Intelligent operation and maintenance method and system based on big data analysis

Abstract

The invention discloses an intelligent operation and maintenance method and system based on big data analysis, and the method comprises the steps: obtaining a historical operation and maintenance log which comprises the historical operation and maintenance log of equipment; determining P devices influencing operation and maintenance duration change according to the historical operation and maintenance log, and predicting device operation and maintenance duration prediction values of the P devices in a second preset time period according to device operation and maintenance duration sequence data of the P devices in the first preset time period in the historical operation and maintenance log; determining abnormal fluctuation characteristics according to the predicted equipment operation and maintenance duration predicted values of the P pieces of equipment in the second preset time period and real equipment operation and maintenance duration data of the P pieces of equipment in the second preset time period; and according to the abnormal fluctuation characteristics, carrying out resource redistribution on the inspection period and the spare part inventory of the P devices, and determining an operation and maintenance resource configuration scheme. By adopting the technical scheme, the problem that resource allocation is rigid due to an existing equipment operation and maintenance method is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of equipment operation and maintenance, in particular to an intelligent operation and maintenance method and system based on big data analysis. BACKGROUND

[0002] With the development of industrial automation and Internet of Things technology, various types of equipment bear more and more important functions in enterprise operation. In order to ensure the stable operation of the equipment, it is usually necessary to carry out periodic inspection, maintenance and spare parts replacement and other operation and maintenance operations. At present, the operation and maintenance strategy of most enterprises depends on the preset periodic plan or manual experience judgment, and lacks the dynamic perception ability of the relationship between the equipment operation state and the operation and maintenance behavior.

[0003] In practical application, the operation and maintenance resource configuration is usually set by a unified standard, such as fixed inspection period, average allocation of spare parts inventory, etc. Although this method is convenient for management, it is difficult to realize the flexibility and pertinence of resource allocation when facing different equipment types, running environments and fault frequency differences. Some high-load running or fault-prone equipment may affect system stability due to insufficient maintenance. In addition, operation and maintenance data are scattered in different management systems in the form of logs, lacking effective integration mechanism and analysis means, resulting in that operation and maintenance decision mainly depends on subjective judgment, and it is difficult to provide strong data support for resource allocation.

[0004] For the problem of resource allocation rigidity caused by the existing equipment operation and maintenance method in the related art, no effective solution has been proposed so far. SUMMARY

[0005] The embodiments of the present application provide an intelligent operation and maintenance method and system based on big data analysis, to at least solve the problem of resource allocation rigidity caused by the existing equipment operation and maintenance method.

[0006] According to an aspect of the embodiments of the present application, an intelligent operation and maintenance method based on big data analysis is provided, comprising: acquiring historical operation and maintenance logs of equipment; determining P equipment affecting the change of operation and maintenance time according to the historical operation and maintenance logs, and predicting the equipment operation and maintenance time length prediction value of the P equipment in a second preset time period according to the equipment operation and maintenance time length sequence data of the P equipment in a first preset time period in the historical operation and maintenance logs, wherein the time length of the first preset time period and the second preset time period is equal, and P is an integer greater than or equal to 1; determining abnormal fluctuation characteristics according to the predicted equipment operation and maintenance time length prediction value of the P equipment in the second preset time period and the real equipment operation and maintenance time length data of the P equipment in the second preset time period, wherein the abnormal fluctuation characteristics are used to represent the behavior characteristics causing the change of operation and maintenance time length in the operation and maintenance process; reallocating resources of the inspection period and the spare parts inventory of the P equipment according to the abnormal fluctuation characteristics, and determining an operation and maintenance resource configuration scheme.

[0007] In an example embodiment, determining P devices affecting the operation and maintenance time length change according to the historical operation and maintenance log comprises: decomposing the device operation and maintenance time length sequence data of N devices in the historical operation and maintenance log by using a time series decomposition tool to obtain N decomposed device operation and maintenance time length sequence data, wherein the decomposition processing is used to separate the periodic change part and the long-term trend part, and N is an integer greater than or equal to 2; for the long-term trend part in each decomposed device operation and maintenance time length sequence data, the average slope of the operation and maintenance time length change is calculated by using a sliding window tool; P devices are determined from M devices, wherein the M devices are devices in the N devices whose corresponding average slope is not in a preset range, the P devices are devices in the M devices whose device downtime frequency and maintenance frequency fluctuation have a positive relationship, M is an integer greater than or equal to 1 and less than or equal to N, and P is an integer less than or equal to M.

[0008] In an example embodiment, predicting the device operation and maintenance time length prediction value of the P devices in a second preset time period according to the device operation and maintenance time length sequence data of the P devices in a first preset time period in the historical operation and maintenance log comprises: determining the device operation and maintenance time length prediction value of each device in the P devices in the second preset time period by the following manner to determine the device operation and maintenance time length prediction value of the P devices in the second preset time period, wherein the each device is a target device in the process of performing the following operations: the device operation and maintenance time length sequence data of the target device in the first preset time period is divided into multiple sub-sequence data by a time window division tool, and the multiple sub-sequence data are input into a long short-term memory neural network; the device operation and maintenance time length prediction value of the target device in the second preset time period determined by the long short-term memory neural network is obtained.

[0009] In an example embodiment, determining the abnormal fluctuation feature according to the predicted device operation and maintenance time length prediction value of the P devices in the second preset time period and the real device operation and maintenance time length data of the P devices in the second preset time period comprises: determining a plurality of time points in the second preset time period according to the predicted device operation and maintenance time length prediction value of the P devices in the second preset time period and the real device operation and maintenance time length data of the P devices in the second preset time period, wherein the difference between the device operation and maintenance time length prediction value corresponding to each time point in the plurality of time points and the real value in the real device operation and maintenance time length data is greater than a preset difference value; the downtime length and the resource idle rate of each time point in the plurality of time points are obtained to obtain a target feature set; the plurality of time points are clustered according to the target feature set, and the abnormal fluctuation feature is determined according to the clustering result.

[0010] In an example embodiment, the resource re-allocation of the inspection cycle and the spare parts inventory of the P devices according to the abnormal fluctuation feature comprises: adjusting the inspection cycle of the P devices according to the abnormal fluctuation feature and the downtime duration data corresponding to the P devices, to obtain an adjusted inspection cycle, wherein the historical operation and maintenance log comprises the downtime duration data; and re-allocating the spare parts inventory of the P devices according to the adjusted inspection cycle and the inventory turnover rate.

[0011] In an example embodiment, after determining the operation and maintenance resource configuration scheme, the method further comprises: performing virtual operation on the device operation and maintenance process by using a discrete event simulation tool according to the operation and maintenance resource configuration scheme, generating operation logs under a plurality of simulation scenarios, and determining corresponding downtime frequency records according to the operation logs under each simulation scenario; classifying the plurality of simulation scenarios according to the downtime frequency records corresponding to the plurality of simulation scenarios; in the case that there is a device of a target device type with a failure probability greater than a preset probability in a target scenario, shortening the inspection cycle of the device of the target device type according to the operation state data of the device of the target device type, to determine an updated inspection cycle scheme, wherein the target scenario is a scenario in which the downtime frequency exceeds a preset threshold, and the P devices comprise the device of the target device type; and dynamically adjusting the spare parts inventory of the P devices according to the updated inspection cycle scheme and the inventory turnover data.

[0012] In an example embodiment, the method further comprises: in the case that the updated inspection cycle scheme and the updated spare parts inventory do not satisfy a preset condition, performing simulation processing again, and re-allocating the inspection cycle and the spare parts inventory of the P devices according to the simulation result.

[0013] According to another aspect of the embodiments of the present application, an intelligent operation and maintenance system based on big data analysis is also provided, comprising: an acquisition module configured to acquire historical operation and maintenance logs of a device; a prediction module configured to determine P devices that affect operation and maintenance time length change according to the historical operation and maintenance logs, and predict device operation and maintenance time length prediction values of the P devices in a second preset time period according to device operation and maintenance time length sequence data of the P devices in a first preset time period in the historical operation and maintenance logs, wherein the first preset time period and the second preset time period have equal time lengths, and P is an integer greater than or equal to 1; a determination module configured to determine abnormal fluctuation characteristics according to the predicted device operation and maintenance time length prediction values of the P devices in the second preset time period and real device operation and maintenance time length data of the P devices in the second preset time period, wherein the abnormal fluctuation characteristics are used to represent behavior characteristics that cause operation and maintenance time length change in an operation and maintenance process; and an adjustment module configured to perform resource reallocation on inspection cycles and spare parts inventory of the P devices according to the abnormal fluctuation characteristics, and determine an operation and maintenance resource configuration scheme.

[0014] According to still another aspect of the embodiments of the present application, an electronic device is also provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements steps of the intelligent operation and maintenance method based on big data analysis when executing the computer program.

[0015] According to still another aspect of the embodiments of the present application, a computer readable storage medium is also provided, wherein the computer readable storage medium stores a computer program, and the computer program is configured to execute the intelligent operation and maintenance method based on big data analysis when running.

[0016] According to still another aspect of the embodiments of the present application, a computer program product is also provided, comprising a computer program, and the computer program is configured to execute the intelligent operation and maintenance method based on big data analysis when executed by a processor.

[0017] The present application, by introducing big data processing technology, analyzes the historical operation and maintenance log of the equipment, extracts the key equipment and its operation and maintenance characteristics affecting the operation and maintenance time length change. And on this basis, further use the historical operation and maintenance time length sequence to predict the future operation and maintenance time length trend, realize the change from experience driven to data driven. In addition, by comparing the prediction result with the actual operation and maintenance time length, the behavior characteristics (i.e. abnormal fluctuation characteristics) causing the abnormal fluctuation of the operation and maintenance time length are identified, and the inspection cycle and spare parts inventory are dynamically adjusted according to the identified abnormal fluctuation characteristics, so that the resource allocation can dynamically change according to the equipment running state, and the flexibility and pertinence of the operation and maintenance strategy are improved. The problem that the existing equipment operation and maintenance method causes the resource allocation to be rigid is solved. In addition, by dynamically adjusting the inspection cycle, the timeliness of the inspection of the key equipment and the fault response ability are improved, and the reliability of the equipment operation is enhanced; at the same time, by dynamically allocating the spare parts inventory, the inventory turnover rate is optimized. BRIEF DESCRIPTION OF DRAWINGS

[0018] The drawings incorporated into the specification and forming a part of the specification, show embodiments consistent with the present application, and together with the specification, serve to explain the principles of the present application.

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings required to be used in the embodiments or the prior art description will be briefly introduced. Obviously, for those skilled in the art, other drawings can also be obtained based on these drawings without creative labor.

[0020] Figure 1 is a flow diagram of an intelligent operation and maintenance method based on big data analysis according to an embodiment of the present application;

[0021] Figure 2 is a structural block diagram of an intelligent operation and maintenance system based on big data analysis according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to make the technical personnel in the art better understand the present application, the technical solutions of the embodiments of the present application will be described clearly and completely below with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the present application.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0024] In the present embodiment, an intelligent operation and maintenance method based on big data analysis is provided, Figure 1 is a flowchart of an intelligent operation and maintenance method based on big data analysis according to an embodiment of the present application, as Figure 1 shown, the flow includes the following steps S102-S108:

[0025] Step S102: Obtain the historical operation and maintenance log of the device;

[0026] Optionally, the historical operation and maintenance log includes but is not limited to:

[0027] Operation and maintenance event frequency: such as the number of device downtime, alarm triggering frequency, number of manual intervention, etc.;

[0028] Operation and maintenance response time: such as fault repair time, inspection task completion time, spare parts replacement time, etc.;

[0029] Resource call intensity: such as the number of operation and maintenance personnel called per unit time, tool usage frequency, spare parts replacement quantity, etc.;

[0030] Operation and maintenance task complexity score: a task difficulty coefficient generated based on historical data analysis, used to quantify technical challenges in the operation and maintenance process;

[0031] Device running state index: such as load rate, running time, mean time between failures (MTBF), etc. Indirectly affecting the operation and maintenance demand data.

[0032] Optionally, the historical operation and maintenance log can be obtained from the edge device and stored using a distributed file system, wherein a data sharding tool can be used for partition storage according to timestamp and device identifier. This architecture based on big data processing can effectively support efficient storage and access of massive operation and maintenance data, and improve the scalability and high availability of the system.

[0033] It should be noted that the core principle of the distributed file system is to store data scattered on multiple nodes to ensure high availability and scalability. Assuming that an enterprise has 1000 edge devices, which generate about 10,000 operation and maintenance logs and records every day, through the distributed file system, these data can be stored by timestamp and device identification partitioning to avoid single point of failure and improve data reading and maintenance time.

[0034] Optionally, data sharding tools can be used to further optimize storage and queries. Assuming that data is divided into shards by timestamp, with each device's log further subdivided by device identification. In this way, when querying the operation and maintenance data of a certain device on a certain day, only the specific shard needs to be accessed, significantly reducing query time. This partitioning method also provides a basis for parallel computing for subsequent data processing, saving resources. For the original data set after sharding, the use of data cleaning tools is crucial.

[0035] Optionally, for the original data set after sharding, data cleaning tools can be used to remove duplicate records and invalid fields to ensure data accuracy. If there are missing values in the original data set, linear interpolation method is used to complete the missing data based on time series, thus obtaining the cleaned and completed data set. This process uses big data processing technology to ensure data integrity.

[0036] For example, assuming that there are duplicate operation and maintenance log records in a shard, the cleaning tool can identify and delete duplicates to ensure data accuracy. At the same time, if it is found that a device has missing data in a certain time period, linear interpolation method can be used to complete it. Assuming that a device has 2 work hours and 2.4 work hours on March 1 and March 3, but the data for March 2 is missing, it can be estimated to be 2.2 work hours to complete. This method is based on the continuity of time series based on big data processing technology, ensuring data integrity.

[0037] Optionally, based on the cleaned and completed data set, a relational database management system is used to store historical operation and maintenance logs in a unified field format, and structured query language is used to extract timestamp, device identification and type fields to form standardized historical operation and maintenance logs.

[0038] It should be noted that the cleaned and completed data set is stored in a relational database management system, which can be in a unified field format. For example, the timestamp, device identification and type fields in the operation and maintenance log are standardized for storage, making it easy to extract data quickly through structured query language. Assuming that you need to query the type distribution of a device in the past 30 days, you only need to write a simple query statement to get the result, improving data access and maintenance time.

[0039] Optionally, the historical operation and maintenance logs can be grouped and calculated by the data aggregation tool according to the device identifier and the timestamp, and the total operation and maintenance time of each device is obtained. If the total operation and maintenance time exceeds a preset threshold, it is marked as a high-consumption device, and the classified operation and maintenance time record is obtained. This analysis method based on big data processing helps enterprises quickly identify operation and maintenance resource consumption anomalies.

[0040] For example, for structured historical operation and maintenance logs, the application of the data aggregation tool can realize time length summary. Assuming that after grouping according to the device identifier and the timestamp, it is calculated that the total operation and maintenance time of a device in the past week is 10 man-hours, and the preset threshold is 8 man-hours, then the device is marked as a high-consumption device. This classification method helps enterprises quickly identify anomalies and optimize resource allocation. By marking high-consumption devices, enterprises can further analyze the reasons, such as device aging or improper use, and take targeted measures to reduce maintenance frequency and time.

[0041] It should be noted that the historical operation and maintenance logs obtained through the above acquisition, cleaning, structuring and aggregation process based on big data processing have completeness, consistency and analyzability, providing a solid data support for the fine operation and maintenance management of enterprises.

[0042] Step S104: determining P devices that affect the operation and maintenance time length change according to the historical operation and maintenance logs, and predicting the device operation and maintenance time length prediction value of the P devices in a second preset time period according to the device operation and maintenance time length sequence data of the P devices in a first preset time period in the historical operation and maintenance logs, wherein the time length of the first preset time period and the second preset time period is equal, and P is an integer greater than or equal to 1;

[0043] In an exemplary embodiment, the above-mentioned determination of P devices that affect the operation and maintenance time length change according to the historical operation and maintenance logs means that based on the historical operation and maintenance logs, P key devices that have a significant impact on the overall operation and maintenance time length change are identified, which provides a decision basis for subsequent prediction and resource optimization. Specifically, it can be realized by the following steps S11-S13:

[0044] Step S11: decomposing the device operation and maintenance time length sequence data of N devices in the historical operation and maintenance logs by using a time series decomposition tool to obtain N decomposed device operation and maintenance time length sequence data, wherein the decomposition processing is used to separate the periodic change part and the long-term trend part, and N is an integer greater than or equal to 2;

[0045] It should be noted that in this step, the time series decomposition tool is used to process the historical device operation time sequence data of N devices in the structured historical operation and maintenance log obtained from the edge device. Through the tool, complex data fluctuations are divided into multiple components, including but not limited to periodic variation, long-term trend, and random noise, thereby extracting fluctuation characteristics reflecting the operation rules of the device.

[0046] For example, when processing structured historical operation and maintenance logs, the application of the time series decomposition tool can help identify the periodic variation and long-term trend of the operation time data. Assuming that the historical operation time records of a device show that the operation time is higher at the beginning of each month and tends to be stable in the middle of the month, this regular fluctuation can be extracted as a periodic variation part by the decomposition tool; at the same time, the long-term trend part may show that the operation time has been rising continuously in the past year, reflecting potential problems such as device aging or increased usage frequency.

[0047] Step S12: For each long-term trend part of the decomposed device operation time sequence data, a sliding window tool is used to calculate the average slope of the operation time change;

[0048] It should be noted that after obtaining the decomposed operation time sequence data of N devices, the average slope of the operation time change needs to be calculated for the long-term trend part using the sliding window tool.

[0049] The principle of the sliding window is to smooth fluctuations by averaging the data within a fixed time period, thereby identifying the trend direction. For example, with a 30-day window, the average slope of the data of a device in the past three windows is continuously positive and exceeds the preset threshold range (such as an increase of more than 20 working hours per month), which is marked as an abnormal trend. This process helps enterprises to identify devices with high risk of loss of control in advance, providing a target candidate set for subsequent in-depth analysis.

[0050] Step S13: Determine P devices from M devices, wherein the M devices are devices in the N devices whose average slope does not lie within the preset range, the P devices are devices in the M devices whose device downtime frequency and maintenance frequency fluctuation have a positive relationship, M is an integer greater than or equal to 1 and less than or equal to N, and P is an integer less than or equal to M.

[0051] It should be noted that the M devices are devices whose average slope deviates from the preset threshold range, i.e., devices with obvious trend abnormalities.

[0052] Optionally, the downtime records of the M devices can be analyzed to extract the correspondence between the device downtime and the maintenance frequency fluctuation using a data correlation tool. If the analysis result shows that the maintenance frequency fluctuation amplitude of a certain device significantly increases (for example, increases by about 20% after each downtime) after multiple downtimes, it is determined that there is a positive correlation between the downtime frequency and the maintenance frequency fluctuation, and the device is marked as a high-impact device, and is finally determined as one of the P key devices.

[0053] For example, a certain device has been down 5 times in the past three months, and the maintenance frequency fluctuation amplitude increases by about 20% after each downtime. Through correlation analysis, it can be determined that the downtime frequency and the maintenance frequency fluctuation are positively related, and therefore the device is marked as a high-impact device. Such devices have a significant driving effect on the change of the overall operation and maintenance time, and should be prioritized for resource allocation optimization.

[0054] It should be noted that the above steps S11-S13 use big data processing technology to deeply mine massive device operation and maintenance time data. Through a time series decomposition tool, periodic fluctuations and long-term trends are extracted from the historical data of the N devices, and M devices with abnormal trends are identified. In combination with the correlation analysis of downtime frequency and maintenance frequency fluctuation, P high-impact devices are selected. This process realizes the full-process automation from data-driven to key device identification, significantly improves the prediction accuracy and operation and maintenance resource allocation operation and maintenance time, and provides accurate data support for subsequent optimization.

[0055] In an exemplary embodiment, the above-mentioned prediction of the device operation and maintenance time of the P devices in the second preset time period based on the device operation and maintenance time sequence data of the P devices in the first preset time period is a prediction of the future operation and maintenance time trend based on historical operation and maintenance time data using big data analysis and deep learning technology. Specifically, it includes the following steps S21-S22 to determine the device operation and maintenance time prediction value of each device in the P devices in the second preset time period to determine the device operation and maintenance time prediction value of the P devices in the second preset time period. In the process of performing the following operations, the each device is a target device:

[0056] Step S21: divide the device operation and maintenance time sequence data of the target device in the first preset time period into multiple sub-sequence data by a time window division tool, and input the multiple sub-sequence data into a long short-term memory neural network;

[0057] It should be noted that for each identified key device affecting the operation and maintenance time length change (i.e. the target device in the P devices), a time window division tool can be used to divide the device operation and maintenance time length sequence data of the target device in the first preset time period into multiple fixed-length time window segments. For example, the continuous historical data is divided into several sub-sequences with 30 days as a window unit, so as to capture the periodic fluctuation amplitude and short-term fluctuation characteristics. This process helps to extract the timing regularity of device changes and provides structured input for subsequent model training. Then, the sub-sequence data is input into a long short-term memory neural network (LSTM) for training. The LSTM can effectively capture the long-term dependencies in the operation and maintenance time length sequence through the forgetting gate mechanism and input gate update control information flow, thereby improving the prediction accuracy.

[0058] Step S22: obtaining the device operation and maintenance time length prediction value of the target device in the second preset time period determined by the long short-term memory neural network.

[0059] It should be noted that after the LSTM model is trained, the LSTM model can output the prediction value of the target device in the future second preset time period (such as the next 30 days) based on the learned historical fluctuation pattern. For example, the fluctuation of a certain multi-cluster device in the past several time windows shows an upward trend, and the prediction tool estimates that the operation and maintenance time length may increase to 90 man-hours in the next 30 days, while the historical average is 80 man-hours, with a deviation range of 12.5%. If the deviation is within an acceptable range, the prediction value can be used as an important reference for subsequent resource scheduling and operation and maintenance time length control.

[0060] It should be noted that the above steps S21-S22 are based on time window division and LSTM deep learning model, and use big data processing technology to accurately predict the operation and maintenance time length of the key device. By dividing the historical operation and maintenance time length sequence into multiple sub-sequences and inputting them into the long short-term memory neural network, the timing characteristics and long-term dependencies of the data are fully mined, and the prediction accuracy is improved.

[0061] Step S106: determining an abnormal fluctuation feature based on the predicted device operation and maintenance time length prediction value of the P devices in the second preset time period and the real device operation and maintenance time length data of the P devices in the second preset time period, wherein the abnormal fluctuation feature is used to represent the behavior characteristics that cause the operation and maintenance time length to change during the operation and maintenance process;

[0062] It should be noted that in the present application, the abnormal fluctuation feature specifically refers to the abnormal fluctuation phenomenon of the operation and maintenance time length caused by certain specific operation and maintenance behaviors or device state changes during the device operation and maintenance process. These behavior characteristics include but are not limited to: device downtime increase, high resource idle rate, unreasonable inspection frequency, excessive spare parts replacement, etc., which directly or indirectly cause the unexpected increase of the operation and maintenance time length.

[0063] For example, in the comparison and analysis between the operation and maintenance duration prediction and the actual data, if a device has a higher operation and maintenance duration deviation at multiple time points, accompanied by a longer downtime and a higher resource idle rate, such behavior can be summarized as a typical "high-frequency abnormal fluctuation feature caused by device aging". This mode reflects the persistent resource waste problem caused by the poor running state of the device.

[0064] In an exemplary embodiment, the above step S106 can be implemented by the following steps S31-S33:

[0065] Step S31: determining a plurality of time points in the second preset time period according to the predicted device operation and maintenance duration prediction values of the P devices in the second preset time period and the real device operation and maintenance duration data of the P devices in the second preset time period, wherein the difference between the device operation and maintenance duration prediction value corresponding to each time point in the plurality of time points and the real value in the real device operation and maintenance duration data is greater than a preset difference value;

[0066] It should be noted that after the prediction of future operation and maintenance duration is completed (such as generating the operation and maintenance duration prediction value of the next three months by the LSTM model), the prediction result is compared with the actual operation and maintenance duration data of the corresponding time point, and the operation and maintenance duration deviation value is calculated. If the operation and maintenance duration deviation value of a certain time point exceeds the preset threshold (for example, the deviation rate exceeds 15%), it is marked as a time point where potential waste occurs. These time points constitute the basic set for subsequent abnormal fluctuation feature analysis.

[0067] For example, assuming that the predicted operation and maintenance duration of a certain month is 10 working hours, while the actual operation and maintenance duration reaches 12 working hours, the deviation rate exceeds 15%, and the time point is marked as an abnormal time point, which needs to be further analyzed for its behavior characteristics.

[0068] Step S32: obtaining the downtime and resource idle rate of each time point in the plurality of time points to obtain a target feature set;

[0069] It should be noted that for each time point marked as abnormal, the operation and maintenance behavior data related thereto is collected, including key indicators such as downtime and resource idle rate, to form a waste point feature set. These features reflect the specific behavior factors that cause the operation and maintenance duration deviation, and are the core dimension for identifying abnormal fluctuation features.

[0070] For example, the device downtime of a certain time point is 5 hours, and the resource idle rate is 30%. These information constitutes the target features of the time point, which are used for subsequent similarity analysis and classification.

[0071] Step S33: clustering the multiple time points according to the target feature set, and determining the abnormal fluctuation feature according to the clustering result.

[0072] It should be noted that after obtaining the target feature set, the K-means clustering tool is used for analysis. First, the difference between different feature points is calculated by the following Euclidean distance formula, and then combined with the initial cluster center selection and iterative clustering process, the time points with similar features are classified into a class, and the high-frequency abnormal fluctuation feature is identified.

[0073] Wherein d is the Euclidean distance, x1, x2 represent the downtime of two points (for example, the downtime of two points is 4 hours and 6 hours respectively), y1, y2 represent the resource idle rate of two points (for example, the resource idle rate of two points is 25% and 35% respectively).

[0074] For example, a clustering result shows that a group of time points are concentrated in the region with longer downtime and higher resource idle rate, which can be marked as "high-frequency abnormal fluctuation feature caused by device aging". In addition, for isolated points not covered by clustering, non-typical abnormalities such as high maintenance time caused by sudden failure or human operation error can also be identified by isolated point detection tool. For example, assuming that the downtime of a time point is 10 hours, far exceeding the average value of 2 hours of other points, and the idle rate is also abnormally high, it can be marked as an isolated abnormal point. This marking method can help enterprises focus on non-typical problems and avoid potential risks of resource waste.

[0075] It should be noted that the above step S31 quantitatively analyzes the deviation between the predicted value and the actual collected value, identifies the time points with significant deviation characteristics, and provides a key entry point for subsequent behavior feature extraction. Step S32 further extracts running indicators closely related to the maintenance state, such as downtime and resource idle rate, for these time points, constructs a structured feature set, and enhances the observability of device operation anomalies. Step S33 classifies the feature points using clustering algorithm on this basis, and mines the behavior patterns with repetitive characteristics, thereby realizing the automatic abstraction process from raw data to interpretable patterns. This process improves the perception of maintenance behavior trend changes and enhances the identification accuracy of non-typical maintenance activities, providing reliable data support for dynamic optimization of resource allocation.

[0076] Step S108: re-allocating the inspection cycle and spare parts inventory of the P devices according to the abnormal fluctuation feature, and determining the maintenance resource allocation scheme.

[0077] In an exemplary embodiment, the above step S108 can be implemented by the following steps S41-S42:

[0078] Step S41: Adjust the inspection cycle of the P devices according to the abnormal fluctuation characteristics and the downtime duration data corresponding to the P devices, to obtain an adjusted inspection cycle, wherein the historical operation and maintenance log includes the downtime duration data;

[0079] Optionally, for the identified high-frequency abnormal fluctuation characteristics (such as "high-frequency abnormal fluctuation characteristics caused by device aging"), the current inspection cycle and spare parts inventory data are extracted from the corresponding department and device type, and combined with the downtime duration information of the P devices, a linear programming tool is used to adjust the inspection cycle.

[0080] For example, a certain edge server cluster (including the cluster of the above P devices) is marked as a high-frequency abnormal fluctuation characteristic, and its average downtime duration is 4 hours / month, which is much higher than the normal value of 2 hours. By setting the constraint condition (such as downtime duration and device load), the original monthly inspection cycle is shortened to every 3 weeks, forming an adjusted inspection cycle set. This adjustment helps to discover device hidden dangers earlier and reduce the operation and maintenance duration fluctuation caused by sudden failures.

[0081] Step S42: Reallocate the spare parts inventory of the P devices according to the adjusted inspection cycle and the inventory turnover rate.

[0082] It should be noted that after obtaining the adjusted inspection cycle, the spare parts inventory and inventory turnover rate data of the P devices are combined, and an inventory optimization tool is used to dynamically calculate the spare parts allocation.

[0083] For example, assuming that the annual turnover rate of hard disks is 2 times and the annual turnover rate of power modules is 1 time, the inventory optimization tool can recalculate the spare parts inventory based on the replacement demand caused by the increase in inspection frequency. For the above edge server cluster, the original configuration is 50 hard disks and 20 power modules, and after optimization, it is adjusted to 60 hard disks (increased by 10 to cope with the aging risk) and 15 power modules (decreased by 5 to control the inventory operation and maintenance duration). This way realizes the fine management of spare parts inventory and improves the resource utilization operation and maintenance duration.

[0084] In addition, if the device downtime duration exceeds the preset threshold (such as 6 hours / month), the device running state and failure rate can be further combined to adjust the inspection priority using a priority sorting tool, and the operation and maintenance personnel and resources can be redistributed through a scheduling optimization tool to ensure that high-risk devices receive priority maintenance support.

[0085] It should be noted that the above steps S41-S42 are based on the high-frequency abnormal fluctuation feature recognition result, and the equipment inspection period and spare parts inventory are dynamically adjusted by using big data processing technology and optimization algorithm. This process fully integrates historical data analysis, prediction modeling and resource scheduling optimization, reflects the driving role of big data processing in operation and maintenance decision-making, and improves the scientificity of resource allocation and operation and maintenance time.

[0086] The above steps, by introducing big data processing technology, systematically analyze the historical operation and maintenance log of the equipment, extract the key equipment and its operation and maintenance characteristics that affect the operation and maintenance time change. And on this basis, further use the historical operation and maintenance time sequence to predict the future operation and maintenance time trend, realize the change from experience-driven to data-driven. In addition, by comparing the prediction result with the actual operation and maintenance time, the behavior characteristics (i.e. abnormal fluctuation characteristics) causing the abnormal fluctuation of operation and maintenance time are identified, and the inspection period and spare parts inventory are dynamically adjusted according to the identified abnormal fluctuation characteristics, so that the resource allocation can dynamically change according to the equipment running state, improving the flexibility and pertinence of operation and maintenance strategy. The problem that the existing equipment operation and maintenance method causes resource allocation to be rigid is solved. In addition, by dynamically adjusting the inspection period, the timeliness of inspection and fault response ability of key equipment are improved, and the reliability of equipment operation is enhanced; at the same time, by dynamically allocating spare parts inventory, the inventory turnover rate is optimized.

[0087] In an exemplary embodiment, after determining the operation and maintenance resource allocation scheme, the method further comprises the following steps S51-S54:

[0088] Step S51: According to the operation and maintenance resource allocation scheme, a discrete event simulation tool is used to virtually run the equipment operation and maintenance process, generate operation logs under multiple simulation scenarios, and determine corresponding downtime frequency records according to the operation logs under each simulation scenario;

[0089] It should be noted that based on the adjusted inspection period and spare parts inventory data, a discrete event simulation tool is used to virtually run the edge device operation and maintenance process. This tool discretizes time and events to simulate the state changes of the equipment at different time points, generates operation logs close to real scenarios, and records the occurrence time and repair time of each fault.

[0090] For example, for a department's edge router cluster, the inspection period is set to every 2 weeks, and the spare parts inventory is 30 core modules. The simulation tool can generate operation logs under multiple scenarios, such as a scenario where the monthly operation and maintenance time is 50 man-hours and the downtime frequency is 3 times / month.

[0091] Step S52: Classify the multiple simulation scenarios according to the downtime frequency records corresponding to the multiple simulation scenarios;

[0092] It should be noted that the simulation results are classified by the data comparison tool, and high-risk and low-risk scenarios are divided according to the preset threshold. For example, scenarios with operation time longer than 40 hours / month or downtime frequency higher than 2 times / month are classified as high-risk scenarios. If a scenario shows that the edge router cluster has an operation time of 55 hours / month and a downtime frequency of 4 times / month, which is obviously higher than the threshold, it is marked as a target combination for priority adjustment.

[0093] Step S53: In the case where there is a target device type of device with a failure probability greater than a preset probability in the target scenario, shorten the inspection cycle of the target device type of device according to the running state data of the target device type of device, and determine an updated inspection cycle scheme, wherein the target scenario is a scenario with a downtime frequency exceeding a preset threshold, and the P devices include the target device type of device;

[0094] It should be noted that when a target scenario (such as operation time or downtime frequency exceeding the standard) is identified as containing a device type with a high failure probability, a risk assessment tool is used in combination with its running state data (such as device load, historical failure rate, etc.) to determine whether the inspection cycle needs to be temporarily shortened.

[0095] For example, if the failure probability of the edge router cluster is 6%, which is higher than the average of 3%, the inspection cycle can be adjusted from every 2 weeks to every 10 days, forming an updated inspection cycle scheme to discover potential problems more timely.

[0096] Step S54: Dynamically adjust the spare parts inventory of the P devices according to the updated inspection cycle scheme and inventory turnover data.

[0097] It should be noted that based on the updated inspection cycle, in combination with inventory turnover data, an inventory management tool is used to dynamically optimize the spare parts inventory. For example, assuming that the annual turnover rate of the core module of the edge router cluster is 1.5 times, the tool analyzes and recommends increasing the spare parts inventory from 30 to 35 to meet the replacement needs brought by more frequent inspections, and re-planning the resource allocation ratio according to the actual needs of the department to ensure the rationality of resource allocation.

[0098] It should be noted that the above steps use discrete event simulation and big data processing analysis technology to build a virtual verification and dynamic optimization closed loop for operation and maintenance resource allocation, deeply integrating event-driven simulation, data comparison analysis, risk assessment, and dynamic inventory management, etc. Big data processing methods realize fine-tuning of resource allocation schemes. Through simulation experiments, potential problems are discovered in advance, improving the forward-looking and adaptability of operation and maintenance decisions, effectively reducing operation time fluctuations, and enhancing the robustness and intelligent level of the system in complex scenarios.

[0099] In an exemplary embodiment, the method further comprises: in the case that the updated inspection cycle scheme and the updated spare parts inventory do not meet the preset condition, performing simulation processing again, and according to the simulation result, performing resource reallocation on the inspection cycle and the spare parts inventory of the P devices again.

[0100] It should be noted that, after the initial operation and maintenance resource configuration is completed, the application verifies the configuration effect through discrete event simulation. If the simulation result does not reach the expected target (such as the downtime frequency is still higher than the preset threshold, the operation and maintenance time deviation exceeds the budget range), the multiple simulation iteration optimization process is entered, and the continuous optimization of resource configuration is realized.

[0101] Optionally, when the updated inspection cycle scheme and the spare parts inventory cannot meet the preset condition (for example, the preset downtime frequency is 3 times per month, and the current simulation result shows that it is still 4 times per month; or the operation and maintenance time control target is 8 working hours, but the current simulation operation and maintenance time is 8.8 working hours), the scene simulation data based on the simulation output is classified and processed by using a data comparison tool, the corresponding operation and maintenance time deviation value and isolated abnormal point are extracted, and a preliminary abnormal distribution record is formed.

[0102] Subsequently, the risk assessment tool is used to dynamically sort the inspection priority, the inspection frequency is adjusted for the device category with high operation and maintenance time deviation value, and a new inspection schedule is generated. If there are still isolated abnormal points in the inspection schedule, the inventory management tool is further used to reallocate the spare parts inventory in combination with the current resource configuration scheme, and adjusted inventory distribution data is obtained.

[0103] It should be noted that the operation and maintenance time control scheme can be virtually run for multiple rounds by using a simulation iteration calculation tool, and the scene simulation data is compared and analyzed after each run to determine whether the final operation and maintenance time control result meets the expected range. For example, after the first iteration, the operation and maintenance time is reduced to 9.5 working hours, and the downtime frequency is reduced to 4 times per month; after the second iteration, the operation and maintenance time is further reduced to 8.8 working hours, and the frequency is reduced to 3.5 times per month; after 5 iterations, the final operation and maintenance time is controlled to 8.2 working hours, and the frequency is reduced to 3 times per month, which reaches the threshold.

[0104] Through the above multiple simulation iteration mechanism, the resource configuration is changed from "static planning" to "dynamic optimization", and the operation and maintenance time control scheme finally formed by using big data processing technology has high adaptability and executability.

[0105] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be realized by means of software on a general hardware platform as necessary, and of course can also be realized by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solution of the present application or the part of the prior art that makes a contribution can be embodied in the form of a software product, which is stored in a storage medium (such as a ROM / RAM, a magnetic disk, or an optical disc) and includes a plurality of instructions for causing a terminal device (which can be a mobile phone, a computer, a server, or a network device) to execute the method of each embodiment of the present application.

[0106] In the embodiment, an intelligent operation and maintenance system based on big data analysis is also provided, which is used to implement the above embodiments and preferred embodiments, and will not be described again. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, hardware or a combination of software and hardware is also possible and is contemplated.

[0107] Figure 2 is a structural block diagram of an intelligent operation and maintenance system based on big data analysis according to an embodiment of the present application, and the system comprises:

[0108] The acquisition module 202 is configured to acquire historical operation and maintenance logs of the device.

[0109] The prediction module 204 is configured to determine P devices that affect the operation and maintenance time length change according to the historical operation and maintenance logs, and predict device operation and maintenance time length prediction values of the P devices in a second preset time period according to device operation and maintenance time length sequence data of the P devices in a first preset time period in the historical operation and maintenance logs, wherein the time length of the first preset time period and the second preset time period is equal, and P is an integer greater than or equal to 1.

[0110] The determination module 206 is configured to determine an abnormal fluctuation feature according to the predicted device operation and maintenance time length prediction values of the P devices in the second preset time period and real device operation and maintenance time length data of the P devices in the second preset time period, wherein the abnormal fluctuation feature is used to represent a behavior feature that causes the operation and maintenance time length to change during the operation and maintenance process.

[0111] The adjustment module 208 is configured to perform resource reallocation on the inspection cycle and spare parts inventory of the P devices according to the abnormal fluctuation feature, and determine an operation and maintenance resource configuration scheme.

[0112] The system, by introducing big data processing technology, analyzes the historical operation and maintenance logs of the equipment, extracts the key equipment and its operation and maintenance characteristics that affect the operation and maintenance time length, and further predicts the future operation and maintenance time length trend based on the historical operation and maintenance time length sequence, realizes the change from experience driving to data driving. In addition, by comparing the prediction result with the actual operation and maintenance time length, the behavior characteristics (i.e. abnormal fluctuation characteristics) causing the abnormal fluctuation of the operation and maintenance time length are identified, and the inspection cycle and spare parts inventory are dynamically adjusted according to the identified abnormal fluctuation characteristics, so that the resource allocation can dynamically change according to the equipment operation state, improving the flexibility and pertinence of the operation and maintenance strategy. The existing equipment operation and maintenance method can cause the problem of rigid resource allocation. In addition, by dynamically adjusting the inspection cycle, the timeliness of the inspection of the key equipment and the fault response capability are improved, and the reliability of the equipment operation is enhanced; at the same time, by dynamically allocating the spare parts inventory, the inventory turnover rate is optimized.

[0113] In an exemplary embodiment, the prediction module 204 is further configured to: decompose, by using a time series decomposition tool, the equipment operation and maintenance time length sequence data of the N equipment in the historical operation and maintenance logs to obtain N decomposed equipment operation and maintenance time length sequence data, wherein the decomposition processing is used to separate the periodic change part and the long-term trend part, and N is an integer greater than or equal to 2; for the long-term trend part in each decomposed equipment operation and maintenance time length sequence data, calculate the average slope of the operation and maintenance time length change by using a sliding window tool; determine P equipment from M equipment, wherein the M equipment is the equipment in the N equipment whose corresponding average slope is not in a preset range, the P equipment is the equipment in the M equipment whose equipment downtime frequency and maintenance frequency fluctuation have a positive relationship, M is an integer greater than or equal to 1 and less than or equal to N, and P is an integer less than or equal to M.

[0114] In an exemplary embodiment, the prediction module 204 is further configured to determine the equipment operation and maintenance time length prediction value of each equipment in the P equipment in the second preset time period by the following method to determine the equipment operation and maintenance time length prediction value of the P equipment in the second preset time period, wherein in the process of performing the following operations, the equipment is a target equipment: divide, by using a time window division tool, the equipment operation and maintenance time length sequence data of the target equipment in the first preset time period into a plurality of subsequence data, and input the plurality of subsequence data into a long short-term memory neural network; and obtain the equipment operation and maintenance time length prediction value of the target equipment in the second preset time period determined by the long short-term memory neural network.

[0115] In an example embodiment, the determining module 206 is further configured to determine a plurality of time points according to the predicted device operation and maintenance time length prediction value of the P devices in a second preset time period and the real device operation and maintenance time length data of the P devices in the second preset time period, wherein each time point in the plurality of time points corresponds to a device operation and maintenance time length prediction value and a real value in the real device operation and maintenance time length data, and a difference between the device operation and maintenance time length prediction value and the real value is greater than a preset difference value; obtain a downtime length and a resource idle rate of each time point in the plurality of time points to obtain a target feature set; perform clustering processing on the plurality of time points according to the target feature set, and determine an abnormal fluctuation feature according to a clustering result.

[0116] In an example embodiment, the adjusting module 208 is further configured to adjust a patrol cycle of the P devices according to the abnormal fluctuation feature and the downtime length data corresponding to the P devices, to obtain an adjusted patrol cycle, wherein the historical operation and maintenance log includes the downtime length data; and perform resource reallocation on a spare part inventory of the P devices according to the adjusted patrol cycle and an inventory turnover rate.

[0117] In an example embodiment, the system further includes an updating module configured to, after determining the operation and maintenance resource configuration scheme, perform virtual operation on a device operation and maintenance process by using a discrete event simulation tool according to the operation and maintenance resource configuration scheme, generate an operation log in a plurality of simulation scenarios, and determine a corresponding downtime frequency record according to the operation log in each simulation scenario; classify the plurality of simulation scenarios according to the corresponding downtime frequency records of the plurality of simulation scenarios; in a case where there is a device of a target device type with a failure probability greater than a preset probability in a target scenario, shorten a patrol cycle of the device of the target device type according to operation state data of the device of the target device type, to determine an updated patrol cycle scheme, wherein the target scenario is a scenario in which a downtime frequency exceeds a preset threshold, and the P devices include the device of the target device type; and perform dynamic adjustment on a spare part inventory of the P devices according to the updated patrol cycle scheme and inventory turnover data.

[0118] In an example embodiment, the updating module is further configured to, in a case where the updated patrol cycle scheme and the updated spare part inventory do not satisfy a preset condition, perform simulation processing again, and perform resource reallocation on the patrol cycle and the spare part inventory of the P devices again according to a simulation result.

[0119] Embodiments of the present application also provide a computer readable storage medium having a computer program stored therein, wherein the computer program is configured to execute the steps in any of the method embodiments described above when running.

[0120] Optionally, in the embodiment, the storage medium can be configured to store a computer program for executing the following steps.

[0121] In an example embodiment, the computer readable storage medium can include, but is not limited to, a U disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media capable of storing computer programs.

[0122] The specific examples in the embodiment can refer to the examples described in the above embodiments and example embodiments, which will not be repeated here.

[0123] The embodiment of the present application further provides a computer program product, comprising a computer program, which is executed by a processor to perform the steps in any of the method embodiments.

[0124] The embodiment of the present application further provides another computer program product, comprising a non-volatile computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to perform the steps in any of the method embodiments.

[0125] The embodiment of the present application further provides an electronic device, which comprises a memory and a processor, the memory stores a computer program, and the processor is configured to perform the steps in any of the method embodiments by the computer program.

[0126] The specific examples in the embodiment can refer to the examples described in the above embodiments and example embodiments, which will not be repeated here.

[0127] Obviously, those skilled in the art should understand that the modules or steps of the present application described above can be realized by general computing devices, which can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, which can be realized by program codes executable by computing devices, so that they can be stored in storage devices and executed by computing devices, and in some cases, the steps shown or described can be executed in different order, or they can be manufactured into individual integrated circuit modules, or multiple modules or steps can be manufactured into a single integrated circuit module. Thus, the present application is not limited to any specific combination of hardware and software.

[0128] The above merely describes the preferred embodiments of the present application, and it should be pointed out that, for those skilled in the art, some improvements and refinements can be made without departing from the principles of the present application, and these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. An intelligent operation and maintenance method based on big data analysis, characterized in that, The method comprises the following steps: obtaining historical operation and maintenance logs of devices; determining P devices that affect the operation and maintenance time length change according to the historical operation and maintenance logs, and predicting the device operation and maintenance time length prediction value of the P devices in a second preset time period according to the device operation and maintenance time length sequence data of the P devices in a first preset time period in the historical operation and maintenance logs, wherein the time length of the first preset time period and the second preset time period is equal, and P is an integer greater than or equal to 1; determining an abnormal fluctuation feature according to the predicted device operation and maintenance time length prediction value of the P devices in the second preset time period and the real device operation and maintenance time length data of the P devices in the second preset time period, wherein the abnormal fluctuation feature is used to represent the behavior feature that causes the operation and maintenance time length to change during operation and maintenance; reallocating resources of the inspection cycle and spare parts inventory of the P devices according to the abnormal fluctuation feature to determine an operation and maintenance resource configuration scheme.

2. The method of claim 1, wherein, Determining P devices that affect the operation and maintenance time length change according to the historical operation and maintenance logs comprises the following steps: performing decomposition processing on the device operation and maintenance time length sequence data of N devices in the historical operation and maintenance logs by using a time series decomposition tool to obtain N decomposed device operation and maintenance time length sequence data, wherein the decomposition processing is used to separate the periodic change part and the long-term trend part, and N is an integer greater than or equal to 2; for the long-term trend part in each decomposed device operation and maintenance time length sequence data, calculating the average slope of the operation and maintenance time length change by using a sliding window tool; determining P devices from M devices, wherein the M devices are devices whose corresponding average slopes are not in a preset range among the N devices, the P devices are devices whose device downtime frequency and maintenance frequency fluctuation have a positive relationship among the M devices, M is an integer greater than or equal to 1 and less than or equal to N, and P is an integer less than or equal to M.

3. The method of claim 1, wherein, Predicting the device operation and maintenance time length prediction value of the P devices in a second preset time period according to the device operation and maintenance time length sequence data of the P devices in a first preset time period in the historical operation and maintenance logs comprises the following steps: determining the device operation and maintenance time length prediction value of each device in the P devices in the second preset time period by the following method to determine the device operation and maintenance time length prediction value of the P devices in the second preset time period, wherein the device is a target device in the process of performing the following operations: segmenting the device operation and maintenance time length sequence data of the target device in the first preset time period into a plurality of sub-sequence data by using a time window division tool, and inputting the plurality of sub-sequence data into a long short-term memory neural network; obtaining the device operation and maintenance time length prediction value of the target device in the second preset time period determined by the long short-term memory neural network.

4. The method of claim 1, wherein, Determining an abnormal fluctuation feature according to the predicted device operation and maintenance time length prediction value of the P devices in the second preset time period and the real device operation and maintenance time length data of the P devices in the second preset time period comprises the following steps: determine a plurality of time points in the second preset time period according to the predicted P device operation and maintenance time length prediction values of the P devices in the second preset time period and the real device operation and maintenance time length data of the P devices in the second preset time period, wherein the difference between the device operation and maintenance time length prediction value corresponding to each time point in the plurality of time points and the real value in the real device operation and maintenance time length data is greater than a preset difference value; obtain the downtime and resource idle rate of each time point in the plurality of time points to obtain a target feature set; perform clustering processing on the plurality of time points according to the target feature set, and determine an abnormal fluctuation feature according to the clustering result.

5. The method of claim 1, wherein, redistribute resources of the P devices according to the abnormal fluctuation feature, including: adjust the inspection cycle of the P devices according to the abnormal fluctuation feature and the downtime data corresponding to the P devices, to obtain an adjusted inspection cycle, wherein the historical operation and maintenance log includes the downtime data; redistribute the spare parts inventory of the P devices according to the adjusted inspection cycle and the inventory turnover rate.

6. The method of claim 1, wherein, After determining the operation and maintenance resource configuration scheme, the method further includes: perform virtual operation on the device operation and maintenance process by using a discrete event simulation tool according to the operation and maintenance resource configuration scheme, generate operation logs in a plurality of simulation scenarios, and determine corresponding downtime frequency records according to the operation logs in each simulation scenario; classify the plurality of simulation scenarios according to the corresponding downtime frequency records; in the case that there is a target device type of device with a failure probability greater than a preset probability in a target scenario, shorten the inspection cycle of the target device type of device according to the operation state data of the target device type of device to determine an updated inspection cycle scheme, wherein the target scenario is a scenario in which the downtime frequency exceeds a preset threshold, and the P devices include the target device type of device; and dynamically adjust the spare parts inventory of the P devices according to the updated inspection cycle scheme and inventory turnover data.

7. The method of claim 6, wherein, The method further includes: in the case that the updated inspection cycle scheme and the updated spare parts inventory do not satisfy a preset condition, perform simulation processing again, and redistribute resources of the inspection cycle and the spare parts inventory of the P devices again according to the simulation result.

8. An intelligent operation and maintenance system based on big data analysis, characterized in that, including: an acquisition module configured to acquire historical operation and maintenance logs of devices; a prediction module configured to determine P devices that affect operation and maintenance time length changes according to the historical operation and maintenance logs, and predict device operation and maintenance time length prediction values of the P devices in a second preset time period according to device operation and maintenance time length sequence data of the P devices in a first preset time period in the historical operation and maintenance logs, wherein the time length of the first preset time period and the second preset time period is equal, and P is an integer greater than or equal to 1; The determining module is configured to determine an abnormal fluctuation feature according to the predicted device operation and maintenance time length prediction value of the P devices in a second preset time period and the real device operation and maintenance time length data of the P devices in the second preset time period, wherein the abnormal fluctuation feature is used to represent a behavior feature causing the operation and maintenance time length to change in the operation and maintenance process. The adjusting module is configured to perform resource reallocation on the inspection cycle and spare part inventory of the P devices according to the abnormal fluctuation feature, and determine an operation and maintenance resource configuration scheme.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, wherein the computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 7.

10. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the steps of the method in any one of claims 1 to 7.

Images