An operation and maintenance management system and method based on industrial Internet of Things

Abstract

The present invention discloses an operation and maintenance management system and method based on the industrial Internet of Things, belonging to the field of operation and maintenance management technology. It includes a real-time equipment monitoring module, a fault diagnosis and early warning module, a predictive maintenance module, a maintenance work order management module and a data analysis and reporting module wirelessly connected to a central controller; the real-time equipment monitoring module is used to collect equipment operation status data and intuitively display the real-time operation status of the equipment, including a real-time equipment data processing unit; the real-time equipment data processing unit is used to perform a first preprocessing on the collected real-time operation status data, and generate effective real-time operation status data as a first preprocessing result. The present invention ensures that the collected real-time operation status data meets the actual work requirements through the first preprocessing, removes unnecessary noise and invalid data, and provides a high-quality data basis for subsequent analysis, fault diagnosis and decision-making.

Description

Technical Field

[0001] The present invention relates to the field of operation and maintenance management technology, and in particular to an operation and maintenance management system and method based on industrial Internet of Things. Background Art

[0002] The Industrial Internet of Things (IIoT) is the product of the deep integration of current information technology and industrial manufacturing. It uses the core technologies of the Internet of Things, such as sensors, communication technology, cloud computing, and big data analysis, to achieve data interaction and collaboration between devices, thereby improving production efficiency, optimizing resource utilization, and realizing functions such as remote monitoring and predictive maintenance.

[0003] The operation and maintenance management system based on the Industrial Internet of Things is a system that uses the Internet of Things technology to optimize the operation and maintenance process of monitoring and management of industrial equipment and facilities. Thanks to the continuous development of the Industrial Internet of Things technology, the operation and maintenance management system based on the Industrial Internet of Things has become more intelligent, efficient and reliable. It not only monitors the status of equipment in real time through a variety of sensors, but also integrates cloud computing and big data analysis to analyze and predict equipment failures. At the same time, it can also respond to failures in a timely manner through system remote monitoring for operation and maintenance, greatly improving industrial production efficiency.

[0004] However, while the IIoT-based operation and maintenance management system is developing rapidly, it is also facing some challenges and problems: due to the lack of sufficient processing of the massive data generated by IIoT devices, inconsistent data formats and protocols between different devices and systems, and the existing data processing algorithms may not be able to quickly adapt to new data patterns or changes, resulting in poor data quality and unable to lay the foundation for subsequent analysis. At the same time, due to poor data quality and the high complexity of industrial equipment and systems, fault diagnosis is inaccurate, fault diagnosis and location response speed are slow, and further the accuracy of fault prediction is low, affecting operation and maintenance management decisions.

[0005] There is an urgent need for an enterprise operation and maintenance management system that can accurately diagnose and predict faults and accurately and quickly resolve faults. Summary of the invention

[0006] In order to solve the problems existing in the above background technology, the present invention adopts the following technical solutions:

[0007] On the one hand, the present application provides an operation and maintenance management system based on the industrial Internet of Things, including a real-time equipment monitoring module, a fault diagnosis and early warning module, a predictive maintenance module, a maintenance work order management module and a data analysis and reporting module wirelessly connected to a central controller; the real-time equipment monitoring module is used to collect equipment operation status data and intuitively display the real-time operation status of the equipment, including a real-time equipment data processing unit; the real-time equipment data processing unit is used to perform a first preprocessing on the collected real-time operation status data to generate effective real-time operation status data as a first preprocessing result; wherein the first preprocessing includes:

[0008] Acquire the collected real-time operating status data, which includes temperature parameters, pressure parameters, vibration parameters, electrical parameters and operating parameters.

[0009] Set time period thresholds for real-time operation status data. Arrange the real-time operation status data by timestamp through programming or data processing tools, and filter out data collected within the set effective working time period.

[0010] Perform data filtering to remove data outside the time period threshold in real-time operation status data.

[0011] The data filtered by time period are integrated to generate a first preprocessing result.

[0012] Preferably, the fault diagnosis and early warning module is used to analyze the equipment operation status data, warn of equipment abnormalities and faults, automatically diagnose the cause of the fault and provide a diagnosis report; wherein the analysis of the equipment operation status data specifically includes the following steps:

[0013] S100, obtaining effective real-time operation status data of the device;

[0014] S200, performing second preprocessing and feature extraction on the effective real-time operation status data;

[0015] S300 , classifying the effective real-time operating status data after the second preprocessing and feature extraction to generate a classification result data set.

[0016] Preferably, the second preprocessing comprises the following steps:

[0017] S201, detecting and removing abnormal values ​​in effective real-time operation status data;

[0018] S202, deduplication of the effective real-time operation status data after removing the abnormal values, and removing duplicate data;

[0019] S203, removing noise from the effective real-time operating status data after deduplication;

[0020] S204: Integrate and normalize the data processed in step S203 to generate a unified data format.

[0021] Preferably, the feature extraction includes:

[0022] Define the features to be extracted;

[0023] Acquire effective real-time operation status data after the second preprocessing;

[0024] Calculate and extract features, including calculating and extracting statistical features, frequency domain features and time series features;

[0025] Combine the extracted features into a feature vector.

[0026] Preferably, the step S300 uses a K-means algorithm to classify the effective real-time operating status data into three categories: normal, minor fault, and major fault, specifically including:

[0027] Determine the K value,

[0028] Use K-means algorithm to cluster the feature vector data after feature extraction.

[0029] Analyze the clustered data to generate a classification result data set.

[0030] Preferably, the warning device abnormality and failure, automatic diagnosis of the failure cause and provision of a diagnosis report include:

[0031] Fault diagnosis and early warning, including inputting the classification result data set into a pre-trained fault diagnosis and early warning model for fault diagnosis and early warning;

[0032] Perform fault root cause analysis and location based on fault diagnosis and early warning results, including fault classification, cause analysis, and correlation analysis;

[0033] Generate a fault diagnosis report based on the fault type and root cause analysis and location results.

[0034] Preferably, the predictive maintenance module is used to predict potential failures and maintenance requirements based on the historical failure data and equipment operation status data of the equipment, and formulate a maintenance plan in advance; specifically, it includes:

[0035] Analyze the historical failure data and historical operation status data of the equipment, establish data models, and identify failure modes and trends;

[0036] Based on the current operating status and historical data, machine learning algorithms are used to predict the failure probability of equipment, and clustering and association rule technologies are used to mine and identify maintenance needs to generate prediction results, which include failure probability and maintenance needs;

[0037] Based on the prediction results, a maintenance plan is automatically generated to remind maintenance personnel to perform inspections and maintenance;

[0038] The prediction result is compared with the actual fault data to evaluate the accuracy of the prediction, and the data classification rules in step S300 are verified and optimized based on the comparison result.

[0039] On the other hand, the present application also provides an operation and maintenance management method based on industrial Internet of Things, comprising the following steps:

[0040] S1. Real-time collection of equipment operation status data;

[0041] S2. Performing a first preprocessing on the equipment operation status data to generate effective real-time operation status data as a first preprocessing result; including:

[0042] Set time period thresholds for real-time operation status data.

[0043] Perform data filtering to remove data outside the time period threshold in real-time operation status data.

[0044] Integrate the data filtered by time period to generate a first preprocessing result;

[0045] S3, performing second preprocessing and feature extraction on the effective real-time operation status data;

[0046] S4, classifying the effective real-time operating status data after the second preprocessing and feature extraction, and generating a classification result data set;

[0047] S5. Input the classification result data set into a pre-trained fault diagnosis and prediction model for fault diagnosis and prediction.

[0048] Preferably, the method further includes: step S6, comparing and correcting the fault diagnosis and prediction results with the actual fault data, and verifying and optimizing the data classification rules in step S4.

[0049] Preferably, in step S5, during fault diagnosis and prediction, key features are extracted and matched to corresponding classification result data sets, and then analyzed and diagnosed through big data or artificial algorithms to output fault diagnosis and prediction results.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] The present invention ensures that the collected real-time operating status data meets actual work requirements through the first preprocessing, removes unnecessary noise and invalid data, and provides a high-quality data basis for subsequent analysis, fault diagnosis and decision-making.

[0052] The equipment real-time monitoring module of the present invention collects the equipment operation status data through these sensors, performs a first preprocessing, removes unnecessary noise and invalid data, and sets alarms, which can reduce the subsequent data pressure of the system, improve data quality, realize real-time monitoring of the equipment, timely discover potential faults, optimize maintenance strategies, and improve the reliability and operation efficiency of the equipment.

[0053] The present invention can effectively clean the real-time operation status data of the operation and maintenance management system by removing outliers, duplicate data and noise through the second preprocessing, thereby improving data quality and laying a foundation for subsequent data analysis and fault diagnosis; through feature extraction, effective statistical features and frequency domain features are extracted, which not only improves the availability of data, but also provides important information for subsequent modeling and fault diagnosis.

[0054] The present invention classifies effective real-time operation status data through the K-means algorithm, which can effectively identify the operation status of the equipment and help timely discover and deal with potential fault problems. This process not only improves the utilization rate of data, but also provides an important basis for the maintenance and management of equipment.

[0055] The present invention compares the prediction results with the actual fault data, verifies and optimizes the data classification rules according to the comparison results to adapt to the changes in the equipment operating status and new fault modes, thereby improving not only the data quality of the operation and maintenance management system, but also the operation and maintenance management efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0057] Figure 1 A schematic diagram of the structure of an operation and maintenance management system based on the industrial Internet of Things provided by the present invention;

[0058] Figure 2 A schematic diagram of a first preprocessing process provided by an embodiment of the present invention;

[0059] Figure 3 A schematic diagram of a process for analyzing device operating status data provided by an embodiment of the present invention;

[0060] Figure 4 A process diagram of a predictive maintenance module provided by an embodiment of the present invention;

[0061] Figure 5A schematic diagram of a flow chart of an operation and maintenance management method based on industrial Internet of Things provided by the present invention. DETAILED DESCRIPTION

[0062] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. It should be noted that relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0063] Embodiment 1

[0064] See also Figure 1 The present invention provides an operation and maintenance management system based on the industrial Internet of Things, including a real-time equipment monitoring module wirelessly connected to a central controller, a fault diagnosis and early warning module, a predictive maintenance module, a maintenance work order management module, and a data analysis and reporting module.

[0065] The real-time equipment monitoring module is used to collect equipment operation status data and intuitively display the real-time operation status of the equipment, so that users can understand the equipment operation status in real time; the fault diagnosis and early warning module is used to analyze equipment operation status data, timely warn equipment abnormalities and failures, automatically diagnose the cause of the failure and provide a diagnosis report; the predictive maintenance module is used to predict potential failures and maintenance needs based on the equipment's historical failure data and equipment operation status data, and formulate maintenance plans in advance; the maintenance work order management module is used to automatically generate work orders when equipment fails and assign them to maintenance personnel, supporting maintenance personnel to receive work orders, view information, and record processes and results through mobile devices to achieve full process management; the data analysis and reporting module is used to statistically analyze equipment operation, failure, maintenance and other data, generate a variety of reports, and provide data support for decision-making.

[0066] Among them, the equipment real-time monitoring module collects equipment operation status data by installing multiple sensors on key equipment. The multiple sensors include temperature sensors, pressure sensors, vibration sensors, etc., and are connected to the central controller through wireless communications (such as Wi-Fi, LoRa, NB-IoT, etc.). Equipment operation status data includes temperature parameters, pressure parameters, vibration parameters, electrical parameters and operation parameters. Temperature parameters include surface temperature and fluid temperature when the equipment is running; pressure parameters include operating pressure and vacuum pressure when the equipment is running; vibration parameters include acceleration and frequency when the equipment is running. Electrical parameters include current and voltage when the equipment is running. Operation parameters include speed and operation time when the equipment is running. Equipment operation status data can also include other parameters such as humidity and flow.

[0067] The device real-time monitoring module includes a device real-time data acquisition unit, a device real-time data processing unit, a device real-time data transmission unit and a device real-time data display unit. The device real-time data acquisition unit is used to collect the real-time operation status data of the device through a variety of sensors. The device real-time data processing unit is used to perform a first preprocessing on the collected real-time operation status data and generate effective real-time operation status data as a first preprocessing result. The device real-time data transmission unit is used to transmit the first preprocessing result to the cloud or local server in real time through the Internet of Things gateway. The device real-time data display unit is used to display the first preprocessing result of the device in real time on a user interface or dashboard.

[0068] In an embodiment of the present invention, Figure 2 As shown, the first preprocessing includes:

[0069] Acquire the collected real-time operating status data, including temperature parameters, pressure parameters, vibration parameters, electrical parameters, and operating parameters;

[0070] Set time period thresholds for real-time operation status data. For example, the time period threshold is set as an effective working time period, and the working time period includes daytime working time period (such as 08:00 - 18:00) and nighttime working time period (such as 18:00 -08:00). Determine the effective working time period based on the actual operation and operation requirements of the equipment. For example, for equipment that requires more manual participation, such as mechanical manufacturing equipment, construction equipment, and traffic management equipment, the daytime working time period is the effective working time period; for some industrial equipment, energy equipment, and cold chain equipment that have high energy consumption but high degree of automation, the nighttime working time period is the effective working time period. For example, for some equipment that requires intermittent operation and maintenance, such as equipment in sewage treatment plants and concrete pumps, the effective working time period includes the equipment operation time period during the day and at night.

[0071] Perform data filtering to remove data outside the time period threshold in the real-time operation status data. Use programming or data processing tools to arrange the real-time operation status data by timestamp and filter out data collected within the set effective working time period. Any data with a timestamp outside the effective working time period will be marked as invalid and excluded to ensure that subsequent analysis is based only on effective operation status data.

[0072] The data filtered by the time period are integrated to generate the first preprocessing result, which is the effective real-time operation status data of the equipment; the integration can be, for example, removing the blanks or missing values ​​formed after removing the timestamp from the real-time operation status data of each equipment, and integrating them into the time period data with the required format. The effective real-time operation status data of the equipment includes effective temperature parameters, effective pressure parameters, effective vibration parameters, effective electrical parameters, and effective operation parameters.

[0073] The present invention ensures that the collected real-time operating status data meets actual work requirements through the first preprocessing, removes unnecessary noise and invalid data, and provides a high-quality data basis for subsequent analysis, fault diagnosis and decision-making.

[0074] The device real-time monitoring module also includes an alarm unit, which is used to automatically generate an alarm and notify the user to handle or troubleshoot the abnormality when the first pre-processing result of the monitored device exceeds a preset alarm threshold. The alarm unit includes a threshold for setting the operating parameters of the device.

[0075] The equipment real-time monitoring module of the present invention collects the equipment operation status data through these sensors, performs a first preprocessing, removes unnecessary noise and invalid data, and sets alarms, which can reduce the subsequent data pressure of the system, improve data quality, realize real-time monitoring of the equipment, timely discover potential faults, optimize maintenance strategies, and improve the reliability and operation efficiency of the equipment.

[0076] Among them, the fault diagnosis and early warning module is used to analyze the equipment operation status data, timely warn of equipment abnormalities and failures, automatically diagnose the cause of the failure and provide a diagnosis report.

[0077] Analyze equipment operating status data, such as Figure 3 As shown, the specific steps include:

[0078] S100, obtaining effective real-time operating status data of the equipment, including effective temperature parameters, effective pressure parameters, effective vibration parameters, effective electrical parameters, and effective operating parameters.

[0079] Example:

[0080] Valid temperature parameters, such as temperature data when the device is running at night (for example, 18:00 - 23:00), are recorded as 25°C, 30°C, etc.

[0081] Effective pressure parameters, such as pressure data when the equipment is running at night, are recorded as 1.2 MPa, 1.0 MPa, etc.

[0082] Valid vibration parameters, such as vibration data when the device is running at night, are recorded as 0.5 g, 0.3 g, etc.

[0083] Effective electrical parameters, such as current data when the equipment is running at night, are recorded as 10 A, 12 A, etc.

[0084] Valid operating parameters, such as the speed data of the equipment when it is running at night, are recorded as 1500 RPM, 1600 RPM, etc.

[0085] S200: Perform second preprocessing and feature extraction on the effective real-time operating status data.

[0086] The second preprocessing of the effective real-time operation status data includes data cleaning of the effective real-time operation status data to remove abnormal values, duplicate data and noise of the effective real-time operation status data to ensure the system data quality.

[0087] In one embodiment of the present invention, removing outliers, duplicate data and noise from valid real-time operating status data specifically includes:

[0088] S201, detect and remove abnormal values ​​in the valid real-time operation status data, which are usually values ​​outside the expected range. For example, the valid temperature data is 23°C, 25°C, 24°C, 22°C, 100°C, 24°C; then 100°C here is obviously an abnormal value, and the abnormal value is directly removed after being detected. After removal, the remaining data is: 23°C, 25°C, 24°C, 22°C, 24°C.

[0089] In an embodiment of the present invention, detecting and removing abnormal values ​​in valid real-time operating status data includes:

[0090] Step one, divide the effective real-time operating status data into effective real-time operating status data subsets, specifically divide the data according to effective temperature parameters, effective pressure parameters, effective vibration parameters, effective electrical parameters and effective operating parameters, to form effective temperature parameter subsets, effective pressure parameter subsets, effective vibration parameter subsets, effective electrical parameter subsets and effective operating parameter subsets.

[0091] Step 2: Calculate the Z-score value of each data in each valid real-time operation status data subset and the IQR value of each valid real-time operation status data subset. The Z-score value is calculated using the Z-score method, and the IQR value is calculated using the interquartile range (IQR) method. Specifically:

[0092] The Z-score calculation includes:

[0093] Step 1: Calculate the value of each data x in each valid real-time running status data subset. i The mean and standard deviation std;

[0094] Step 2: Calculate the Z-score value of each data based on the mean and standard deviation. The specific formula is:

[0095] z_score i = (x i - mean i ) / std i , i represents the amount of data in each valid real-time running status data subset.

[0096] The IQR value calculation includes:

[0097] Step 1, sorting the data in each valid real-time operation status data subset;

[0098] Step 2, calculating the first quartile Q1 and the third quartile Q3 of each valid real-time operation status data subset according to the sorted data;

[0099] Step 3: Calculate the IQR value based on the first quartile Q1 and the third quartile Q3. The specific formula is: IQR=Q3−Q1.

[0100] Step 3: Identify and remove outliers based on the Z-score value of each data in each valid real-time operation status data subset and the IQR value of each valid real-time operation status data subset. Specifically include:

[0101] Step 1: Identify outliers according to the Z-score value of each data in each valid real-time running status data subset. Usually, data with a Z-score value greater than 3 or less than -3 is regarded as the first outlier.

[0102] Step 2: Identify outliers in the data subset according to the IQR value of each valid real-time operating status data subset. The threshold of the outlier is (Q1 - 1.5 *IQR, Q3 + 1.5 *IQR). Data outside this range is marked as the second outlier.

[0103] Step 3, compare the first abnormal value and the second abnormal value of each valid real-time operation status data subset. When the first abnormal value is the same as the second abnormal value, the data is determined to be abnormal and the abnormal data is removed. When the first abnormal value is different from the second abnormal value, steps 1 to 3 are performed again for secondary judgment. If the secondary judgment result is still different, the data is determined to be non-abnormal and does not need to be removed.

[0104] The present invention detects double abnormal data by using the Z-score method and the IQR method, and removes the abnormal data when both detection methods are abnormal, thereby further improving the quality of the data and ensuring the accuracy and reliability of subsequent analysis.

[0105] S202, de-duplicate the valid real-time operation status data after removing the abnormal values, and remove duplicate data. Duplicate data refers to the same data recorded multiple times at the same time point, which may be caused by multiple collections or data merging. De-duplicate data can be de-duplicated by timestamp and data content, specifically including:

[0106] First, remove the data with the same timestamp from the same valid real-time operating status data;

[0107] Secondly, check whether there are rows with exactly the same content in the valid real-time running status data, for example, by checking based on specific fields, etc. If so, remove the duplicate rows;

[0108] Then, the data after removing outliers is integrated to generate the latest or most relevant equipment operation status data record.

[0109] For example, the valid temperature parameter data before deduplication is: 23°C (08:00), 25°C (08:01), 25°C (08:01), 24°C (08:02). According to the timestamp, 25°C (08:01) is a duplicate data, so the duplicate timestamp data is directly removed. The data after removal is: 23°C (08:00), 25°C (08:01), 24°C (08:02).

[0110] S203, removing noise from the effective real-time operating status data after deduplication, where noise refers to random errors or unnecessary information in the data, which may be caused by sensor errors. In an embodiment of the present invention, commonly used denoising methods, such as moving average method and median filtering and other smoothing techniques, can be used to remove noise.

[0111] Example: Valid pressure data is:

[0112] 1.0 MPa, 1.1 MPa, 1.0 MPa, 1.2 MPa, 0.8 MPa, 1.3 MPa, of which 0.8 MPa may be noise;

[0113] Use a moving average method (e.g. 3-point moving average):

[0114] Calculate the new data with the moving average:

[0115] (1.0 + 1.1 + 1.0) / 3 = 1.03 MPa,

[0116] (1.1 + 1.0 + 1.2) / 3 = 1.10 MPa,

[0117] (1.0 + 1.2 + 0.8) / 3 = 1.00 MPa,

[0118] (1.2 + 0.8 + 1.3) / 3 = 1.13 MPa,

[0119] The smoothed data is:

[0120] 1.03 MPa, 1.10 MPa, 1.00 MPa, 1.13 MPa.

[0121] Preferably, in some embodiments of the present invention, noise removal also includes processing missing values, where missing values ​​are missing parts in a data set, such as spaces, etc., which may affect analysis results.

[0122] The present invention removes abnormal values, duplicate data and noise through the second preprocessing, which can effectively clean the real-time operation status data of the operation and maintenance management system, thereby improving data quality and laying a foundation for subsequent data analysis and fault diagnosis.

[0123] Preferably, the second preprocessing further includes integrating and normalizing the cleaned data to generate a unified data format.

[0124] Perform feature extraction on effective real-time operation status data, including extracting key features from effective real-time operation status data to ensure system data quality. Feature extraction specifically includes:

[0125] Step 1: define the features to be extracted. In an embodiment of the present invention, the features to be extracted include statistical features: mean, maximum, minimum, standard deviation, kurtosis and skewness; frequency domain features: spectrum analysis (such as Fourier transform), main frequency and harmonic frequency; time series features: autocorrelation coefficient, lag characteristics.

[0126] Step 2: Obtain the effective real-time operation status data after the second preprocessing. For example, the data is:

[0127] Valid temperature parameters: 23°C, 25°C, 24°C, 22°C, 24°C,

[0128] Effective pressure parameters: 1.0 MPa, 1.1 MPa, 1.0 MPa, 1.2 MPa, 1.1 MPa,

[0129] Effective vibration parameters: 0.3 g, 0.4 g, 0.3 g, 0.5 g, 0.4 g,

[0130] Valid electrical parameters: 10 A, 10 A, 11 A, 12 A, 11 A,

[0131] Valid operating parameters: 1500 RPM, 1500 RPM, 1600 RPM, 1550 RPM, 1500 RPM.

[0132] Step 3: Calculate and extract features, including calculating and extracting statistical features, frequency domain features and time series features. For example, the above data, taking effective vibration parameters as an example, after calculating and extracting features, is:

[0133] Extract statistical features of effective vibration parameters,

[0134] Average value: (0.3+0.4+0.3+0.5+0.4) / 5=0.38g,

[0135] Maximum value: 0.5g,

[0136] Minimum value: 0.3g,

[0137] Standard deviation: ≈0.08g,

[0138] The frequency domain features of effective vibration parameters are extracted, and the effective vibration parameters are subjected to Fourier transform to extract the frequency domain features.

[0139] Step 4: Combine the extracted features into a feature vector for subsequent analysis.

[0140] Example: Feature vector = [average temperature, maximum temperature, standard deviation temperature, average pressure, maximum pressure, standard deviation pressure, average vibration, maximum vibration, standard deviation vibration, average current, maximum current, standard deviation current, average speed, maximum speed, standard deviation speed]

[0141] The present invention extracts effective statistical features and frequency domain features through feature extraction, which not only improves the availability of data but also provides important information for subsequent modeling and fault diagnosis.

[0142] S300, classify the effective real-time operation status data after the second preprocessing and feature extraction, and generate a classification result data set. In an embodiment of the present invention, the K-means algorithm is used to classify the effective real-time operation status data into three categories: normal, minor fault and major fault. Specifically including:

[0143] Step 1: Determine the K value. In the K-means algorithm, the K value represents the number of categories to be divided. In the embodiment of the present invention, K is 3, which means that the data is divided into three categories: normal, minor fault, and major fault.

[0144] Step 2: Use the K-means algorithm to cluster the feature vector data after feature extraction, including:

[0145] Step 1: Initialize the cluster centers and randomly select K data points as initial cluster centers (centroids).

[0146] Step 2: Assign data points. For each data point, calculate its distance from each cluster center (usually using Euclidean distance) and assign the data point to the cluster center with the closest distance.

[0147] Step 3: Update the cluster center and calculate the mean of all data points in each cluster to update the cluster center.

[0148] Step 4: Repeat steps 2 and 3, and repeat the data point allocation and cluster center update until the cluster center no longer changes significantly (i.e., converges) or reaches a predetermined number of iterations.

[0149] Step 3: Analyze the clustered data to generate a classification result data set, including:

[0150] Normal: The characteristic values ​​of this type of data are within the expected range and the equipment is in good operating condition.

[0151] Minor fault: The characteristic values ​​of this type of data slightly deviate from the normal range and may require monitoring.

[0152] Serious failure: The characteristic values ​​of this type of data are obviously deviated from the normal range. The equipment may have serious problems and need to be dealt with immediately.

[0153] Example:

[0154] Normal: The equipment is operating well and all parameters are within normal ranges. For example, the temperature is 25°C, the pressure is 1.0MPa, and the vibration is 0.3 g.

[0155] Minor fault: The equipment has a minor abnormality and some parameters are close to critical values. For example, the temperature is 28°C, the pressure is 1.1 MPa, and the vibration is 0.6 g.

[0156] Serious fault: The equipment is in a serious abnormal state and the parameters are beyond the safe range. For example, the temperature is 35°C, the pressure is 1.5 MPa, and the vibration is 1.0 g.

[0157] After classification using the K-means algorithm, the results may be:

[0158] Normal: temperature 25°C, pressure 1.0 MPa, vibration 0.3 g,

[0159] Minor fault: temperature 28°C, pressure 1.1 MPa, vibration 0.6 g,

[0160] Serious failure: temperature 35°C, pressure 1.5 MPa, vibration 1.0 g.

[0161] The present invention classifies effective real-time operation status data through the K-means algorithm, which can effectively identify the operation status of the equipment and help timely discover and deal with potential fault problems. This process not only improves the utilization rate of data, but also provides an important basis for the maintenance and management of equipment.

[0162] Among them, timely warning of equipment abnormalities and failures, automatic diagnosis of failure causes and provision of diagnostic reports, including:

[0163] S400, fault diagnosis and early warning, including inputting the classification result data set into a pre-trained fault diagnosis and early warning model for fault diagnosis and early warning. The classification result data set is analyzed by the fault diagnosis and early warning model, abnormal patterns that are significantly different from normal behavior are identified, and fault alarms are issued in a timely manner. In an embodiment of the present invention, the fault diagnosis and early warning model can be performed using commonly used models, such as support vector machines (SVMs), random forests or neural networks, and gradient boosting machines (GBMs) based on machine learning. For example, models such as long short-term memory networks (LSTMs) and autoencoders based on deep learning.

[0164] S500, fault root cause analysis and location, including fault classification, cause analysis and association analysis. Classify the fault data obtained in the fault diagnosis and early warning steps to determine the fault type (such as mechanical fault, software fault, electrical fault, etc.); use statistical methods or machine learning models to identify the analyzed equipment operation status data, identify abnormal operation status data points, and use fault tree analysis (FTA) or machine learning models (such as decision trees, neural networks) to analyze the relationship between features and faults, infer the cause of the fault, and locate the specific equipment or operation status sequence (such as time series); use association rules to analyze the relationship between fault type, fault cause and equipment status, and further confirm the root cause of the fault.

[0165] S600. Generate a fault diagnosis report. Based on the fault type and root cause analysis and location results, generate a corresponding fault diagnosis report. The fault diagnosis report includes fault data, fault analysis process, fault cause and fault location, and also includes fault solutions. Among them, fault solutions include offline equipment processing and online system resource allocation. Offline equipment processing includes dispatching orders to maintenance personnel for offline equipment testing, restarting equipment, etc., and online system resource allocation includes adjusting configurations, expanding resources, etc.; at the same time, the generated fault diagnosis report is stored in the database, and historical cases of similar fault types are searched from the database for comparison and optimization, and the best successful solution is proposed; according to the best solution, an automated repair process is designed, and the system automatically performs repair operations according to the best solution.

[0166] The present invention realizes intelligent diagnosis and analysis of equipment failures through a systematic process of fault diagnosis and early warning module, combined with machine learning and data analysis technology. This process can not only quickly locate the root cause of the fault, but also provide effective guidance and suggestions, thereby improving the reliability and flexibility of the system. Through continuous learning and optimization, the module can adapt to the ever-changing environment and improve the efficiency of fault handling.

[0167] Among them, the predictive maintenance module is used to predict potential failures and maintenance needs based on the historical failure data and equipment operation status data of the equipment, and formulate maintenance plans in advance; Figure 4 As shown, specifically including:

[0168] Step 1: Analyze the historical failure data and historical operating status data of the equipment, establish a data model, and identify failure modes and trends.

[0169] Step 2: Based on the current operating status and historical data, apply machine learning algorithms (such as regression analysis, time series analysis, etc.) to predict the failure probability of the equipment, apply clustering and association rule technology to mine and identify maintenance needs, and generate prediction results. The prediction results include failure probability and maintenance needs.

[0170] Step 3: Based on the prediction results, a maintenance plan is automatically generated to remind maintenance personnel to perform necessary inspections and maintenance.

[0171] Preferably, the predictive maintenance module further includes:

[0172] Step 4: Compare the prediction results with the actual fault data to evaluate the accuracy of the prediction. At the same time, verify and optimize the data classification rules in step S300 based on the comparison results, including adjusting and optimizing the classification rules and algorithm parameters, etc., to adapt to changes in the equipment operating status and new fault modes.

[0173] The present invention compares the prediction results with the actual fault data, verifies and optimizes the data classification rules in step S300 according to the comparison results to adapt to changes in the equipment operating status and new fault modes, thereby not only improving the data quality of the operation and maintenance management system, but also improving the operation and maintenance management efficiency.

[0174] Among them, the maintenance work order management module is used to automatically generate work orders when equipment fails and assign them to maintenance personnel. It supports maintenance personnel to receive work orders, view information, and record processes and results through mobile devices to achieve full process management; specifically, it includes:

[0175] Work order generation: Once the system detects a new type of fault, it automatically generates a maintenance work order, including information such as fault type, location, priority, etc.

[0176] Work order allocation: The system automatically allocates work orders based on the maintenance personnel's skills and workload.

[0177] Work order processing: Maintenance personnel can receive work orders and view equipment-related information (including faults, operating status, etc.) through mobile application terminals, perform maintenance, and record the maintenance process and results after maintenance.

[0178] Feedback and closed loop: The system automatically feeds back the maintenance process and results to the fault diagnosis and early warning module and the predictive maintenance model to optimize subsequent diagnosis and prediction models.

[0179] The data analysis and reporting module is used to statistically analyze data on equipment operation, failure, maintenance, etc., generate a variety of reports, and provide data support for decision-making. Specifically, it includes:

[0180] Data summary: Regularly summarize data such as equipment operating status, fault records, maintenance records, etc.

[0181] Report generation: Generate various reports (such as operation efficiency report, fault analysis report, maintenance cost report, etc.) based on the summary data.

[0182] Decision support: Visual reports help management make data-driven decisions and improve operational efficiency.

[0183] Embodiment 2

[0184] The present invention also provides an operation and maintenance management method based on industrial Internet of Things, which is applied to the aforementioned operation and maintenance management system based on industrial Internet of Things. Specifically, Figure 5 As shown, the method comprises the following steps:

[0185] S1. Real-time collection of equipment operation status data. Equipment operation status data includes temperature parameters, pressure parameters, vibration parameters, electrical parameters and operation parameters. Temperature parameters include surface temperature and fluid temperature when the equipment is running; pressure parameters include operating pressure and vacuum pressure when the equipment is running; vibration parameters include acceleration and frequency when the equipment is running. Electrical parameters include current and voltage when the equipment is running. Operation parameters include speed and operation time when the equipment is running. Equipment operation status data can also include other parameters such as humidity and flow.

[0186] S2. Perform a first preprocessing on the device operation status data to generate effective real-time operation status data as a first preprocessing result. Specifically including:

[0187] Set time period thresholds for real-time operation status data. For example, the time period threshold is set as an effective working time period, and the working time period includes daytime working time period (such as 08:00 - 18:00) and nighttime working time period (such as 18:00 -08:00). Determine the effective working time period based on the actual operation and operation requirements of the equipment. For example, for equipment that requires more manual participation, such as mechanical manufacturing equipment, construction equipment, and traffic management equipment, the daytime working time period is the effective working time period; for some industrial equipment, energy equipment, and cold chain equipment that have high energy consumption but high degree of automation, the nighttime working time period is the effective working time period. For example, for some equipment that requires intermittent operation and maintenance, such as equipment in sewage treatment plants and concrete pumps, the effective working time period includes the equipment operation time period during the day and at night.

[0188] Perform data filtering to remove data outside the time period threshold in the real-time operation status data. Use programming or data processing tools to arrange the real-time operation status data by timestamp and filter out data collected within the set effective working time period. Any data with a timestamp outside the effective working time period will be marked as invalid and excluded to ensure that subsequent analysis is based only on effective operation status data.

[0189] The data filtered by the time period are integrated to generate the first preprocessing result, which is the effective real-time operation status data of the equipment; the integration can be, for example, removing the blanks or missing values ​​formed after removing the timestamp from the real-time operation status data of each equipment, and integrating them into the time period data with the required format. The effective real-time operation status data of the equipment includes effective temperature parameters, effective pressure parameters, effective vibration parameters, effective electrical parameters, and effective operation parameters.

[0190] The present invention ensures that the collected real-time operating status data meets actual work requirements through the first preprocessing, removes unnecessary noise and invalid data, and provides a high-quality data basis for subsequent analysis, fault diagnosis and decision-making.

[0191] S3. Perform a second preprocessing and feature extraction on the effective real-time operating status data. Among them, the second preprocessing of the effective real-time operating status data includes data cleaning of the effective real-time operating status data, removing outliers, duplicate data and noise from the effective real-time operating status data, and ensuring the quality of system data. The second preprocessing also includes integrating and normalizing the cleaned data to generate a unified data format. Among them, feature extraction is performed on the effective real-time operating status data, including extracting key features from the effective real-time operating status data to ensure the quality of system data. The specific second preprocessing and feature extraction process is the same as that in Example 1 and will not be repeated here.

[0192] S4, classify the effective real-time operation status data after the second preprocessing and feature extraction, and generate a classification result data set. In an embodiment of the present invention, the K-means algorithm is used to classify the effective real-time operation status data into three categories: normal, minor fault, and major fault. The specific data classification process is the same as that of the first embodiment, and will not be repeated here.

[0193] S5. Input the classification result data set into the pre-trained fault diagnosis and prediction model for fault diagnosis and prediction. During fault diagnosis and prediction, extract key features and match them to the corresponding classification result data set, then analyze and diagnose through big data or artificial algorithms, and output the fault diagnosis and prediction results.

[0194] Among them, fault diagnosis and prediction also include fault root cause analysis and location and generating a fault diagnosis report. The specific fault diagnosis and prediction process is the same as that of the first embodiment and will not be repeated here.

[0195] S6. Compare and correct the fault diagnosis and prediction results with the actual fault data to verify and optimize the data classification rules in step S4.

[0196] Through the above steps, the operation and maintenance management method based on the Industrial Internet of Things can achieve real-time monitoring, fault diagnosis and prediction of equipment operation status. This method ensures high quality and accuracy of data through multi-level data processing and analysis, thereby improving the reliability and maintenance efficiency of equipment operation and maintenance management.

[0197] The foregoing is merely a specific embodiment of the present invention, which enables those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features claimed herein.

Claims

1. An operation and maintenance management system based on industrial Internet of Things, characterized by: It includes a real-time equipment monitoring module, a fault diagnosis and early warning module, a predictive maintenance module, a maintenance work order management module, and a data analysis and reporting module connected to the central controller by wireless communication; The equipment real-time monitoring module is used to collect equipment operation status data and intuitively display the equipment real-time operation status, including the equipment real-time data processing unit; The equipment real-time data processing unit is used to perform a first preprocessing on the collected real-time operation status data to generate effective real-time operation status data as a first preprocessing result; wherein the first preprocessing includes: Acquire the collected real-time operating status data, which includes temperature parameters, pressure parameters, vibration parameters, electrical parameters and operating parameters. Set time period thresholds for real-time operation status data. Arrange the real-time operation status data by timestamp through programming or data processing tools, and filter out data collected within the set effective working time period. Perform data filtering to remove data outside the time period threshold in real-time operation status data. The data filtered by the time period are integrated to generate a first preprocessing result; The fault diagnosis and early warning module is used to analyze the equipment operation status data, warn of equipment abnormalities and faults, automatically diagnose the cause of the fault and provide a diagnosis report; wherein the analysis of the equipment operation status data specifically includes the following steps: S100, obtaining effective real-time operation status data of the device; S200, performing second preprocessing and feature extraction on the effective real-time operation status data; S300, classifying the effective real-time operating status data after the second preprocessing and feature extraction, and generating a classification result data set; The second preprocessing comprises the following steps: S201, detecting and removing abnormal values ​​in effective real-time operation status data; S202, deduplication of the effective real-time operation status data after removing the abnormal values, and removing duplicate data; S203, removing noise from the effective real-time operating status data after deduplication; S204, integrating and normalizing the data processed in step S203 to generate a unified data format; in, Step S201 includes: Step 1: dividing the effective real-time operation status data into effective real-time operation status data subsets, specifically dividing the data according to effective temperature parameters, effective pressure parameters, effective vibration parameters, effective electrical parameters and effective operation parameters to form effective temperature parameter subsets, effective pressure parameter subsets, effective vibration parameter subsets, effective electrical parameter subsets and effective operation parameter subsets; Step 2: Calculate the Z-score value of each data in each valid real-time operation status data subset and the IQR value of each valid real-time operation status data subset; wherein the Z-score value is calculated using the Z-score method, and the IQR value is calculated using the interquartile range (IQR) method; The Z-score calculation includes: Step 1: Calculate the value of each data x in each valid real-time operation status data subset. i The mean and standard deviation std; Step 2: Calculate the Z-score value of each data based on the mean and standard deviation. The specific formula is: z_score i = (x i - mean i ) / std i , i represents the amount of data in each valid real-time operation status data subset; The IQR value calculation includes: Step 1, sorting the data in each valid real-time operation status data subset; Step 2, calculating the first quartile Q1 and the third quartile Q3 of each valid real-time operation status data subset according to the sorted data; Step 3: Calculate the IQR value based on the first quartile Q1 and the third quartile Q3. The specific formula is: IQR=Q3−Q1; Step 3: Identify and remove outliers based on the Z-score value of each data in each valid real-time operation status data subset and the IQR value of each valid real-time operation status data subset; specifically, include: Step 1, identifying outliers according to the Z-score value of each data in each valid real-time running status data subset, and treating data with a Z-score value greater than 3 or less than -3 as the first outlier; Step 2: Identify outliers in the data subset according to the IQR value of each valid real-time operation status data subset. The threshold of the outlier is (Q1 - 1.5 *IQR, Q3 + 1.5 *IQR). Data outside this range is marked as the second outlier. Step 3, compare the first abnormal value and the second abnormal value of each valid real-time operation status data subset. When the first abnormal value is the same as the second abnormal value, the data is determined to be abnormal and the abnormal data is removed; when the first abnormal value is different from the second abnormal value, execute steps 1 to 3 again for a second judgment. If the results of the second judgment are still different, the data is determined to be non-abnormal and does not need to be removed.

2. According to claim 1, an operation and maintenance management system based on industrial Internet of Things is characterized by: The feature extraction comprises: Define the features to be extracted; Acquire effective real-time operation status data after the second preprocessing; Calculate and extract features, including calculating and extracting statistical features, frequency domain features and time series features; Combine the extracted features into a feature vector.

3. The operation and maintenance management system based on industrial Internet of Things according to claim 1 is characterized in that: The data classification in step S300 uses the K-means algorithm to classify the effective real-time operation status data into three categories: normal, minor fault and major fault, specifically including: Determine the K value, Use K-means algorithm to cluster the feature vector data after feature extraction. Analyze the clustered data to generate a classification result data set.

4. The operation and maintenance management system based on industrial Internet of Things according to claim 1 is characterized in that: The warning device abnormality and failure, automatic diagnosis of the failure cause and provision of a diagnosis report include: Fault diagnosis and early warning, including inputting the classification result data set into a pre-trained fault diagnosis and early warning model for fault diagnosis and early warning; Perform fault root cause analysis and location based on fault diagnosis and early warning results, including fault classification, cause analysis, and correlation analysis; Generate a fault diagnosis report based on the fault type and root cause analysis and location results.

5. The operation and maintenance management system based on industrial Internet of Things according to claim 1 is characterized in that: The predictive maintenance module is used to predict potential failures and maintenance needs based on the equipment's historical failure data and equipment operating status data, and to formulate maintenance plans in advance; Specifically include: Analyze the historical failure data and historical operation status data of the equipment, establish data models, and identify failure modes and trends; Based on the current operating status and historical data, machine learning algorithms are used to predict the failure probability of equipment, and clustering and association rule technologies are used to mine and identify maintenance needs to generate prediction results, which include failure probability and maintenance needs; Based on the prediction results, a maintenance plan is automatically generated to remind maintenance personnel to perform inspections and maintenance; The prediction result is compared with the actual fault data to evaluate the accuracy of the prediction, and the data classification rules in step S300 are verified and optimized based on the comparison result.

6. An operation and maintenance management method based on industrial Internet of Things, characterized by: Applied to an operation and maintenance management system based on industrial Internet of Things as claimed in any one of claims 2 to 5, the operation and maintenance management method based on industrial Internet of Things comprises the following steps: S1. Real-time collection of equipment operation status data; S2. Performing a first preprocessing on the device operation status data to generate effective real-time operation status data as a first preprocessing result; including: Set time period thresholds for real-time operation status data. Perform data filtering to remove data outside the time period threshold in real-time operation status data. Integrate the data filtered by time period to generate a first preprocessing result; S3, performing second preprocessing and feature extraction on the effective real-time operation status data; S4, classifying the effective real-time operating status data after the second preprocessing and feature extraction, and generating a classification result data set; S5. Input the classification result data set into a pre-trained fault diagnosis and prediction model for fault diagnosis and prediction.

7. The operation and maintenance management method based on industrial Internet of Things according to claim 6 is characterized in that: Also includes: Step S6: compare and calibrate the fault diagnosis and prediction results with the actual fault data to verify and optimize the data classification rules in step S4.

8. An operation and maintenance management method based on industrial Internet of Things according to claim 6 or 7, characterized in that: In step S5, during fault diagnosis and prediction, key features are extracted and matched to corresponding classification result data sets, and then analyzed and diagnosed through big data or artificial algorithms to output fault diagnosis and prediction results.

Images