Device operation and maintenance data management method and system

Abstract

The application discloses a kind of equipment operation and maintenance data management method and system, it is related to industrial equipment maintenance technical field, including the operation parameter of industrial equipment is formed acquisition data group;Acquisition data group is input support vector machine model, and output equipment abnormal probability value and equipment remaining life prediction value;According to equipment abnormal probability value, equipment state grade is generated, and maintenance priority matrix is established by the combination relationship of equipment state grade and equipment remaining life prediction value;According to maintenance priority matrix, select the equipment to be maintained, extract the characteristic parameter of acquisition data group, match the standard maintenance scheme in maintenance scheme library, and generate the customized maintenance scheme for the equipment to be maintained.The application combines data driving with intelligent algorithm, improves equipment operation and maintenance efficiency, effectively reduces maintenance cost, and prolongs the service life of equipment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial equipment maintenance, in particular to an equipment operation and maintenance data management method and system. BACKGROUND

[0002] The operation and maintenance management of industrial equipment has gradually changed from traditional passive repair and planned maintenance to intelligent and predictive maintenance. The existing technology mainly adopts a device state evaluation method based on threshold monitoring and a fixed cycle maintenance strategy, collects equipment operation data by deploying a sensor network, and formulates maintenance rules in combination with expert experience. At the same time, machine learning technology has made significant progress in equipment fault diagnosis and prediction, including support vector machines, deep learning algorithms, which are widely used in equipment health state evaluation and residual life prediction, providing data support for equipment maintenance decision-making.

[0003] However, the existing equipment operation and maintenance management technology still has the following shortcomings: first, the equipment state evaluation method relies too much on fixed thresholds and experience rules, and is difficult to adapt to the dynamic changes of equipment operating conditions, resulting in frequent false positives and false negatives; second, the maintenance decision lacks a scientific priority division mechanism, and fails to consider the equipment state level and residual life prediction results comprehensively, affecting the rational allocation of maintenance resources; third, the standard maintenance scheme is often difficult to meet the individual needs of different equipment, lacks a scheme optimization and adjustment mechanism based on historical data, and reduces the maintenance effect; finally, the collection, analysis and application process of equipment operation and maintenance data is fragmented, and a closed-loop data management system has not been formed, affecting the continuous improvement of maintenance efficiency.

[0004] In view of the above problems, the present application provides an equipment operation and maintenance data management method and system, which realizes equipment state evaluation and life prediction by establishing a support vector machine model, constructs a maintenance priority matrix to guide maintenance decision-making, and realizes individual customization of maintenance schemes based on a data-driven approach, forming a complete equipment operation and maintenance data management closed loop, effectively solving the technical problems of inaccurate equipment state evaluation, lack of scientific basis for maintenance decision-making, and low individualization degree of maintenance schemes in the prior art. SUMMARY

[0005] In view of the problems of existing equipment operation and maintenance data management technology, such as excessive dependence on fixed thresholds for equipment state evaluation, lack of scientific priority division mechanism for maintenance decision-making, insufficient individualization of maintenance schemes, and incomplete operation and maintenance data management system, the present application is proposed.

[0006] Therefore, the problem to be solved by the present application is how to realize accurate prediction of equipment state and life through a support vector machine model, establish a maintenance priority evaluation mechanism, and realize intelligent customization of maintenance schemes, and finally form a closed-loop equipment operation and maintenance data management system.

[0007] To solve the above technical problems, the present application provides the following technical solutions:

[0008] In a first aspect, the embodiments of the present application provide a device operation and maintenance data management method, which comprises: collecting operation parameters of an industrial device to form a collection data set; inputting the collection data set into a support vector machine model to output a device anomaly probability value and a device remaining life prediction value; generating a device state level according to the device anomaly probability value, and establishing a maintenance priority matrix through a combination relationship of the device state level and the device remaining life prediction value; selecting a device to be maintained according to the maintenance priority matrix, extracting feature parameters of the collection data set, matching a standard maintenance scheme in a maintenance scheme library, and generating a customized maintenance scheme for the device to be maintained.

[0009] As a preferred scheme of the device operation and maintenance data management method of the present application, the generation method of the customized maintenance scheme is as follows: selecting a device to be maintained according to a priority level order of the maintenance priority matrix, and preferentially processing a high-priority device; extracting feature parameters from a collection data set of the device to be maintained, and performing feature matching on the feature parameters and a standard maintenance scheme in a maintenance scheme library, wherein the maintenance scheme library adopts a hierarchical structure; based on a matching result, determining an optimal matching standard maintenance scheme through feature similarity calculation, and combining operation history records, maintenance history records and spare part replacement records of the device to be maintained to perform individual adjustment on the standard maintenance scheme, thereby forming the customized maintenance scheme for the device to be maintained.

[0010] As a preferred scheme of the device operation and maintenance data management method of the present application, the generation of the device state level according to the device anomaly probability value and the establishment of the maintenance priority matrix through a combination relationship of the device state level and the device remaining life prediction value comprise: generating the device state level according to the device anomaly probability value, and dividing a device operation state into a normal operation state, a sub-healthy state and a fault warning state through a preset threshold division method; at the same time, dividing the device remaining life prediction value into an abundant life interval, a transition life interval and an emergency life interval according to a life consumption degree; establishing the maintenance priority matrix according to an interval combination of the device state level and the device remaining life prediction value, wherein the maintenance priority matrix adopts a three-dimensional structure; determining a maintenance priority level through cross mapping of the state level and the life interval, and adjusting the priority level through a weight coefficient between different dimensional combinations.

[0011] As a preferred scheme of the equipment operation and maintenance data management method, when the equipment abnormal probability value is less than the first threshold value, the equipment operation state is divided into a normal operation state; if the equipment is in the normal operation state and belongs to the sufficient life interval, a low-priority maintenance task is determined, and the maintenance task urgency is defined as routine maintenance; when the equipment abnormal probability value is greater than the first threshold value and less than the second threshold value, the equipment operation state is divided into a sub-health state; if the equipment is in the sub-health state and belongs to the transition life interval, a medium-priority maintenance task is determined, and the maintenance task urgency is defined as planned processing; when the equipment abnormal probability value is greater than the second threshold value, the equipment operation state is divided into a fault warning state; if the equipment is in the fault warning state and belongs to the emergency life interval, a high-priority maintenance task is determined, and the maintenance task urgency is defined as immediate processing; when the equipment remaining life prediction value is greater than the life threshold value A, it belongs to the sufficient life interval; when the equipment remaining life prediction value is less than the life threshold value A and greater than the life threshold value B, it belongs to the transition life interval; when the equipment remaining life prediction value is less than the life threshold value B, it belongs to the emergency life interval; when the combination of the equipment state level and the remaining life interval does not match, the priority is adjusted through a weight coefficient.

[0012] As a preferred scheme of the equipment operation and maintenance data management method, the collected data set is input into a support vector machine model to output an equipment abnormal probability value and an equipment remaining life prediction value, including: inputting the standardized collected data set into a support vector machine model according to a time sequence structure for analysis and prediction, wherein the support vector machine model includes a state diagnosis model and a life prediction model; the state diagnosis model adopts a radial basis kernel function to construct a feature space mapping relationship, and performs nonlinear classification on sample data in the feature space by minimizing a structural risk criterion, and calculates the equipment abnormal probability value according to a distance function of a classification hyperplane; the life prediction model adopts a support vector regression structure, selects an epsilon-insensitive loss function to construct an optimization objective, extracts temperature trend features, pressure fluctuation features and vibration spectrum features in the collected data set to establish a regression prediction model, and trains the regression prediction model in combination with a life attenuation law in historical operation data to output the equipment remaining life prediction value.

[0013] As a preferred scheme of the equipment operation and maintenance data management method, the specific formula of the equipment abnormal probability value is as follows:

[0014] ;

[0015] wherein, is the equipment abnormal probability value, is an input feature vector, is a support vector quantity, is a Lagrange multiplier, is a class label of the i th support vector, is an input feature vector of the i th support vector, is an RBF kernel function parameter, is a probability conversion coefficient, is a mean value of the feature vector.

[0016] The specific formula of the equipment remaining life prediction value is as follows:

[0017] ;

[0018] wherein, is an equipment remaining life prediction value, is an equipment nominal life, is a number of degradation features, is a weight coefficient of the i th feature, is a measured value of the i th feature, is an ideal value of the i th feature, is a life attenuation coefficient, is a current value of a key environmental factor, is a standard value of the environmental factor, is a tolerance coefficient of the environmental factor.

[0019] As a preferred scheme of the equipment operation and maintenance data management method, the forming method of the collection data group is that the temperature data, the pressure data and the vibration data are time-aligned according to the collection time stamps of the operation parameters, and the aligned operation parameters are divided into a plurality of collection data groups according to a preset time interval; the temperature data, the pressure data and the vibration data in the collection data group are respectively preprocessed to generate a standardized collection data group.

[0020] In a second aspect, an equipment operation and maintenance data management system is provided, which comprises: a collection module configured to collect operation parameters of an industrial equipment to form a collection data group; an output module configured to input the collection data group into a support vector machine model to output an equipment abnormality probability value and an equipment remaining life prediction value; an establishment module configured to generate an equipment state level according to the equipment abnormality probability value, and establish a maintenance priority matrix through a combination relationship of the equipment state level and the equipment remaining life prediction value; and a generation module configured to select a to-be-maintained equipment according to the maintenance priority matrix, extract feature parameters of the collection data group, match a standard maintenance scheme in a maintenance scheme library, and generate a customized maintenance scheme for the to-be-maintained equipment.

[0021] In a third aspect, an embodiment of the present application provides a computer device, comprising a memory and a processor, and the memory stores a computer program, wherein the computer program instructs the processor to implement the steps of the device operation and maintenance data management method according to the first aspect of the present application.

[0022] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, which stores a computer program, wherein the computer program instructs a processor to implement the steps of the device operation and maintenance data management method according to the first aspect of the present application.

[0023] The present application has the following beneficial effects: by collecting the operation parameters of the industrial equipment to form a standardized collection data set, and inputting the collection data set into a support vector machine model for analysis, the dual prediction of the equipment abnormal state and the remaining life is realized; by combining the equipment state level and the predicted value of the remaining life to establish a three-dimensional structure of the maintenance priority matrix, the accurate grading of the equipment maintenance task is realized; finally, the equipment to be maintained is selected based on the maintenance priority matrix, and the standard maintenance scheme is individually adjusted in combination with the equipment historical record, so that a targeted customized maintenance scheme is formed, the present application combines data driving and intelligent algorithm, improves the equipment operation and maintenance efficiency, effectively reduces the maintenance cost, and prolongs the service life of the equipment. BRIEF DESCRIPTION OF DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor. Among them:

[0025] Fig. 1 The flowchart of the device operation and maintenance data management method of embodiment 1.

[0026] Fig. 2 The system block diagram of the device operation and maintenance data management method of embodiment 1. DETAILED DESCRIPTION

[0027] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail in combination with the drawings of the specification.

[0028] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, but the present application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited to the specific embodiments disclosed below.

[0029] Second, the "one embodiment" or "an embodiment" referred to herein means a particular feature, structure, or characteristic including an implementation that can be included in at least one implementation of the application. The appearances of "in one embodiment" or "in an embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily all referring to a single, alternative embodiment, or a single, alternative implementation.

[0030] Embodiment 1

[0031] With reference to Figs. 1-2 For the first embodiment of the application, the embodiment provides a device operation and maintenance data management method, comprising,

[0032] S1: Collecting the running parameters of the industrial equipment to form a collection data set.

[0033] Specifically, the industrial equipment includes processing equipment, conveying equipment and detection equipment on the production line, wherein the running parameters include temperature data, pressure data and vibration data; the temperature data is collected by an infrared temperature measurement sensor according to a sampling period of 500 ms to obtain the temperature values of each monitoring point of the equipment; the pressure data is collected by a pressure sensor according to a sampling period of 200 ms to obtain the pressure values of each pressure measurement point of the equipment; and the vibration data is collected by an acceleration sensor according to a sampling period of 100 ms to obtain the vibration acceleration values and vibration displacement values of each vibration measurement point of the equipment.

[0034] Further, the temperature data, the pressure data and the vibration data are time-aligned according to the collection time stamps of the running parameters, and the aligned running parameters are divided into a plurality of collection data sets according to a preset time interval.

[0035] Further, the temperature data, the pressure data and the vibration data in each collection data set are respectively preprocessed to generate standardized collection data sets, wherein the preprocessing includes removing abnormal mutation values, supplementing missing values and data normalization processing.

[0036] S2: Inputting the collection data set into a support vector machine model to output an equipment abnormal probability value and an equipment remaining life prediction value.

[0037] Specifically, the standardized collection data set is inputted into the support vector machine model according to the time sequence structure for analysis and prediction, wherein the support vector machine model includes a state diagnosis model and a life prediction model.

[0038] Further, the state diagnosis model adopts a radial basis kernel function to construct a feature space mapping relationship, and performs nonlinear classification on the sample data in the feature space by minimizing the structural risk criterion, and calculates the equipment abnormal probability value based on the distance function of the classification hyperplane, and the specific formula is as follows:

[0039] ;

[0040] wherein, is the device abnormality probability value, is the input feature vector, is the number of support vectors, is the Lagrange multiplier, is the class label of the ith support vector, is the input feature vector of the ith support vector, is the RBF kernel function parameter, is the probability conversion coefficient, is the mean of the feature vector.

[0041] It should be noted that the kernel function parameter and the slack variable of the radial basis kernel function are obtained by grid search and cross-validation method.

[0042] Further, the life prediction model adopts a support vector regression structure, selects an epsilon-insensitive loss function to construct an optimization objective, extracts temperature trend features, pressure fluctuation features and vibration spectrum features in the collected data set to establish a regression prediction model, and combines the life attenuation law in the historical operation data to train the regression prediction model, and outputs a device remaining life prediction value, and the specific formula is as follows:

[0043] ;

[0044] wherein, is the device remaining life prediction value, is the device nominal life, is the number of degradation features, is the weight coefficient of the ith feature, is the measured value of the ith feature, is the ideal value of the ith feature, is the life attenuation coefficient, is the current value of the key environmental factor, is the standard value of the environmental factor, is the tolerance coefficient of the environmental factor.

[0045] It should be noted that the insensitive coefficient is determined by the grid search method.

[0046] S3: generating a device state level according to the device abnormality probability value, and establishing a maintenance priority matrix through a combination relationship of the device state level and the device remaining life prediction value.

[0047] Specifically, a device state level is generated according to the device abnormality probability value, and a device operating state is divided into a normal operating state, a sub-healthy state and a fault warning state through a preset threshold division method.

[0048] Further, when the device abnormality probability value is less than the first threshold value, then the device running state is classified as a normal running state; if the device is in the normal running state and belongs to the abundant life interval, then a low-priority maintenance task is determined, and the maintenance task urgency is defined as routine maintenance; when the device abnormality probability value is greater than the first threshold value and less than the second threshold value, then the device running state is classified as a sub-health state; if the device is in the sub-health state and belongs to the transition life interval, then a medium-priority maintenance task is determined, and the maintenance task urgency is defined as planned processing; when the device abnormality probability value is greater than the second threshold value, then the device running state is classified as a fault warning state; if the device is in the fault warning state and belongs to the emergency life interval, then a high-priority maintenance task is determined, and the maintenance task urgency is defined as immediate processing.

[0049] It should be noted that the first threshold value is determined by analyzing the abnormality probability distribution of the device in the normal running state, and combining statistical analysis of the normal fault frequency in the historical data; and the second threshold value is determined by transition analysis between the sub-health state and the fault warning state of the device.

[0050] Further, the device remaining life prediction value is further divided into the abundant life interval, the transition life interval and the emergency life interval according to the life consumption degree.

[0051] Specifically, when the device remaining life prediction value is greater than the life threshold value A, then it belongs to the abundant life interval; when the device remaining life prediction value is less than the life threshold value A and greater than the life threshold value B, then it belongs to the transition life interval; and when the device remaining life prediction value is less than the life threshold value B, then it belongs to the emergency life interval.

[0052] It should be noted that the life threshold value A is determined by analyzing the average life data of the device in the normal running state, the historical fault occurrence frequency and the health state estimation of the device; and the life threshold value B is determined by comprehensively considering the fault risk of the device and the decline trend of the remaining life.

[0053] Further, a maintenance priority matrix is established according to the interval combination of the device state level and the device remaining life prediction value, wherein the maintenance priority matrix adopts a three-dimensional structure.

[0054] It should be noted that the horizontal dimension represents the device state level, the vertical dimension represents the remaining life interval, and the depth dimension represents the urgency of the maintenance task.

[0055] Further, the maintenance priority level is determined by cross-mapping of the state level and the life interval, and the priority is adjusted between different dimension combinations by a weight coefficient.

[0056] Specifically, when the combination of the equipment state level and the remaining life interval does not match, priority adjustment is performed through a weight coefficient; if a higher state level but sufficient life appears, the state level is given priority; if a normal state but urgent life appears, the remaining life is given priority.

[0057] It should be noted that the normal operation state corresponds to a low priority in the sufficient life interval, the sub-health state corresponds to a medium priority in the transition life interval, and the fault warning state corresponds to a high priority in the urgent life interval.

[0058] S4: Selecting the equipment to be maintained according to the maintenance priority matrix, extracting the characteristic parameters of the collected data set, matching the standard maintenance scheme in the maintenance scheme library, and generating a customized maintenance scheme for the equipment to be maintained.

[0059] Specifically, the equipment to be maintained is selected according to the priority level order of the maintenance priority matrix, and high-priority equipment is processed first; equipment with the same priority is sorted according to the abnormal probability value.

[0060] Further, the characteristic parameters are extracted from the collected data set of the equipment to be maintained, and the characteristic parameters are matched with the standard maintenance scheme in the maintenance scheme library, wherein the maintenance scheme library adopts a hierarchical structure.

[0061] It should be noted that the characteristic parameters include temperature characteristic parameters, pressure characteristic parameters, and vibration characteristic parameters, the temperature characteristic parameters include temperature mean value, temperature fluctuation rate, and temperature gradient change trend; the pressure characteristic parameters include pressure peak value, pressure pulsation frequency, and pressure fluctuation range; the vibration characteristic parameters include amplitude spectrum, phase spectrum, and envelope spectrum characteristics; the hierarchical structure includes equipment type layer, fault mode layer, and maintenance measure layer.

[0062] Further, based on the matching result, the optimal matching standard maintenance scheme is determined through characteristic similarity calculation, and the standard maintenance scheme is personalized adjusted combined with the operation history record, maintenance history record, and spare part replacement record of the equipment to be maintained, to form a customized maintenance scheme for the equipment to be maintained.

[0063] It should be noted that the personalized adjustment includes maintenance cycle adjustment, maintenance project optimization, and maintenance resource allocation, and the customized maintenance scheme includes maintenance time arrangement, maintenance project list, spare part demand list, personnel skill requirement, and maintenance quality standard.

[0064] Further, the embodiment also provides a wireless network high reliability transmission system based on dual-link redundancy, comprising: an acquisition module configured to acquire operation parameters of an industrial device to form an acquisition data set; an output module configured to input the acquisition data set into a support vector machine model to output a device anomaly probability value and a device remaining life prediction value; an establishment module configured to generate a device state level according to the device anomaly probability value, and establish a maintenance priority matrix through a combination relationship of the device state level and the device remaining life prediction value; and a generation module configured to select a device to be maintained according to the maintenance priority matrix, extract a characteristic parameter of the acquisition data set, match a standard maintenance scheme in a maintenance scheme library, and generate a customized maintenance scheme for the device to be maintained.

[0065] The embodiment also provides a computer device suitable for the device operation and maintenance data management method, comprising a memory and a processor; the memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions to implement the device operation and maintenance data management method proposed in the above embodiment.

[0066] The computer device can be a terminal, and the computer device comprises a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The communication interface of the computer device is configured to perform wired or wireless communication with external terminals. The wireless communication can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device. The input device of the computer device can also be an external keyboard, touchpad or mouse, etc.

[0067] To sum up, the present application forms a standardized acquisition data set by acquiring operation parameters of an industrial device, and inputs the acquisition data set into a support vector machine model for analysis, thereby realizing double prediction of device abnormal state and remaining life. A three-dimensional maintenance priority matrix is established through a combination relationship of device state level and remaining life prediction value, thereby realizing accurate grading of device maintenance tasks. Finally, a device to be maintained is selected based on the maintenance priority matrix, and a standard maintenance scheme is personalized adjusted in combination with device historical records, thereby forming a targeted customized maintenance scheme. The present application combines data driving and intelligent algorithms, thereby improving device operation and maintenance efficiency, effectively reducing maintenance cost, and prolonging device service life.

[0068] Embodiment 2

[0069] Referring to Table 1, for the second embodiment of the present application, the embodiment provides a device operation and maintenance data management method. In order to verify the beneficial effects of the present application, economic benefit calculation and simulation experiments are carried out for scientific demonstration.

[0070] Specifically, 6 key pump devices of a certain chemical plant are selected as experimental objects, and a 3-month operation and maintenance management experiment is carried out. These pump devices are of the same type, all being centrifugal pumps of a certain brand, with a nominal service life of 50,000 hours, but the time of use is different, and the running conditions are different. Before the experiment, a temperature sensor (precision ±0.1℃), a pressure sensor (precision ±0.01MPa) and a vibration sensor (sampling frequency 10kHz) are installed on each pump device. The sampling interval is set to 10 minutes. Industrial-grade data collectors are used to collect device operating parameters, and data is transmitted in real time to a data processing server through an OPC protocol.

[0071] Further, in the data preprocessing stage, the collected temperature, pressure and vibration data are time-aligned, outliers and missing values are removed, and standardized. Standardization uses the Z-score method to convert the data to a standard normal distribution with a mean of 0 and a standard deviation of 1. For vibration data, additional frequency spectrum analysis is performed to extract frequency spectrum features in the range of 0-1000Hz.

[0072] Further, the training of the support vector machine model uses historical operating data, including normal operating data, device failure data and maintenance records within 2 years. The state diagnosis model uses an RBF kernel function, with a kernel parameter γ set to 0.1 and a relaxation variable C set to 1.0. The model parameters are optimized through 5-fold cross-validation. The life prediction model uses an ε-SVR structure, with ε set to 0.1, and the optimal parameter combination is determined using a grid search method.

[0073] Specifically, two threshold values are used for device state level division: the first threshold value is set to 0.3, and the second threshold value is set to 0.7. The division of the remaining life interval uses the following standards: life threshold A is 50% of the nominal life of the device, and life threshold B is 20% of the nominal life of the device. The maintenance priority matrix uses a 3x3x3 three-dimensional structure, which determines the priority by cross-mapping the device state level and the remaining life interval, and introduces a weight coefficient for dynamic adjustment. The maintenance scheme library uses a three-layer structure: the first layer is the device type, the second layer is the fault type, and the third layer is the specific maintenance scheme. Feature matching uses a cosine similarity algorithm, with a similarity threshold set to 0.8. When generating a customized maintenance scheme, the running history record, maintenance history record and spare parts replacement record of the device are considered to adjust the standard maintenance scheme accordingly.

[0074] Further, as shown in Table 1, from the distribution of anomaly probability values, the method can accurately identify devices in different operating states. For example, the anomaly probability value of Pump-003 is 0.82, which is significantly higher than the second threshold value 0.7, and is correctly classified as a fault warning state; while the anomaly probability values of Pump-001 and Pump-004 are 0.25 and 0.28 respectively, which are lower than the first threshold value 0.3, indicating that the two devices are in good operating condition. This accurate state identification capability is due to the radial basis kernel function used by the support vector machine model, which can effectively capture the nonlinear characteristics of device operating data.

[0075] Table 1 Comparison of device performance and maintenance indicators

[0076]

[0077] Further, in terms of remaining life prediction, the method shows high accuracy. Taking Pump-005 as an example, the device has been running for 41,000 hours, close to 82% of the nominal life (50,000 hours), and the method predicts its remaining life as 4,800 hours, which is highly consistent with the actual operating condition. By comparing the data of Pump-002 and Pump-005, it can be found that although both are in sub-healthy state, due to the difference in remaining life prediction value (8500 hours and 4800 hours respectively), the system allocates a higher maintenance priority to Pump-005, reflecting the rationality of the method in maintenance decision-making.

[0078] Specifically, the maintenance scheme matching degree reflects the degree of fit between the customized maintenance scheme and the actual needs of the device. The data shows that the maintenance scheme matching degree of all devices exceeds 88%, among which Pump-003 in the fault warning state reaches a high matching degree of 94.6%, which shows that the method can generate highly personalized maintenance schemes according to the specific conditions of the device. The fault prevention rate indicator generally maintains a high level of more than 90%, proving that the method has a significant effect in preventive maintenance.

[0079] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present application, which should be covered by the scope of the claims of the present application.

Claims

1. An equipment operation and maintenance data management method, characterized in that: Comprising, Collecting operation parameters of industrial equipment to form a collection data set; Inputting the collection data set into a support vector machine model to output an equipment abnormality probability value and an equipment remaining life prediction value; Generating an equipment state level according to the equipment abnormality probability value, and establishing a maintenance priority matrix through a combined relationship of the equipment state level and the equipment remaining life prediction value; Selecting a to-be-maintained equipment according to the maintenance priority matrix, extracting a characteristic parameter of the collection data set, matching a standard maintenance scheme in a maintenance scheme library, and generating a customized maintenance scheme for the to-be-maintained equipment; Inputting the standardized collection data set into a support vector machine model according to a time sequence structure for analysis and prediction, wherein the support vector machine model comprises a state diagnosis model and a life prediction model; The state diagnosis model adopts a radial basis kernel function to construct a feature space mapping relationship, and performs nonlinear classification on sample data in the feature space by minimizing a structural risk criterion, and calculates an equipment abnormality probability value according to a distance function of a classification hyperplane; The life prediction model adopts a support vector regression structure, selects an epsilon-insensitive loss function to construct an optimization objective, establishes a regression prediction model by extracting a temperature trend feature, a pressure fluctuation feature and a vibration spectrum feature in the collection data set, and trains the regression prediction model in combination with a life attenuation law in historical operation data to output an equipment remaining life prediction value; A specific formula of the equipment abnormality probability value is as follows: ； wherein, is a device anomaly probability value, is an input feature vector, is a number of support vectors, is a Lagrange multiplier, is a class label of the i-th support vector, is an input feature vector of the i-th support vector, is an RBF kernel function parameter, is a probability conversion coefficient, is a mean of the feature vector; A specific formula of the equipment remaining life prediction value is as follows: ； wherein, is a predicted value of the remaining life of the device, is a nominal life of the device, is a number of degradation features, is a weight coefficient of the ith feature, is a measured value of the ith feature, is an ideal value of the ith feature, is a life decay coefficient, is a current value of a key environmental factor, is a standard value of the environmental factor, is a tolerance coefficient of the environmental factor.

2. The device operation and maintenance data management method of claim 1, wherein: A generation method of the customized maintenance scheme is as follows: Selecting a to-be-maintained equipment according to a priority level order of the maintenance priority matrix, and preferentially processing a high-priority equipment; Extracting a characteristic parameter from a collection data set of the to-be-maintained equipment, and performing characteristic matching on the characteristic parameter and a standard maintenance scheme in a maintenance scheme library, wherein the maintenance scheme library adopts a hierarchical structure; Based on a matching result, determining an optimal matching standard maintenance scheme through characteristic similarity calculation, and forming a customized maintenance scheme for the to-be-maintained equipment by individually adjusting the standard maintenance scheme in combination with operation history records, maintenance history records and spare part replacement records of the to-be-maintained equipment.

3. The device operation and maintenance data management method of claim 2, wherein: Generating an equipment state level according to an equipment abnormality probability value, and establishing a maintenance priority matrix through a combined relationship of the equipment state level and an equipment remaining life prediction value, comprising: Generating an equipment state level according to an equipment abnormality probability value, and dividing an equipment operation state into a normal operation state, a sub-healthy state and a fault warning state through a preset threshold division method; Meanwhile, dividing the equipment remaining life prediction value into an abundant life interval, a transition life interval and an urgent life interval according to a life consumption degree; Establishing a maintenance priority matrix according to an interval combination of the equipment state level and the equipment remaining life prediction value, wherein the maintenance priority matrix adopts a three-dimensional structure; Determining a maintenance priority level through cross mapping of the state level and the life interval, and adjusting the priority level through a weight coefficient between different dimensional combinations.

4. The device operation data management method of claim 3, wherein: Further comprising, When the device anomaly probability value is less than the first threshold value, the device running state is classified as a normal running state; if the device is in the normal running state and belongs to the sufficient life interval, a low-priority maintenance task is determined, and the maintenance task urgency is defined as routine maintenance; When the device anomaly probability value is greater than the first threshold value and less than the second threshold value, the device running state is classified as a sub-health state; if the device is in the sub-health state and belongs to the transition life interval, a medium-priority maintenance task is determined, and the maintenance task urgency is defined as planned processing; When the device anomaly probability value is greater than the second threshold value, the device running state is classified as a fault warning state; If the device is in the fault warning state and belongs to the emergency life interval, a high-priority maintenance task is determined, and the maintenance task urgency is defined as immediate processing; When the device remaining life prediction value is greater than the life threshold value A, it belongs to the sufficient life interval; when the device remaining life prediction value is less than the life threshold value A and greater than the life threshold value B, it belongs to the transition life interval; when the device remaining life prediction value is less than the life threshold value B, it belongs to the emergency life interval; When the combination of the device state level and the remaining life interval does not match, the priority is adjusted through a weight coefficient.

5. The device operation data management method of claim 1, wherein: The forming method of the collection data set is, The temperature data, pressure data and vibration data are time-aligned according to the collection time stamps of the running parameters, and the aligned running parameters are divided into a plurality of collection data sets according to a preset time interval; The temperature data, pressure data and vibration data in the collection data set are respectively pre-processed to generate a standardized collection data set.

6. An equipment operation data management system based on the equipment operation data management method according to any one of claims 1 to 5, characterized by: It comprises, A collection module for collecting running parameters of an industrial device to form a collection data set; An output module for inputting the collection data set into a support vector machine model to output a device anomaly probability value and a device remaining life prediction value; An establishment module for generating a device state level according to the device anomaly probability value, and establishing a maintenance priority matrix through the combination of the device state level and the device remaining life prediction value; A generation module for selecting a device to be maintained according to the maintenance priority matrix, extracting feature parameters of the collection data set, matching a standard maintenance scheme in a maintenance scheme library, and generating a customized maintenance scheme for the device to be maintained.

7. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that: The processor executes the computer program to realize the steps of the device operation and maintenance data management method of any one of claims 1-5.

8. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to realize the steps of the device operation and maintenance data management method of any one of claims 1-5.

Images