Air compressor fault monitoring and operation and maintenance management method and system based on AI technology

Abstract

The invention discloses an air compressor fault monitoring and operation and maintenance management method and system based on the AI technology, and relates to the technical field of air compressors, the key points of the scheme are as follows: based on the operation data of an air compressor, reading pressure and temperature data to carry out data standardization processing, and utilizing a CNN network to carry out feature extraction on the pressure and temperature data; and identifying an abnormal fluctuation mode of sudden increase or sudden decrease, and monitoring the change trend of the abnormal mode to obtain abnormal mode tracking information. The method is used for carrying out continuous monitoring and feature extraction on pressure and temperature data, an abnormal fluctuation mode can be effectively recognized, the development trend of the abnormal fluctuation mode can be predicted, short-term data abnormity can be captured through application of the long-term and short-term memory network, long-term data changes can be analyzed, prediction of potential risks in the future is provided, and the method is suitable for popularization and application. The accuracy and timeliness of fault prediction are improved, so that a maintenance team can carry out intervention before the problem develops to a serious fault, and the downtime caused by the fault is remarkably reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of air compressors, and particularly to a method and system for air compressor fault monitoring and operation and maintenance management based on AI technology. Background Art

[0002] An air compressor is a device used to compress gas. The air compressor is similar in structure to a water pump. Most air compressors are reciprocating pistons, rotary vanes or rotary screws, involving the compression of air or other gases for various industrial and commercial uses. This device plays a core role in the manufacturing, construction, medical and energy industries, and is used to provide power, control the process flow or conduct gas transmission. The key aspects of air compressor technology include improving energy efficiency, reducing operating costs, enhancing the reliability and safety of the system.

[0003] Among them, the method for air compressor fault monitoring and operation and maintenance management based on AI technology uses artificial intelligence technology to enhance the efficiency of air compressor fault detection and maintenance management. By analyzing a large amount of data collected from the air compressor, it predicts equipment failures and performance degradation, enabling the maintenance team to intervene in advance, reduce downtime, and at the same time support the implementation of predictive maintenance strategies to reduce maintenance costs and improve the operating efficiency of the equipment.

[0004] Traditional technologies rely on periodic inspections for fault monitoring and maintenance management, cannot reflect the current operating state of the equipment in real time, and are difficult to capture immediately occurring faults, resulting in major production interruptions and high maintenance costs. The lack of real-time monitoring and prediction capabilities means that fault handling starts only after the problem has occurred, resulting in delayed maintenance responses and being unable to effectively prevent the occurrence of faults. In addition, the inability to conduct long-term data trend analysis makes predictive maintenance almost impossible to implement, resulting in premature or late maintenance of the equipment before reaching the optimal maintenance time point, affecting the overall operating efficiency and lifespan of the equipment. The lack of accurate fault prediction and classification also leads to improper allocation of maintenance resources, increasing unnecessary operations and costs. Summary of the Invention

[0005] The present invention provides a method and system for air compressor fault monitoring and operation and maintenance management based on AI technology.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A method for air compressor fault monitoring and operation and maintenance management based on AI technology includes the following steps:

[0008] S1: Based on the operating data of the air compressor, read the pressure and temperature data for data standardization processing, use the CNN network to extract features from the pressure and temperature data, identify abnormal fluctuation patterns of sudden increase or decrease, and monitor the change trend of the abnormal patterns to obtain abnormal pattern tracking information;

[0009] S2: Based on the abnormal pattern tracking information, extract the key abnormal signals of rapid pressure drop and enhanced vibration, compare them with the historical fault data of the air compressor, identify the current fault category, and update the fault data record of the air compressor to obtain a fault type database;

[0010] S3: Based on the fault type database, use a long short-term memory network to analyze time series data, identify the long-term change trends of pressure and temperature, predict the change of the operating state of the air compressor in the future time period, and then optimize the prevention ability of potential problems according to the prediction results to obtain fault event prediction data;

[0011] S4: Based on the fault event prediction data, analyze the operating data of the lubrication system and air filtration device of the air compressor, determine the key maintenance points through continuous monitoring of vibration signals and pressure readings, allocate maintenance resources with reference to the maintenance points, and notify the maintenance personnel to intervene in real time to obtain an optimized maintenance configuration.

[0012] The improvement of the present invention is that the steps of data standardization processing are specifically as follows:

[0013] S111: Based on the operating data of the air compressor, extract the original pressure and temperature data, calculate the mean and standard deviation of each parameter to obtain the mean and standard deviation of the temperature and pressure data;

[0014] S112: Based on the mean and standard deviation of the temperature and pressure data, use the formula:

[0015]

[0016] Process the temperature and pressure data to obtain the standardized temperature and pressure data Z, where X represents the original data point, which is a single pressure or temperature measurement value, μ is the mean, representing the average value of the pressure or temperature data, and σ is the standard deviation, representing the standard deviation of the pressure or temperature data.

[0017] The improvement of the present invention is that the steps of obtaining the abnormal pattern tracking information are specifically as follows:

[0018] S121: Use the CNN network to extract features from the standardized pressure and temperature data, identify the abnormal fluctuation patterns of sudden increase or decrease in the data, and obtain preliminary abnormal feature information;

[0019] S122: Perform time series analysis on the preliminary abnormal feature information, using the formula:

[0020]

[0021] Calculate the trend degree TQ, monitor the change trend of the abnormal pattern, and obtain the abnormal pattern tracking information, where fq i represents the eigenvalue of the i-th data point within the window, μq represents the mean value of all data points within the window, σq represents the standard deviation of all data points within the window, and n is the window size.

[0022] The improvement of the present invention is that the step of identifying the current fault category is specifically as follows:

[0023] S211: Based on the abnormal pattern tracking information, extract the data points with a sharp pressure drop and enhanced vibration, and separate the key abnormal signals to obtain abnormal signal data;

[0024] S212: Based on the abnormal signal data, use the formula:

[0025]

[0026] Calculate the distance from each data point to the cluster center, determine the data points belonging to the fault mode, and obtain the potential fault mode information, where DS k represents the Euclidean distance from the data point to the cluster center, xs i is the value of a single data point, specifically referring to the specific value in the abnormal signal, such as the measured value of pressure or vibration intensity, represents the position of the k-th cluster center, n s is the total number of data points, that is, the number of abnormal signal data points participating in the clustering analysis;

[0027] S213: Based on the potential fault mode information, compare it with the historical fault data of the air compressor to determine the situation of matching the known fault mode, and obtain the current fault category.

[0028] The improvement of the present invention is that the step of obtaining the fault type database is specifically as follows:

[0029] S221: According to the determined current fault category, extract the data logs matching the current fault, integrate the data logs, and obtain the integrated fault records;

[0030] S222: Based on the integrated fault records, use association rule learning to mine the potential relationships and patterns between the fault data, and update the fault data of the air compressor to obtain the fault type database.

[0031] The improvement of the present invention is that the step of predicting the change of the operating state of the air compressor is specifically as follows:

[0032] S311: Based on the fault type database, use the long short-term memory network to analyze the time series data, identify the change trends of pressure and temperature, and obtain the change trend recognition result;

[0033] S312: Based on the change trend recognition result, combined with the current real-time monitoring data, use the formula:

[0034]

[0035] Predict the operating state of the air compressor within a future time period to obtain an operating state prediction log, where P future represents the pressure trend within the future time period, T current represents the temperature data at the current time point, S recent represents the sum of the pressure data within a recent period of time, n s is the total number of data points, α U represents the weight coefficient of the temperature data, β U represents the weight coefficient of the pressure data.

[0036] The improvement of the present invention is that the step of obtaining the fault event prediction data is specifically as follows:

[0037] S321: According to the prediction result, combined with the fault event data and the current operating parameters of the air compressor, perform fault matching analysis, and compare the predicted pressure trend with the known fault modes to obtain a preliminary fault matching result;

[0038] S322: Based on the preliminary fault matching result, optimize the maintenance process and response ability for potential problems of the air compressor to obtain fault event prediction data.

[0039] The improvement of the present invention is that the step of obtaining the maintenance optimization configuration is specifically as follows:

[0040] S411: Based on the fault event prediction data, analyze the operating data of the lubrication system and air filter device of the air compressor, including vibration signals and pressure readings, to obtain operating state analysis information;

[0041] S412: Based on the operating state analysis information, use the formula:

[0042]

[0043] Classify the maintenance points by priority to determine the key maintenance points, where DK critical represents the decision value of the key maintenance point, VD i represents the vibration signal intensity at the i-th measurement point, N D is the total number of measurement points, γ D is an adjustment coefficient used to adjust the influence degree of the vibration data to adapt to the specific equipment state and response history;

[0044] S413: Based on the key maintenance points, perform resource allocation and maintenance scheduling, notify the maintenance team in real time, and intervene according to the maintenance priority to obtain the optimized maintenance configuration.

[0045] An air compressor fault monitoring and operation and maintenance management system based on AI technology, the system includes:

[0046] The data monitoring module, based on the operation data of the air compressor, reads the pressure and temperature data for data standardization processing, identifies abnormal fluctuation patterns of sudden increase or decrease, and monitors the change trend of the abnormal patterns to obtain abnormal pattern tracking information;

[0047] The fault detection module, based on the abnormal pattern tracking information, extracts key abnormal signals of sharp pressure drop and enhanced vibration, compares them with the historical fault data of the air compressor, and identifies the current fault category to obtain a fault type database;

[0048] The trend analysis module, based on the fault type database, identifies the long-term change trends of pressure and temperature, predicts the change of the operation state of the air compressor in the future time period, and then optimizes the prevention ability for potential problems according to the prediction results to obtain fault event prediction data;

[0049] The maintenance scheduling module, based on the fault event prediction data, determines the key maintenance points by continuously monitoring the vibration signal and pressure readings, allocates maintenance resources with reference to the maintenance points, notifies the maintenance personnel to intervene in real time, and obtains the optimized maintenance configuration.

[0050] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0051] In the present invention, by continuously monitoring and feature extracting the pressure and temperature data, abnormal fluctuation patterns can be effectively identified and their development trends can be predicted. The application of the long short-term memory network enables not only capturing short-term data anomalies but also analyzing long-term data changes, providing predictions of future potential risks, improving the accuracy and timeliness of fault prediction, enabling the maintenance team to intervene before the problem develops into a serious fault, significantly reducing the downtime caused by faults, effectively improving the maintenance efficiency and equipment reliability, and at the same time reducing the operation and maintenance costs by optimizing the maintenance resource configuration and notifying the maintenance personnel in a timely manner. Description of the Drawings

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. The drawings are only used for the purpose of showing the implementation manners and are not considered as a limitation to the present invention.

[0053] Figure 1 It is the main step flow chart in the embodiments of the present invention;

[0054] Figure 2 This is the flowchart of data standardization processing in the embodiments of the present invention;

[0055] Figure 3 This is the flowchart of obtaining abnormal mode tracking information in the embodiments of the present invention;

[0056] Figure 4 This is the flowchart of identifying the current fault category in the embodiments of the present invention;

[0057] Figure 5 This is the flowchart of obtaining the fault type database in the embodiments of the present invention;

[0058] Figure 6 This is the flowchart of predicting the change of the operating state of the air compressor in the embodiments of the present invention;

[0059] Figure 7 This is the flowchart of obtaining the fault event prediction data in the embodiments of the present invention;

[0060] Figure 8 This is the flowchart of obtaining the maintenance optimization configuration in the embodiments of the present invention. Detailed implementation manners

[0061] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without making creative efforts belong to the scope of protection of the present invention.

[0062] Unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs; the terms used in the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The terms "including" and "having" in the specification and claims of the present invention and any variations thereof are intended to cover non-exclusive inclusion.

[0063] In the description of the embodiments of the present invention, technical terms such as "first" and "second" are only used to distinguish different objects and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity, specific order or primary-secondary relationship of the indicated technical features. In the description of the embodiments of the present invention, "a plurality of" means two or more unless otherwise specifically defined.

[0064] In the description of the embodiments of the present invention, the term "and / or" is merely an association relationship describing associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Additionally, in this article, the character " / " generally represents an "or" relationship between the associated objects before and after.

[0065] In the description of the embodiments of the present invention, the term "plural" refers to two or more (including two). Similarly, "multiple groups" refers to two or more groups (including two groups), and "multiple pieces" refers to two or more pieces (including two pieces).

[0066] In the description of the embodiments of the present invention, the orientation or positional relationship indicated by technical terms such as "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the embodiments of the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation on the embodiments of the present invention.

[0067] In the description of the embodiments of the present invention, unless otherwise clearly specified and limited, technical terms such as "installation", "connection", "coupling", "fixation", etc. should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or integrated; it can also be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present invention can be understood according to specific circumstances.

[0068] Embodiment

[0069] The embodiments of the present invention provide an air compressor fault monitoring and operation and maintenance management method based on AI technology, as Figure 1 shown, including the following steps:

[0070] S1: Based on the operation data of the air compressor, read the pressure and temperature data for data standardization processing, use the CNN network to extract features from the pressure and temperature data, identify abnormal fluctuation patterns of sudden increase or decrease, and monitor the change trend of the abnormal patterns to obtain abnormal pattern tracking information;

[0071] S2: Based on the abnormal pattern tracking information, extract key abnormal signals of sharp pressure drop and enhanced vibration, compare them with the historical fault data of the air compressor, identify the current fault category, and update the fault data record of the air compressor to obtain a fault type database;

[0072] S3: Based on the fault type database, use the long short-term memory network to analyze the time series data, identify the long-term change trends of pressure and temperature, predict the changes in the operating state of the air compressor within the future time period, and then optimize the prevention ability for potential problems according to the prediction results to obtain the fault event prediction data;

[0073] S4: Based on the fault event prediction data, analyze the operating data of the lubrication system and air filtration device of the air compressor, determine the key maintenance points through continuous monitoring of vibration signals and pressure readings, allocate maintenance resources with reference to the maintenance points, and notify the maintenance personnel to intervene in real time to obtain the maintenance optimization configuration.

[0074] The abnormal mode tracking information includes pressure fluctuation characteristics, temperature fluctuation characteristics, and change trend analysis results. The fault type database specifically includes fault identifiers, historical matching records, and data update logs. The fault event prediction data includes trend prediction results, state change analysis results, and prevention response adjustment information. The maintenance optimization configuration includes key maintenance point calibration, resource allocation situation, and personnel intervention notification.

[0075] As Figure 2 shown, the steps of data standardization processing are specifically as follows:

[0076] S111: Based on the operating data of the air compressor, extract the original pressure and temperature data, calculate the mean and standard deviation of each parameter to obtain the mean and standard deviation of the temperature and pressure data;

[0077] Extract the original pressure and temperature data from the recording device, set the sampling frequency according to the time series, extract continuous pressure and temperature values, sort the data according to the timestamp to ensure continuity, and then check whether there are missing values or abnormal points in the data. The missing values can be filled by interpolation methods, and the abnormal points are screened and removed by calculating the relationship between the current value and the standard deviation of the data within its time window. The data after removal is normalized and sorted again to ensure compliance with time correlation. Immediately calculate the mean and standard deviation of the processed data, respectively use the sliding window method to solve the local mean and global mean for the pressure and temperature data, and synchronously calculate the standard deviation of its distribution. Finally, obtain the mean and standard deviation of the temperature and pressure data.

[0078] S112: Based on the mean and standard deviation of the temperature and pressure data, use the formula:

[0079]

[0080] Process the temperature and pressure data to obtain the standardized temperature and pressure data Z. Here, X represents the original data point, which is a single pressure or temperature measurement value. μ is the mean, representing the average value of the pressure or temperature data, and σ is the standard deviation, representing the standard deviation of the pressure or temperature data.

[0081] For a certain pressure or temperature data point X, its mean μ and standard deviation σ are respectively from the statistical results of the extracted and calculated original pressure and temperature data. If the mean of the pressure is 80 and the standard deviation is 10, and the mean of the temperature is 300 and the standard deviation is 15, substitute them into the formula for calculation:

[0082] If a certain pressure data point is 85, then the standardized value is:

[0083]

[0084] If a certain temperature data point is 315, then the standardized value is:

[0085]

[0086] This result shows that through the standardization formula, the standardized results of pressure and temperature, 0.5 and 1.0 respectively, are obtained, representing the degree of deviation from their respective distributions. The results will be used as the input data for the subsequent steps.

[0087] As Figure 3 shown, the specific steps for obtaining the abnormal pattern tracking information are as follows:

[0088] S121: Use the CNN network to extract features from the standardized pressure and temperature data, identify the abnormal fluctuation patterns of sudden increases or decreases in the data, and obtain the preliminary abnormal feature information.

[0089] By reading the standardized data matrix of pressure and temperature, taking it as the input data, fragmenting it and passing it into the first convolutional layer of the CNN model. The convolutional layer performs local feature extraction on the input data by using predefined weights and generates the initial feature map. Then, perform pooling operations on the feature map to reduce the feature dimension through max pooling or average pooling while retaining the main features. Subsequently, in the fully connected layer, expand all the pooled feature maps into a one-dimensional vector to capture the global correlation between different regional features, screen the non-linear features through the activation function and optimize their distribution. Finally, obtain the feature vector corresponding to each input data segment, including the eigenvalue and its fluctuation pattern characteristics. Identify the sudden increase or decrease fluctuation patterns in the data through this process and generate the preliminary abnormal feature information.

[0090] S122: Conduct time series analysis on the preliminary abnormal feature information, using the formula:

[0091]

[0092] Calculate the trend degree TQ, monitor the change trend of the abnormal pattern, and obtain the abnormal pattern tracking information, where fq i represents the eigenvalue of the i-th data point within the window, μq represents the mean of all data points within the window, σq represents the standard deviation of all data points within the window, and n is the window size;

[0093] Consider a specific window with a window size n of 5, and the eigenvalues fq within the window i are 10, 12, 9, 11, 8. The mean μq of the eigenvalues within the window is 10, and the standard deviation σq is 1.58. Substitute into the formula for calculation:

[0094]

[0095] TQ = 0;

[0096] This result indicates that within this window, the average deviation degree of the eigenvalues is zero, indicating that there is no significant abnormal trend within this window.

[0097] As Figure 4 shown, the steps to identify the current fault category are specifically as follows:

[0098] S211: Based on the abnormal pattern tracking information, extract the data points with a sharp pressure drop and increased vibration, and separate the key abnormal signals to obtain the abnormal signal data;

[0099] First, from the collected pressure and vibration data, by comparing the deviation of the current pressure value from the historical mean and the abnormal increase in the vibration amplitude, filter out the initial data points belonging to the abnormality. The data points need to pass the threshold judgment, and the threshold is set based on the statistical analysis of the known fault states in the historical data. For example, a sharp pressure drop is judged at a certain rate of decrease or a multiple lower than the historical mean, while an increase in vibration can be determined by exceeding a certain critical value of vibration acceleration or frequency amplitude. Subsequently, the preliminary abnormal signals after screening are sorted through time window division, marked and separated with signal intensity and time interval as features, and finally form the abnormal signal data for subsequent fault analysis.

[0100] S212: Based on the abnormal signal data, use the formula:

[0101]

[0102] Calculate the distance from each data point to the cluster center, judge the data points belonging to the fault mode, and obtain the potential fault mode information, where DS k represents the Euclidean distance from the data point to the cluster center, xs iis the value of a single data point, specifically referring to the specific numerical value in the abnormal signal, such as the measured value of pressure or vibration intensity, represents the position of the k-th clustering center, n s is the total number of data points, that is, the number of abnormal signal data points participating in the clustering analysis;

[0103] Let the total number of data points n s be 5, and the abnormal signal values xs i are [120, 130, 115, 123, 118] respectively, and the clustering center is 125, and the calculation process is as follows:

[0104]

[0105] DS k ≈14.247;

[0106] This result indicates that the calculated average distance value is approximately 14.247, indicating the average distance between the data points and the clustering center, which is used to evaluate the matching degree of the fault mode.

[0107] S213: Based on the potential fault mode information, compare it with the historical fault data of the air compressor to determine the situation of matching known fault modes and obtain the current fault category;

[0108] Compare the features extracted from the abnormal signal item by item with the historical fault database of the air compressor. First, normalize the information of all fault modes in the historical database, including the pressure change curve, vibration frequency range, and time characteristics when the fault occurs. Then, calculate the similarity between the key features of the extracted abnormal signal, such as the pressure change amplitude, descent rate, and the amplitude of vibration enhancement, and the known modes in the database respectively. The calculation of similarity can be carried out by the multi-dimensional Euclidean distance method. Finally, based on the similarity score ranking, select the historical mode that best matches the abnormal signal as the recognition result of the current fault, and record the matching result as the current fault category.

[0109] Such as Figure 5 shown, the specific steps for obtaining the fault type database are as follows:

[0110] S221: According to the determined current fault category, extract the data logs matching the current fault, integrate the data logs, and obtain the integrated fault records;

[0111] Filter data logs that match the current fault category from historical data, using keyword matching technology and timestamp-based log filtering to ensure that the extracted data fully conforms to the feature range of the current fault category, while eliminating irrelevant or redundant log entries. The log records include multi-dimensional data points such as pressure, vibration, and temperature. Subsequently, organize the extracted log data, sort the data in time series to ensure the integrity and clarity of the causal relationship of events, and further use the normalization method to standardize the parameter values in the data, thereby eliminating the dimensional differences between different data sources and laying a foundation for the next step of analysis. Finally, organize the integrated fault records into a unified format for further mining and analysis in the future.

[0112] S222: Based on the integrated fault records, use association rule learning to mine the potential relationships and patterns among the fault data, and update the fault data of the air compressor to obtain a fault type database;

[0113] By mining the potential relationships among the fault data, using the association rule learning method, discretize the multi-dimensional parameters in the fault records, group the continuous variables into discrete intervals, such as dividing the vibration amplitude into three categories: "low", "medium", and "high". Then calculate the frequent item sets, identify the combination patterns that frequently co-occur among multiple parameters by scanning the integrated fault records, sort them according to the support and confidence, extract the most significant association rules, and finally compare the parameters of the rules with the existing parameters in the database to update the fields and association information of the current air compressor fault database.

[0114] As Figure 6 shown, the prediction steps for the change of the operating state of the air compressor are specifically as follows:

[0115] S311: Based on the fault type database, use the long short-term memory network to analyze the time series data, identify the change trends of pressure and temperature, and obtain the change trend recognition results;

[0116] Extract the time series data, which includes pressure and temperature parameters. By cleaning the original data, eliminate the obvious outliers and supplement the missing data to ensure the integrity and accuracy of the data. Subsequently, standardize the time series data, convert the values with different units into a unified proportional range to eliminate the influence of data scale differences. Then, input the cleaned and standardized data into the long short-term memory network (LSTM) model in batches. The model processes the historical information through the memory gate and the forget gate, captures the long-term dependencies and short-term change characteristics in the time series data, and finally completes the analysis of the change trends of pressure and temperature to obtain the change trend recognition results.

[0117] S312: Based on the change trend recognition results, combined with the current real-time monitoring data, use the formula:

[0118]

[0119] Predict the operating state of the air compressor within a future time period to obtain an operating state prediction log, where P future represents the pressure trend within the future time period, and T current represents the temperature data at the current time point, and S recent represents the sum of the pressure data within a recent period of time, and n s is the total number of data points, and α U represents the weight coefficient of the temperature data, and β U represents the weight coefficient of the pressure data;

[0120] The current temperature T current is 25 °C, and the sum of the pressure data within a recent period of time S recent is 1500 psi, and the total number of data points n s is 30, and the weight coefficient α U is 0.6, and β U is 0.4, and substitute into the formula for calculation:

[0121]

[0122] P future ≈5.916;

[0123] This result indicates that the predicted future pressure trend is 5.916 psi, indicating that within the next time period, the operating state of the air compressor will tend to this average pressure level.

[0124] As Figure 7 shown, the steps for obtaining the fault event prediction data are specifically as follows:

[0125] S321: According to the prediction result, combine the fault event data and the current operating parameters of the air compressor, perform fault matching analysis, and compare the predicted pressure trend with the known fault patterns to obtain a preliminary fault matching result;

[0126] Extract the pressure and temperature related parameters in the fault event data, compare the parameters with the currently predicted pressure trend in the form of a time series, calculate the difference between the two and define the fault threshold using the matching rules. Then, use the current operating parameters of the air compressor (including the real-time pressure value, temperature fluctuation range, and operating cycle) to verify whether the change trend of the predicted data is close to the historical fault pattern. Through the matching and comparison of multi-dimensional data, screen out the corresponding fault types, and finally generate a preliminary fault matching result.

[0127] S322: Based on the preliminary fault matching result, optimize the maintenance process and response ability for potential problems of the air compressor to obtain the fault event prediction data;

[0128] Optimize the maintenance process for potential problems of air compressors. First, extract the historical maintenance records of all failure types and the corresponding preventive measures, and obtain the required response time, key components, and maintenance parameters from the matching failure types. Then, adjust the priority of the preventive plan according to the operating status data of the current air compressor, and conduct response analysis on the problems that occur, including calculating the expected component life, the rate of decline in operating performance, and the required maintenance frequency. Through the above steps, finally generate the prediction data of failure events comprehensively.

[0129] As Figure 8 shown, the specific steps for obtaining the maintenance optimization configuration are as follows:

[0130] S411: Based on the failure event prediction data, analyze the operating data of the lubrication system and air filtration device of the air compressor, including vibration signals and pressure readings, to obtain the operating status analysis information;

[0131] By obtaining the operating data of the lubrication system and air filtration device of the air compressor, the data includes vibration signals and pressure readings. To ensure the data quality, it is necessary to perform preliminary processing on the collected signals, such as denoising the vibration signals and normalizing the pressure readings. Then, sort the data in chronological order, map the data from different sources to a unified time axis. Next, conduct statistical analysis on the sorted data, calculate the fluctuation range of the vibration signals and the change rate of the pressure readings, so as to form a set of comprehensive operating status indicators, and the indicators are further integrated into the operating status analysis information to provide support for subsequent maintenance.

[0132] S412: Based on the operating status analysis information, use the formula:

[0133]

[0134] Classify the priorities of the maintenance points to determine the key maintenance points, where DK critical represents the decision value of the key maintenance point, VD i represents the vibration signal intensity at the i-th measurement point, N D is the total number of measurement points, γ D is the adjustment coefficient, which is used to adjust the influence degree of the vibration data to adapt to the specific equipment status and response history;

[0135] Considering the vibration signal data set, assume that the total number of measurement points N D is 5, and the vibration signal intensities VD i are [2.0, 3.1, 1.9, 2.5, 3.0] respectively, and the adjustment coefficient γ D is 0.8. The calculation process is as follows:

[0136]

[0137] DK critical = 0.8·(6.494);

[0138] DK critical ≈ 5.2;

[0139] The result shows that the determined value of the calculated key maintenance point is approximately 5.2, which is obtained by quantitatively analyzing vibration data and is used to determine the priority of maintenance points, thereby providing a basis for resource allocation and maintenance scheduling.

[0140] S413: Based on the key maintenance points, perform resource allocation and maintenance scheduling, notify the maintenance team in real time, and intervene according to the maintenance priority to obtain an optimized maintenance configuration;

[0141] Classify the maintenance priorities and match the existing maintenance resources according to the classification results. For example, based on the risk level of the maintenance points and the operating parameters of the equipment, determine the number of maintenance personnel to be dispatched and the tools required, then notify the relevant team in real time to receive maintenance tasks through wireless communication devices, and at the same time upload the allocation situation of maintenance resources to the maintenance management system for subsequent recording and tracking to ensure that the maintenance team can complete the intervention tasks within the specified time range and complete the optimized maintenance configuration.

[0142] An air compressor fault monitoring and operation and maintenance management system based on AI technology, the system includes:

[0143] The data monitoring module, based on the operation data of the air compressor, reads the pressure and temperature data for data standardization processing, identifies abnormal fluctuation patterns of sudden increase or decrease, and monitors the change trend of the abnormal patterns to obtain abnormal pattern tracking information;

[0144] The fault detection module, based on the abnormal pattern tracking information, extracts key abnormal signals of sharp pressure drop and enhanced vibration, compares them with the historical fault data of the air compressor, and identifies the current fault category to obtain a fault type database;

[0145] The trend analysis module, based on the fault type database, identifies the long-term change trends of pressure and temperature, predicts the change of the operating state of the air compressor in the future time period, and then optimizes the prevention ability of potential problems according to the prediction results to obtain fault event prediction data;

[0146] The maintenance scheduling module, based on the fault event prediction data, determines the key maintenance points by continuously monitoring the vibration signals and pressure readings, allocates maintenance resources with reference to the maintenance points, notifies the maintenance personnel to intervene in real time, and obtains an optimized maintenance configuration.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and the description of the present invention. In particular, as long as there is no structural conflict, the technical features mentioned in each embodiment can be combined in any way. The present invention is not limited to the specific embodiments disclosed in the text, but includes all technical solutions that fall within the scope of the claims.

Claims

1. An air compressor fault monitoring and operation and maintenance management method based on AI technology, characterized in that, Including the following steps: S1: Based on the operating data of the air compressor, read the pressure and temperature data for data standardization processing. Use the CNN network to extract features from the pressure and temperature data, identify abnormal fluctuation patterns of sudden increase or decrease, and monitor the change trend of the abnormal patterns to obtain abnormal pattern tracking information; S2: Based on the abnormal pattern tracking information, extract key abnormal signals of sharp pressure drop and enhanced vibration, compare them with the historical fault data of the air compressor, identify the current fault category, and update the fault data record of the air compressor to obtain a fault type database; S3: Based on the fault type database, use the long short-term memory network to analyze time series data, identify the long-term change trends of pressure and temperature, and predict the change of the operating state of the air compressor in the future time period. Then optimize the prevention ability for potential problems according to the prediction results to obtain fault event prediction data; S4: Based on the fault event prediction data, analyze the operating data of the lubrication system and air filtration device of the air compressor. By continuously monitoring the vibration signal and pressure reading, determine the key maintenance points, allocate maintenance resources with reference to the maintenance points, and notify the maintenance personnel to intervene in real time to obtain a maintenance optimization configuration.

2. The air compressor fault monitoring and operation and maintenance management method based on AI technology according to claim 1, wherein The steps of the data standardization processing are specifically as follows: S111: Based on the operating data of the air compressor, extract the original pressure and temperature data, calculate the mean and standard deviation of each parameter to obtain the mean and standard deviation of the temperature and pressure data; S112: Based on the mean and standard deviation of the temperature and pressure data, use the formula: Process the temperature and pressure data to obtain the standardized temperature and pressure data Z. Among them, X represents the original data point, which is a single pressure or temperature measurement value, μ is the mean, representing the average value of the pressure or temperature data, and σ is the standard deviation, representing the standard deviation of the pressure or temperature data.

3. The air compressor fault monitoring and operation and maintenance management method based on AI technology according to claim 1, characterized in that The steps of obtaining the abnormal pattern tracking information are specifically as follows: S121: Use the CNN network to extract features from the standardized pressure and temperature data, identify abnormal fluctuation patterns of sudden increase or decrease in the data, and obtain preliminary abnormal feature information; S122: Perform time series analysis on the preliminary abnormal feature information, using the formula: Calculate the trend degree TQ, monitor the change trend of the abnormal pattern, and obtain the abnormal pattern tracking information, where fq i represents the eigenvalue of the i-th data point within the window, μq represents the mean value of all data points within the window, σq represents the standard deviation of all data points within the window, and n is the window size.

4. The method for monitoring and operation and maintenance management of air compressor faults based on AI technology according to claim 1, wherein, The steps of identifying the current fault category are specifically as follows: S211: Based on the abnormal pattern tracking information, extract the data points of sharp pressure drop and enhanced vibration, and separate the key abnormal signals to obtain abnormal signal data; S212: Based on the abnormal signal data, use the formula: Calculate the distance from each data point to the cluster center, determine the data points belonging to the fault mode, and obtain the potential fault mode information, where DS k represents the Euclidean distance from the data point to the cluster center, and xs i is the value of a single data point, specifically referring to the specific value in the abnormal signal, such as the measured value of pressure or vibration intensity. represents the position of the k-th cluster center, and n s is the total number of data points, that is, the number of abnormal signal data points participating in the clustering analysis; S213: Based on the potential fault mode information, compare it with the historical fault data of the air compressor to determine the situation of matching known fault modes, and obtain the current fault category.

5. The method for monitoring and operation and maintenance management of air compressor faults based on AI technology according to claim 1, characterized in that, The steps of obtaining the fault type database are specifically as follows: S221: According to the determined current fault category, extract the data logs matching the current fault, integrate the data logs to obtain the integrated fault records; S222: Based on the integrated fault records, use association rule learning to mine the potential relationships and patterns between the fault data, and update the fault data of the air compressor to obtain a fault type database.

6. The method for monitoring air compressor faults and operation and maintenance management based on AI technology according to claim 1, characterized in that The specific prediction steps for the operating state change of the air compressor are as follows: S311: Based on the fault type database, use a long short-term memory network to analyze time series data, identify the change trends of pressure and temperature, and obtain a change trend identification result; S312: Based on the change trend identification result, combined with the current real-time monitoring data, use the formula: Predict the operating state of the air compressor within a future time period to obtain an operating state prediction log, where P future represents the pressure trend within the future time period, and T current represents the temperature data at the current time point, and S recent represents the sum of the pressure data within the most recent period of time, and n s is the total number of data points, and α U represents the weight coefficient of the temperature data, and β U represents the weight coefficient of the pressure data.

7. The method for monitoring air compressor faults and operation and maintenance management based on AI technology according to claim 1, wherein, The specific steps for obtaining the fault event prediction data are as follows: S321: According to the prediction result, combined with the fault event data and the current operating parameters of the air compressor, perform fault matching analysis, and compare the predicted pressure trend with the known fault patterns to obtain a preliminary fault matching result; S322: Based on the preliminary fault matching result, optimize the maintenance process and response ability for potential problems of the air compressor to obtain fault event prediction data.

8. The air compressor fault monitoring and operation and maintenance management method based on AI technology according to claim 1, wherein, The specific steps for obtaining the maintenance optimization configuration are as follows: S411: Based on the fault event prediction data, analyze the operating data of the lubrication system and air filtration device of the air compressor, including vibration signals and pressure readings, to obtain operating state analysis information; S412: Based on the operating state analysis information, use the formula: Classify the maintenance points by priority to determine the key maintenance points, where DK critical represents the decision value of the key maintenance point, VD i represents the vibration signal intensity at the i-th measurement point, N D is the total number of measurement points, γ D is an adjustment coefficient used to adjust the influence degree of vibration data to adapt to specific equipment states and response histories; S413: Based on the key maintenance points, perform resource allocation and maintenance scheduling, notify the maintenance team in real time, and intervene according to the maintenance priority to obtain a maintenance optimization configuration.

9. An air compressor fault monitoring and operation and maintenance management system based on AI technology, characterized in that, Execute according to the air compressor fault monitoring and operation and maintenance management method based on AI technology described in any one of claims 1-8. The system includes: The data monitoring module, based on the operating data of the air compressor, reads the pressure and temperature data for data standardization processing, identifies abnormal fluctuation patterns of sudden increase or decrease, and monitors the change trend of the abnormal patterns to obtain abnormal pattern tracking information; The fault detection module, based on the abnormal pattern tracking information, extracts key abnormal signals of sharp pressure drop and enhanced vibration, compares them with the historical fault data of the air compressor, and identifies the current fault category to obtain a fault type database; The trend analysis module, based on the fault type database, identifies the long-term change trends of pressure and temperature, predicts the change of the operating state of the air compressor in the future time period, and then optimizes the prevention ability for potential problems according to the prediction result to obtain fault event prediction data; The maintenance scheduling module, based on the fault event prediction data, continuously monitors the vibration signals and pressure readings to determine the key maintenance points, allocates maintenance resources with reference to the maintenance points, and notifies the maintenance personnel in real time to intervene to obtain a maintenance optimization configuration.

Images