: a data processing module, a model fine-tuning module, a fault detection module, and a fault analysis module; 所述数据处理模块，用于从工业系统中采集多源故障数据，并进行预处理和特征提取；The data processing module is used to collect multi-source fault data from the industrial system and perform preprocessing and feature extraction; 所述模型微调模块，用于使用故障特征数据集对预训练好的大语言模型进行微调；The model fine-tuning module is used to fine-tune the pre-trained large language model using the fault feature dataset; 所述故障检测模块，用于利用微调后的大语言模型和故障特征数据集构建故障检测模型，并进行训练和评估，根据评估结果对模型进行优化；The fault detection module is used to construct a fault detection model using the fine-tuned large language model and the fault feature data set, and to perform training and evaluation, and optimize the model according to the evaluation results; 所述故障分析模块，用于将实时采集的故障数据输入训练好的故障检测模型中，输出初步的故障检测结果，并结合多层次故障分析方法进行多层次分析。The fault analysis module is used to input the real-time collected fault data into the trained fault detection model, output the preliminary fault detection results, and perform multi-level analysis in combination with the multi-level fault analysis method. 4.如权利要求3所述的基于大模型的故障检测系统，其特征在于，所述故障分析模块包括：实时检测单元、多层次分析单元、反馈优化单元；4. The large model-based fault detection system according to claim 3, characterized in that the fault analysis module comprises: a real-time detection unit, a multi-level analysis unit, and a feedback optimization unit; 所述实时检测单元，用于将实时采集的故障数据输入训练好的故障检测模型中，输出初步的故障检测结果；The real-time detection unit is used to input the real-time collected fault data into the trained fault detection model and output preliminary fault detection results; 所述多层次分析单元，用于对初步的故障检测结果进行多层次分析；The multi-level analysis unit is used to perform multi-level analysis on the preliminary fault detection results; 所述反馈优化单元，用于将实时检测结果和多层次分析的结果反馈至故障检测模型进行优化。The feedback optimization unit is used to feed back the real-time detection results and the results of the multi-level analysis to the fault detection model for optimization. 5.一种电子设备，包括存储器和处理器，所述存储器存储有计算机程序，其特征在于，所述处理器执行所述计算机程序时实现权利要求1-2任一项所述的基于大模型的故障检测方法的步骤。5. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein the processor implements the steps of the large model-based fault detection method according to any one of claims 1 to 2 when executing the computer program. 6.一种计算机可读存储介质，其特征在于，其上存储有计算机指令，当计算机指令运行时执行权利要求1-2任一项所述的基于大模型的故障检测方法的步骤。6. A computer-readable storage medium, characterized in that computer instructions are stored thereon, and when the computer instructions are executed, the steps of the large model-based fault detection method described in any one of claims 1-2 are executed.
CN202411732254.2A 2024-11-29 2024-11-29 Fault detection method and system based on large model Active CN119226985B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202411732254.2A CN119226985B (en) 2024-11-29 2024-11-29 Fault detection method and system based on large model

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202411732254.2A CN119226985B (en) 2024-11-29 2024-11-29 Fault detection method and system based on large model

Publications (2)

Publication Number Publication Date
CN119226985A CN119226985A (en) 2024-12-31
CN119226985B true CN119226985B (en) 2025-06-10

Family

ID=94048102

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202411732254.2A Active CN119226985B (en) 2024-11-29 2024-11-29 Fault detection method and system based on large model

Country Status (1)

Country Link
CN (1) CN119226985B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN119879335B (en) * 2025-03-26 2025-07-15 江西理工大学 Fault prediction system and method for heating and ventilation equipment
CN120373581A (en) * 2025-06-26 2025-07-25 深圳市今天国际智能机器人有限公司 Fault prediction method and device applied to AMR (automatic dependent memory) and electronic equipment

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118260603A (en) * 2024-02-07 2024-06-28 中兴系统技术有限公司 Equipment state analysis method, medium and equipment based on intelligent large model
CN118353667A (en) * 2024-04-22 2024-07-16 云仓库(广东)信息科技有限公司 Network security early warning method and system based on deep learning

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AU2017428008A1 (en) * 2017-08-18 2019-12-05 Landmark Graphics Corporation Hybrid optimization of fault detection and interpretation
CN114065613B (en) * 2021-10-27 2022-12-09 中国华能集团清洁能源技术研究院有限公司 Multi-working-condition process industrial fault detection and diagnosis method based on deep migration learning
CN118386024A (en) * 2024-04-17 2024-07-26 万可工业科技(山东)有限公司 Machine tool based on artificial intelligence and fault detection method thereof
CN118466869A (en) * 2024-04-29 2024-08-09 广州市森扬电子科技有限公司 Printer troubleshooting solutions, equipment and storage media based on log analysis
CN118586893A (en) * 2024-06-21 2024-09-03 江苏核电有限公司 A nuclear power plant equipment status intelligent analysis and fault early warning system
CN118748642B (en) * 2024-07-29 2025-07-04 上海凌泽信息科技有限公司 Automatic fault recovery method for Internet of Things communication system based on fault-tolerant design
CN118625791B (en) * 2024-08-08 2024-10-25 山东山矿机械有限公司 A fault-tolerant control system and method for abnormal conveying equipment

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118260603A (en) * 2024-02-07 2024-06-28 中兴系统技术有限公司 Equipment state analysis method, medium and equipment based on intelligent large model
CN118353667A (en) * 2024-04-22 2024-07-16 云仓库(广东)信息科技有限公司 Network security early warning method and system based on deep learning

Also Published As

Publication number Publication date
CN119226985A (en) 2024-12-31

Similar Documents

Publication Publication Date Title
JP7657743B2 (en) Anomaly detection system, anomaly detection method, anomaly detection program, and trained model generation method
US12119114B2 (en) Missing medical diagnosis data imputation method and apparatus, electronic device and medium
CN119226985B (en) Fault detection method and system based on large model
Lindemann et al. Anomaly detection and prediction in discrete manufacturing based on cooperative LSTM networks
CN111767930A (en) IoT time series data anomaly detection method and related equipment
CN116361059B (en) Diagnosis method and diagnosis system for abnormal root cause of banking business
CN116680639B (en) A deep learning-based anomaly detection method for deep-sea submersible sensor data
CN119168075B (en) A method and system for real-time processing and analysis of AI big data
CN117909881A (en) Fault diagnosis method and device for multi-source data fusion pumping unit
CN120011781A (en) A method, device and storage medium for predicting life span and fault monitoring of water conservancy equipment
CN118821835A (en) A data processing, analysis and decision-making method and system based on large model intelligent agent
CN118194156A (en) Knowledge and data fusion driving iteration type nuclear power station water pump fault diagnosis method
Patil et al. Next-Generation Bug Reporting: Enhancing Development with AI Automation
CN115712874A (en) Thermal energy power system fault diagnosis method and device based on time series characteristics
CN119473784A (en) A server cluster anomaly diagnosis method based on big data AI
Colace et al. Unsupervised learning techniques for vibration-based structural health monitoring systems driven by data: a general overview
CN117354168A (en) Hybrid ensemble method for predictive modeling of Internet of things
CN112016240B (en) Prediction method for residual stable service life of incomplete degradation equipment with similar evidence
Singh et al. Software quality analysis based on selective parameters using enhanced ensemble model
CN117113193B (en) Interpretable process monitoring method based on depth state space model and related equipment
CN118312911B (en) Diagnostic fault early warning system and method based on data mining
Srivastava et al. Feature clustering and ensemble learning based approach for software defect prediction
CN119989298A (en) A complex data processing method and system based on large model
CN118897745A (en) Real-time log analysis method, system, device and storage medium based on big data
CN120612007A (en) Enterprise operation method and system based on human efficiency analysis

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
GR01 Patent grant
GR01 Patent grant