Air compressor fault prediction method and system based on large model

Abstract

The invention discloses an air compressor fault prediction method based on a large model, and relates to electric digital data processing, which comprises the following steps: grading collected air compressor data, and sequentially carrying out anomaly analysis processing on each grade of data to obtain an interaction influence value, a matching degree and an association support value between each grade of data; performing standardization and textualization analysis on each level of data, the interaction influence degree, the matching degree and the association support degree through a large model to convert the data, the interaction influence degree, the matching degree and the association support degree into target cue words; and then performing abnormal score calculation on the target prompt word to obtain an abnormal score, and comparing the abnormal score with a set threshold TA to obtain an abnormal detection result. The invention discloses an air compressor fault prediction system based on a large model. According to the method, an efficient, reliable and explainable fault prediction and decision support means is provided in intelligent operation and maintenance of the air compressor, and the intelligent level of equipment management is remarkably improved.

Description

Technical Field

[0001] The present invention relates to electronic digital data processing, and more particularly to a large model-based air compressor fault prediction method and system. Background Art

[0002] In recent years, with the rapid accumulation of industrial equipment operation data and the continuous development of artificial intelligence technology, fault prediction technology based on big data and deep learning has gradually emerged. Traditional fault detection methods usually rely on manual experience and simple threshold settings, which are difficult to cope with the complex operating environment and changing working conditions of industrial equipment. Existing technologies focus more on how to effectively integrate multi-source data, such as time series data, environmental parameters, sensor outputs, etc., and identify fault patterns through machine learning algorithms. However, traditional models have limitations in processing heterogeneous data, real-time response, and interpretability, and are difficult to meet the high standards of intelligent manufacturing for predictive maintenance of equipment. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide an air compressor fault prediction method and system based on a large model in view of the shortcomings of the existing technology, which mainly solves the problems of difficulty in multi-source data fusion, limited anomaly detection and interpretation capabilities, poor model adaptability, data scarcity and weak generalization ability.

[0004] The present invention provides a method for predicting air compressor failure based on a large model, the method comprising:

[0005] The data classification processing step is used to classify the collected air compressor data and perform abnormal analysis on each level of data in turn to obtain the interaction impact value, matching degree and correlation support value between the data at each level;

[0006] The LLM abnormality judgment and interpretation step is used to normalize and textually analyze the data at all levels, interaction influence, matching degree and association support through the large model to convert them into target prompt words; then the target prompt word is subjected to abnormality score calculation to obtain an abnormality score, and the abnormality score is compared with the set threshold T A The results of abnormality detection are obtained by comparison.

[0007] Traditional methods are difficult to effectively integrate a variety of heterogeneous data, resulting in one-sided prediction results. The present invention uses multimodal reprogramming technology to achieve deep integration of multi-source data such as time series, sensors, and text records. In addition, the existing technology is insufficient in explaining complex fault modes. In the present invention, semantic graphs and adaptive prompt design are combined to provide transparent abnormal cause analysis.

[0008] As a further improvement, the classified air compressor data is divided into primary data, secondary data and tertiary data, wherein:

[0009] The primary data is the core data, which is used to analyze whether there is any abnormality in the air compressor. If so, the secondary data is subjected to interactive impact analysis;

[0010] Secondary data is auxiliary data, which is used to calculate the interactive impact value with primary data based on its characteristic mean and variance;

[0011] The third-level data is associated historical data, which is used to analyze the matching degree between the data and the historical failure mode, and calculate the association support value according to the matching degree.

[0012] Furthermore, the analysis of whether the air compressor is abnormal may be performed in the following specific manner:

[0013] Perform time series analysis or trend detection on the time series of the primary data to obtain an abnormal time series; calculate the residual ∈(t) based on the time series of the primary data and its abnormal time series, compare the residual ∈(t) with a preset threshold T, and determine whether there is an abnormality in the core data of the air compressor:

[0014]

[0015] in, X(t) represents the time series of primary data; Indicates abnormal timing.

[0016] Furthermore, the interaction impact value is calculated by the following formula:

[0017] I X,Y =w1*g1(μ Y ,X(t))+w2*g2(σ Y ,X(t));

[0018] Among them, I X,Y is the interaction influence value; w1 and w2 are feature weights; It represents the cumulative impact of the integral calculation feature mean on the operating data within a specific time window;

[0019] Indicates the rate of change of primary data; μ Y is the characteristic mean; σ Y is the variance; X(t) represents the time series of the primary data; the sequence of the secondary data is represented by Y = {y1, y2, ..., y n}.

[0020] Furthermore, the feature mean is calculated by the following formula:

[0021]

[0022] The variance is calculated by the following formula:

[0023]

[0024] Furthermore, the correlation support value is calculated by the following formula:

[0025]

[0026] Where S is the correlation support value; COV(R Z,H , M) represents covariance; R Z,H represents the matching degree between the third-level data Z and the historical failure mode H; M represents the failure mode; σ R and σ M are the standard deviation of the matching degree and failure mode, respectively.

[0027] As a further improvement, the method also includes an optimization and data enhancement step for dynamically adjusting the feature weights and prediction strategies of the large model based on the anomaly scores currently described.

[0028] Furthermore, the optimization and data enhancement steps specifically include:

[0029] The first step is to optimize the feature weights according to the anomaly score:

[0030]

[0031] Among them, w i (t) is the feature weight of the i-th feature at time t; α represents the learning rate; Represents the feature pair anomaly score Contribution function of i Represents the features extracted from the air compressor data;

[0032] Step 2: According to the optimized feature weights, the prompt generation module in the large model automatically adjusts the task guidance content:

[0033]

[0034] Among them, w i is the optimized feature weight; I X,Y is the interaction effect value; R Z,H represents the matching degree; S is the correlation support value;

[0035] The third step is to provide training samples for the optimized feature weights through GANs.

[0036] When faced with changes in equipment operating conditions, traditional models have poor adaptive adjustment capabilities. The present invention uses a real-time weight adjustment mechanism to dynamically optimize the prediction strategy. For the problem of data scarcity, the present invention introduces generative adversarial networks (GANs) to generate synthetic fault data, improve the model training effect and the ability to detect unknown faults, thereby significantly improving the accuracy and reliability of fault prediction.

[0037] Furthermore, the method also includes model output and visualization steps.

[0038] A large model-based air compressor fault prediction system, comprising:

[0039] The data classification processing module is used to classify the collected air compressor data and perform abnormal analysis on each level of data in turn to obtain the interaction impact value, matching degree and correlation support value between the data at each level;

[0040] The LLM abnormality judgment and interpretation module is used to normalize and textually analyze the data at all levels, interaction influence, matching degree and association support through a large model to convert them into target prompt words; then the target prompt words are subjected to abnormality score calculation to obtain abnormality scores, and the abnormality scores are compared with the set threshold T A The results of abnormality detection are obtained by comparison.

[0041] Beneficial Effects

[0042] The advantages of the present invention are:

[0043] 1. Explanation and operability: The present invention can not only detect anomalies, but also provide detailed cause analysis and optimization suggestions, providing direct action guidance for operation and maintenance personnel.

[0044] 2. Data-driven intelligence: The present invention adopts hierarchical processing and fusion of multimodal data, so that the model can make full use of multi-source information and improve the reliability and comprehensiveness of the prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 The figure is a flow chart of the air compressor fault prediction method of the present invention. DETAILED DESCRIPTION

[0046] The present invention will be further described below in conjunction with the embodiments, but it does not constitute any limitation to the present invention. Any limited number of modifications made by anyone within the scope of the claims of the present invention are still within the scope of the claims of the present invention.

[0047] See also Figure 1The present invention provides a method for predicting air compressor faults based on a large model, including a data classification processing step, an LLM abnormality judgment and interpretation step, an optimization and data enhancement step, and a model output and visualization step. Among them, LLM is a large language model, also referred to as a large model. The four steps are specifically introduced as follows.

[0048] About the data classification and processing steps. In this solution, various sensors can be used to obtain data as shown in Table 1. Data collected over a period of time includes 13 types of air compressor data: environment, temperature, air compressor operating load, pipe diameter, leak detector data, air compressor station exhaust status, condensate discharge status, customer site inspection form, customer fault repair record, pressure sensor, flow meter, electric meter, dew point. We perform classification processing based on the characteristics and importance of these data. The specific data classification is shown in Table 1.

[0049]

[0050] Table 1 Data classification

[0051] For primary data, it is core data and needs real-time monitoring and feature extraction. Among them, the time series analysis method can use the autoregressive moving average model, the trend detection can use the Kalman filter, and the threshold setting can use statistical thresholds and other methods. The above methods are not limited, as long as they can achieve the goal. Take the pressure data and the air compressor operating load as an example, as shown in the following formula.

[0052] Assume that the first-level data is x1, x2, ..., x n ,

[0053]

[0054] Where X(t)={x1,x2,...,x n} represents the time series of any one of the primary data of pressure, flow and load; It represents the abnormal time series obtained after time series analysis or trend detection of X(t).

[0055] Anomaly detection is obtained by comparing the residual ∈(t) with the threshold T:

[0056]

[0057] in,

[0058] For example, by monitoring the pressure changes in the air compressor system in real time, time series analysis (such as LSTM or ARIMA) can be used to identify abnormal fluctuations. Abnormal pressure may indicate system leakage or compression efficiency problems. Timely detection of abnormalities can avoid overpressure operation or low-pressure shutdown of the equipment. The load curve is used to analyze whether the equipment is operating within the normal load range. Trend detection methods (such as the Holt-Winters method) can be used to analyze the load change trend and detect overload or underload conditions. The above method can be used to determine the load state of the equipment operation. Overload may cause increased motor losses, while underload may indicate system leakage or insufficient demand. Combined with the above information, after discovering frequent and irregular fluctuations in the operating load curve, further verification through abnormal changes in pressure data can determine whether the air compressor has internal leakage problems.

[0059] For secondary data, it is auxiliary data, taking temperature data and exhaust status and condensate discharge as examples. The sequence of secondary data is expressed as: Y[n] = {y1, y2, ..., y n},

[0060] Its characteristic mean and variance are expressed as:

[0061]

[0062] The interactive impact value between the secondary data and the primary data is calculated based on the characteristic mean and variance of the secondary data:

[0063] I X,Y =f(μ Y ,σ Y ,X(t))=w1*g1(μ Y ,X(t))+w2*g2(σ Y ,X(t)).

[0064] Among them, w1 and w2 are weight parameters used to balance the influence of mean and variance.

[0065] It represents the cumulative impact of the integral calculation mean on the operating data within a specific time window, and can be used to measure the potential impact of high temperature on the increase in equipment pressure.

[0066] It means that the impact of unstable environment on the dynamic response of equipment is quantified by multiplying the environmental variance with the rate of change of equipment data. It is suitable for identifying equipment operation fluctuations caused by environmental changes.

[0067] Finally, the abnormal correlation is obtained by the following formula:

[0068]

[0069] Among them, T YIt is a threshold set based on historical equipment data or expert experience, and is used to distinguish the degree of correlation between normal and abnormal conditions.

[0070] For temperature data, the stability of the air compressor operating temperature is analyzed through statistical analysis (such as mean, variance, etc.) and time series detection (such as moving average), and abnormal conditions of too high or too low temperature are identified. Too high temperature may affect the compression efficiency and cause the equipment overheating protection to be triggered, while too low temperature may affect the condensate discharge. For exhaust status and condensate discharge data, the effectiveness of the equipment cooling and drainage system is judged by using temperature, humidity and dew point data, combined with the exhaust and condensate discharge status of the air compressor station. Use threshold detection to set the normal range of exhaust and condensate. Poor exhaust or improper condensate discharge may cause the internal temperature of the system to rise or moisture to accumulate, affecting the normal operation of the air compressor. Combining the above information, through the analysis of ambient temperature data and exhaust status, it can be found that the exhaust system is not smooth during high temperature, resulting in excessive temperature of the air compressor. Combined with abnormal pressure, it can be inferred that the equipment cooling system is faulty.

[0071] For the third-level data, as related historical data, analysis of inspection records and maintenance data has a high reference value for current faults. The following introduces the processing of the third-level data.

[0072] First, the matching and correlation calculation of the three-level data are as follows:

[0073] R Z,H =Match(Z,H).

[0074] Where R Z,H It represents the matching degree between the tertiary data Z and the historical fault pattern H. Match() can be a distance metric (such as cosine similarity, Euclidean distance) or a statistical matching function, which matches the currently detected data with the historical pattern to identify whether there is a known fault mode.

[0075] Second, perform semantic analysis and fault mode extraction on the third-level data:

[0076] M = SenmanticAnalysis (Z).

[0077] Among them, M represents the fault mode extracted from the three-level data, and the key fault description is identified from the text data such as inspection tables and maintenance records through semantic analysis technology. SenmanticAnalysis() can be a TF-IDF, Bert and other models to extract key information and tags from the text.

[0078] Third, the matching degree R Z,H Combined with the fault mode M, the correlation support value is generated:

[0079]

[0080] Where S is the correlation support value, which provides background support for the current anomaly detection. Z,H ,M) represents the covariance, σ R and σ M are the standard deviation of the matching degree and failure mode, respectively.

[0081] Regarding the LLM abnormality judgment and interpretation step, in this step, LLM is used to make a comprehensive judgment on the various results and data calculated in the previous step. Here, LLM can be a proprietary model for a vertical field (air compressor) or a general large model (the recommended parameter size is above 13B).

[0082] In this step, multi-source data integration is first performed, that is, before LLM abnormality judgment, the results of data processing at all levels are integrated to ensure that LLM can make full use of the indicators calculated previously. The specific method is as follows:

[0083] Input=PreProcess[X,Y,Z,I X,Y ,R Z,H ,S].

[0084] Where X, Y, and Z are primary, secondary, and tertiary data respectively; X,Y is the interaction effect value between the core data and the auxiliary data, which has been calculated in the previous step; R Z,H is the matching degree between the third-level data and the historical failure mode; S represents the associated support value. PreProcess() represents the process of normalizing and textualizing the above data information and converting it into a suitable prompt.

[0085] Then, LLM analyzes the anomaly: LLM receives the integrated input, identifies the anomaly pattern through reasoning and analysis, and makes anomaly judgment based on the interaction effect, matching degree and associated support value, namely:

[0086]

[0087] in, is the output of LLM, i.e., the detected anomaly score, which comprehensively considers various interaction and support indicators. A Compare and thus obtain the result of abnormal judgment.

[0088] Regarding the anomaly analysis of LLM, it is necessary to explain that if the selected LLM is a general model, in this step, the domain support library is needed as the external knowledge base of the model, and the results of the LLM analysis anomaly are analyzed and output through the search enhancement generation technology. If the selected LLM is a special model, there is no need for an external knowledge base. In this scenario, the default model has knowledge in the field of air compressor anomaly detection.

[0089] Let's take a specific example: Based on the above, by processing the three levels of data, the interaction effect value I of pressure and temperature is obtained: X,Y is 12, the background data matches R Z,H is 0.85, and the associated support value S is 0.7. LLM combines these results and scores the detected anomalies is 0.92, exceeding the threshold T A 0.7. Causal analysis shows that the current abnormality is mainly caused by high pressure caused by high temperature and has a high correlation with the similarity of historical condensate drainage failure. The final explanation output is: "High pressure abnormality is detected, the main reason is poor condensate drainage, the correlation support is high, and it is recommended to check the drainage system."

[0090] Regarding the optimization and data enhancement step. This step aims to dynamically adjust the model's attention weight and prediction strategy based on the current anomaly score to improve the model's ability to respond to abnormal data. The specific steps are:

[0091] Step 1: Real-time optimization: In the data classification processing step and the LLM anomaly judgment and interpretation step, each feature has a different impact on anomaly detection, so it is necessary to dynamically adjust the weight of each feature according to the detection results. Suppose each feature in the data classification processing step and the LLM anomaly judgment and interpretation step is F i , whose weight is w i , both are initially set to 1, and the optimization strategy is as follows:

[0092]

[0093] w i (t) is the weight of the i-th feature at the current moment; α represents the learning rate, which is used to control the adjustment amplitude of the weight; Representation feature F i Scoring anomalies The contribution function determines the direction and magnitude of feature weight adjustment. g() can be the Pearson correlation coefficient formula.

[0094] Step 2: Adaptive Prompt Generation and Task Guidance: Based on the optimized feature weights, the Prompt Generation module automatically adjusts the task guidance content to more accurately guide the LLM prediction process. Specifically:

[0095]

[0096] In this step, w i It is the optimized feature weight, ensuring that the generated prompt can better guide the model to focus on important features. For example: pay attention to the impact of high temperature on load and detect whether there is abnormal fluctuation.

[0097] Step 3: GANs data enhancement: Provide more corresponding training samples for the adjusted feature weights to improve the learning ability of the model. The formula is as follows:

[0098]

[0099] Among them, the generator G is not only based on the noise z, but also based on the feature weight w i Generate samples, focusing on simulating abnormal data corresponding to high-weight features. Through the above strategy, focus on generating abnormal samples corresponding to high-weight features. For example, if the weight of high temperature w 温度 Increased, GANs prioritizes generating fault samples related to high temperature.

[0100] About the model output and visualization steps. The output content includes anomaly detection results, causes, and optimization suggestions. Its visualization can be displayed through charts: graphically presenting the operating status of the equipment, the location of the anomaly, and the severity. It also provides real-time monitoring, such as dynamically updated dashboards, showing changes in key indicators. Reports can also be generated: detailed text analysis and operation suggestions, assisting quick decision-making, etc.

[0101] The present invention also discloses an air compressor fault prediction system based on a large model, comprising:

[0102] The data classification processing module is used to classify the collected air compressor data and perform abnormal analysis on each level of data in turn to obtain the interaction impact value, matching degree and correlation support value between the data at each level;

[0103] The LLM abnormality judgment and interpretation module is used to normalize and textually analyze the data at all levels, interaction influence, matching degree and association support through a large model to convert them into target prompt words; then the target prompt words are subjected to abnormality score calculation to obtain abnormality scores, and the abnormality scores are compared with the set threshold T A The results of abnormality detection are obtained by comparison.

[0104] In traditional methods, it is difficult to effectively integrate multiple heterogeneous data, resulting in one-sided prediction results. However, the present invention uses multimodal reprogramming technology to achieve deep integration of multi-source data such as time series, sensors, and text records. In addition, the existing technology is insufficient in explaining complex fault modes. In the present invention, semantic graphs and adaptive prompt design are combined to provide transparent abnormal cause analysis.

[0105] The above is only a preferred embodiment of the present invention. It should be pointed out that for those skilled in the art, several modifications and improvements can be made without departing from the structure of the present invention, which will not affect the effect of the implementation of the present invention and the practicality of the patent.

Claims

1. A method for predicting air compressor failure based on a large model, characterized in that: The method includes: The data classification processing step is used to classify the collected air compressor data and perform abnormal analysis on each level of data in turn to obtain the interaction impact value, matching degree and correlation support value between the data at each level; The LLM abnormality judgment and interpretation step is used to normalize and textually analyze the data at all levels, interaction influence, matching degree and association support through the large model to convert them into target prompt words; then the target prompt word is subjected to abnormality score calculation to obtain an abnormality score, and the abnormality score is compared with the set threshold T A The results of abnormality detection are obtained by comparison.

2. The air compressor fault prediction method based on a large model according to claim 1 is characterized in that: The classified air compressor data is divided into primary data, secondary data and tertiary data, wherein: The primary data is the core data, which is used to analyze whether there is any abnormality in the air compressor. If so, the secondary data is subjected to interactive impact analysis; Secondary data is auxiliary data, which is used to calculate the interactive impact value with primary data based on its characteristic mean and variance; The third-level data is associated historical data, which is used to analyze the matching degree between the data and the historical failure mode, and calculate the association support value according to the matching degree.

3. The air compressor fault prediction method based on a large model according to claim 2 is characterized in that: The specific analysis method for analyzing whether the air compressor is abnormal is as follows: Perform time series analysis or trend detection on the time series of the primary data to obtain an abnormal time series; calculate the residual ∈(t) based on the time series of the primary data and its abnormal time series, compare the residual ∈(t) with a preset threshold T, and determine whether there is an abnormality in the core data of the air compressor: in, X(t) represents the time series of primary data; Indicates abnormal timing.

4. The air compressor fault prediction method based on a large model according to claim 2 is characterized in that: The interaction impact value is calculated by the following formula: I X,Y =w1*g1(μ Y ,X(t))+w2*g2(σ Y ,X(t)); Among them, I X,Y is the interaction influence value; w1 and w2 are feature weights; It represents the cumulative impact of the integral calculation feature mean on the operating data within a specific time window; Indicates the rate of change of primary data; μ Y is the characteristic mean; σ Y is the variance; X(t) represents the time series of the primary data; the sequence of the secondary data is represented by Y = {y1, y2, ..., y n }.

5. The method for predicting air compressor failure based on a large model according to claim 4 is characterized in that: The feature mean is calculated by the following formula: The variance is calculated by the following formula:

6. The method for predicting air compressor failure based on a large model according to claim 2, characterized in that: The association support value is calculated by the following formula: Where S is the correlation support value; COV(R Z,H , M) represents covariance; R Z,H represents the matching degree between the third-level data Z and the historical failure mode H; M represents the failure mode; σ R and σ M are the standard deviation of the matching degree and failure mode, respectively.

7. The method for predicting air compressor failure based on a large model according to any one of claims 1 to 6, characterized in that: The method also includes an optimization and data enhancement step for dynamically adjusting the feature weights and prediction strategy of the large model based on the anomaly score currently described.

8. The large model-based air compressor fault prediction method according to claim 5, characterized in that: The optimization and data enhancement steps specifically include: The first step is to optimize the feature weights according to the anomaly score: Among them, w i (t) is the feature weight of the i-th feature at time t; α represents the learning rate; Represents the feature pair anomaly score Contribution function of i Represents the features extracted from the air compressor data; Step 2: According to the optimized feature weights, the prompt generation module in the large model automatically adjusts the task guidance content: Among them, w i is the optimized feature weight; I X,Y is the interaction effect value; R Z,H represents the matching degree; S is the correlation support value; The third step is to provide training samples for the optimized feature weights through GANs.

9. The large model-based air compressor fault prediction method according to claim 5, characterized in that: The method also includes model output and visualization steps.

10. An air compressor fault prediction system based on a large model, characterized in that: include: The data classification processing module is used to classify the collected air compressor data and perform abnormal analysis on each level of data in turn to obtain the interaction impact value, matching degree and correlation support value between the data at each level; The LLM abnormality judgment and interpretation module is used to normalize and textually analyze the data at all levels, interaction influence, matching degree and association support through a large model to convert them into target prompt words; then the target prompt words are subjected to abnormality score calculation to obtain abnormality scores, and the abnormality scores are compared with the set threshold T A The results of abnormality detection are obtained by comparison.

Images