AI-based enterprise customer operation and maintenance demand prediction management system and method

Abstract

The invention relates to the technical field of customer operation and maintenance demand management, in particular to an enterprise customer operation and maintenance demand prediction management system and method based on AI, and discloses an enterprise customer operation and maintenance prediction demand classification module, a customer set sequence generation module and a customer operation and maintenance intervention guidance module. According to the system, a feature fusion means and a time sequence model are used, operation and maintenance associated features of all clients of an enterprise are predicted regularly, clients with similar operation and maintenance demands are classified in an integrated mode, an operation and maintenance demand unified prediction model of all the clients is built, deep operation and maintenance demand simulation analysis is conducted on all the client sets, and the operation and maintenance demands of all the clients are predicted. The operation and maintenance demand analysis efficiency and precision of each customer are improved, operation and maintenance management is involved in various customer sets according to the operation and maintenance demand sequence, and the customer who firstly needs to participate in operation and maintenance management is accurately and efficiently locked through deep analysis of the operation and maintenance demand of each customer in the customer sets.

Description

Technical Field

[0001] The present invention relates to the technical field of customer operation and maintenance demand management, and more specifically, to an AI-based enterprise customer operation and maintenance demand forecasting management system and method. Background Art

[0002] As enterprises deepen their digital transformation, the scale and complexity of infrastructure such as IT equipment, industrial controllers, and network systems are growing exponentially, posing severe challenges to managing customer operations and maintenance needs. Traditional operations and maintenance models rely on manual experience and regular inspections, making it difficult to cope with dynamic scenarios such as equipment performance degradation, business load fluctuations, and multi-device coupling failures. This leads to high operations and maintenance costs, low response efficiency, and insufficient fault resolution rates. In recent years, data-driven intelligent operations and maintenance technologies have gradually emerged, enabling fault prediction and resource scheduling by collecting equipment operating data. However, existing solutions generally suffer from the following technical bottlenecks:

[0003] The customer classification method is single: classification is based only on static attributes (such as industry and scale), without real-time clustering based on dynamic operation and maintenance demand characteristics (such as failure frequency, equipment aging, business peak cycle, etc.), resulting in classification results that cannot reflect the actual demand differences.

[0004] Lack of in-depth analysis in operation and maintenance decisions: Existing systems mostly remain at the demand forecasting stage, and have not formed a closed loop with operation and maintenance resource scheduling and intervention strategy optimization. They are unable to achieve full-process intelligence of "accurate prediction - efficient configuration - root cause treatment" through demand simulation analysis.

[0005] Existing technologies for managing customer O&M needs suffer from incomplete feature analysis, inefficient resource allocation, and crude strategy development, making them incapable of addressing complex equipment clusters and dynamic business scenarios. Therefore, this paper proposes an AI-based enterprise customer O&M demand forecasting management system and method. Summary of the Invention

[0006] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide an AI-based enterprise customer operation and maintenance demand forecasting management system and method.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] AI-based enterprise customer operation and maintenance demand forecasting and management system, including enterprise customer operation and maintenance forecast demand classification module, customer set sequence generation module, and customer operation and maintenance intervention guidance module;

[0009] The enterprise customer operation and maintenance forecast demand classification module is used to set the enterprise's customer operation and maintenance demand forecast cycle. Based on the customer operation and maintenance demand forecast cycle, the enterprise regularly determines the three-dimensional operation and maintenance feature forecast set for each customer and further generates multiple sets of operation and maintenance demand customers;

[0010] The customer set sequence generation module controls each customer set with operation and maintenance requirements to generate various types of operation and maintenance related prediction fusion features, builds a unified operation and maintenance demand prediction model for each customer set with operation and maintenance requirements, and then determines the comprehensive operation and maintenance demand prediction index for each customer set with operation and maintenance requirements to generate a customer set sequence;

[0011] The customer operation and maintenance intervention guidance module first determines the simulated operation and maintenance recommendation strategy for the first-ranked customer set of operation and maintenance needs, obtains the operation and maintenance root cause legacy value of each customer in the customer set of operation and maintenance needs, sorts each customer in descending order according to the value of the operation and maintenance root cause legacy value, and intervenes in operation and maintenance management for each customer in turn according to the ranking. When all customers in the customer set of operation and maintenance needs have completed operation and maintenance management, the simulated operation and maintenance recommendation strategy for the next-ranked customer set of operation and maintenance needs is determined, and this cycle is repeated.

[0012] Furthermore, the customer's three-dimensional operation and maintenance feature prediction set is regularly determined in the following manner: whenever a customer operation and maintenance demand forecast cycle ends, various types of operation and maintenance related features of the customer are collected, and various types of operation and maintenance related features are combined into an operation and maintenance related feature set in the form of a feature set, and the operation and maintenance related feature sets corresponding to the customer's previous c consecutive customer operation and maintenance demand forecast cycles are collected, and c+1 operation and maintenance related feature sets are combined into an operation and maintenance related time series set in the form of a time series, and the operation and maintenance related time series set is imported into the operation and maintenance related feature prediction model belonging to the customer, and the operation and maintenance related feature prediction model derives a three-dimensional operation and maintenance feature prediction set.

[0013] Furthermore, a method for generating a type of operation and maintenance related prediction fusion feature for a set of operation and maintenance demand customers is as follows: determine all customers included in a set of operation and maintenance demand customers, obtain a three-dimensional operation and maintenance feature prediction set for each customer, select a type of operation and maintenance related feature, extract the type of operation and maintenance related features in each three-dimensional operation and maintenance feature prediction set, fuse all the extracted operation and maintenance related features of the type, and generate the type of operation and maintenance related prediction fusion feature.

[0014] Furthermore, the method for determining the comprehensive forecast index of the operation and maintenance demand of the set of operation and maintenance demand customers is as follows: a unified forecast model of the operation and maintenance demand of a set of operation and maintenance demand customers is controlled to perform T mi Duration simulation, T mi After the duration ends, the unified operation and maintenance demand prediction model outputs various types of core physical quantities of the equipment, performs feature extraction on each type of core physical quantity, extracts each type of core physical features, combines each type of core physical features in the form of a feature set into a comprehensive core physical feature set, imports the comprehensive core physical feature set into the operation and maintenance demand model, and the operation and maintenance demand model derives a comprehensive operation and maintenance demand prediction index.

[0015] Furthermore, the customer set sequence is generated by sorting each operation and maintenance demand customer set in descending order according to the value of the comprehensive operation and maintenance demand prediction index, and generating the customer set sequence according to the sorting.

[0016] Furthermore, the method for determining the simulated operation and maintenance recommendation strategy for the set of customers with operation and maintenance requirements is as follows: a unified prediction model for operation and maintenance requirements corresponding to the set of customers with operation and maintenance requirements is obtained to perform multiple simulated operations and maintenance, and the operation and maintenance strategy corresponding to the simulated operation and maintenance with the smallest comprehensive prediction index of operation and maintenance requirements is marked as the simulated operation and maintenance recommendation strategy.

[0017] Furthermore, the method for obtaining the residual value of the customer's operation and maintenance root cause is as follows: select a customer, obtain the simulated operation and maintenance recommendation strategy of the customer's operation and maintenance demand customer set corresponding to the previous period, extract features of the simulated operation and maintenance recommendation strategy of the current period and the simulated operation and maintenance recommendation strategy of the previous period respectively, extract the simulated operation and maintenance strategy features of the current period and the simulated operation and maintenance strategy features of the previous period, match the simulated operation and maintenance strategy features of the current period and the simulated operation and maintenance strategy features of the previous period into a simulated operation and maintenance feature group, import the simulated operation and maintenance feature group into the operation and maintenance strategy comparison model, and derive an operation and maintenance strategy legacy index AD(s) from the operation and maintenance strategy comparison model, and set a comprehensive prediction of operation and maintenance demand. Threshold index: when the comprehensive forecast index of the customer set of operation and maintenance demand in the customer operation and maintenance demand forecast cycle is ≥ the comprehensive forecast threshold index of operation and maintenance demand, the corresponding customer operation and maintenance demand forecast cycle will be marked as an emergency demand cycle, and the number of emergency demands will be increased by one. All emergency demand cycles will be compared pairwise. When the two compared emergency demand cycles are adjacent customer operation and maintenance demand forecast cycles, the number of emergency demand durations will be increased by one. Finally, the number of emergency demands will be marked as KS(c), and the number of emergency demand durations will be marked as RY(u). The actual possible operation and maintenance demand value EW(a) of the customer is calculated by EW(a)=AD(s)*[KS(c)+RY(u)].

[0018] Furthermore, the AI-based enterprise customer operation and maintenance demand forecasting and management method has the following steps:

[0019] Step 1: Regularly determine the enterprise's three-dimensional operation and maintenance feature prediction set for each customer, and further generate multiple sets of customers with operation and maintenance requirements;

[0020] Step 2: Control each set of customers with operation and maintenance requirements to generate various types of operation and maintenance related prediction fusion features, and build a unified operation and maintenance demand prediction model for each set of customers with operation and maintenance requirements;

[0021] Step 3: Determine the comprehensive forecast index of operation and maintenance demand for each set of customers with operation and maintenance demand, and generate a customer set sequence;

[0022] Step 4: Determine the simulated operation and maintenance recommendation strategy for the top-ranked set of customers with operation and maintenance needs, and obtain the residual value of the operation and maintenance root cause for each customer in the set of customers with operation and maintenance needs;

[0023] Step 5: Sort each customer by the value of the residual value of the root cause of operation and maintenance from large to small, and intervene in operation and maintenance management for each customer in turn based on the ranking.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] The system of the present invention uses feature fusion means and time series models to regularly predict the operation and maintenance related features of each customer of the enterprise, classify customers with similar operation and maintenance needs in groups, and build a unified prediction model for the operation and maintenance needs of various customers. It conducts in-depth operation and maintenance demand simulation analysis on various customer sets, and further improves the efficiency and accuracy of each customer's operation and maintenance demand analysis on the basis of avoiding the waste of operation and maintenance resources caused by the "one-size-fits-all" operation and maintenance model. It intervenes in the operation and maintenance management of various customer sets in sequence of operation and maintenance needs, and through in-depth analysis of the operation and maintenance needs of each customer in the customer set, accurately and efficiently locks in the customers who need to be intervened in operation and maintenance management first. This system significantly improves the intelligence level of enterprise operation and maintenance, and is particularly suitable for industries with large equipment clusters and complex business scenarios, providing customers with all-round protection of "accurate prediction-efficient configuration-root cause governance". BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a principle block diagram of the system of the present invention;

[0027] Figure 2 A flow chart for determining the comprehensive forecast index of operation and maintenance demand for a collection of operation and maintenance demand customers;

[0028] Figure 3 4 is a flowchart of the method of the present invention. DETAILED DESCRIPTION

[0029] Example 1, refer to Figures 1 to 2 , an AI-based enterprise customer operation and maintenance demand forecasting and management system, including an enterprise customer operation and maintenance forecasting demand classification module, a customer set sequence generation module, and a customer operation and maintenance intervention guidance module.

[0030] The enterprise customer operation and maintenance forecast demand classification module sets the enterprise's customer operation and maintenance demand forecast cycle (the duration of the customer operation and maintenance demand forecast cycle is comprehensively set based on the customer's business characteristics, equipment requirements and other attributes). Based on the customer operation and maintenance demand forecast cycle, the enterprise regularly determines the three-dimensional operation and maintenance feature prediction set for each customer (that is, each customer corresponds to a three-dimensional operation and maintenance feature prediction set), and further generates multiple sets of operation and maintenance demand customers.

[0031] The regular determination method of the customer's three-dimensional operation and maintenance feature prediction set is as follows: whenever a customer operation and maintenance demand forecast cycle ends, various types of operation and maintenance related features of the customer are collected (the types of operation and maintenance related features include equipment basic attribute features, equipment operating status features, fault and maintenance features, etc. The above features are core data that directly support operation and maintenance demand analysis, but their impact on operation and maintenance demand includes both direct correlation and indirect driving), and various types of operation and maintenance related features are combined into an operation and maintenance related feature set in the form of feature sets (the operation and maintenance related feature set can be used for model training), and the operation and maintenance related feature sets corresponding to the previous c consecutive customer operation and maintenance demand forecast cycles of the customer are collected, and c+1 operation and maintenance related feature sets are combined into an operation and maintenance related time series set in the form of time series, and the operation and maintenance related time series set is imported into the operation and maintenance related feature prediction model belonging to the customer. The operation and maintenance related feature prediction model derives a three-dimensional operation and maintenance feature prediction set (the derived three-dimensional operation and maintenance feature prediction set is the customer's three-dimensional operation and maintenance feature prediction set).

[0032] Each customer has an operation and maintenance related feature prediction model. Each operation and maintenance related feature prediction model is built based on the LSTM model. In the specific implementation method, the construction process of the operation and maintenance related feature prediction model of customer A is disclosed: build an LSTM model, collect multiple operation and maintenance related time series sets of customer A, use the operation and maintenance related time series sets as basic data, and train the built LSTM model. In this process, a three-dimensional operation and maintenance feature prediction set is given to each operation and maintenance related time series set (the dimensions of the operation and maintenance related time series sets are very different, and directly inputting LSTM may cause the gradient to disappear / explode, so for each type of operation and maintenance in the operation and maintenance related time series set, a three-dimensional operation and maintenance feature prediction set is given). The associated features are normalized separately), the three-dimensional operation and maintenance feature prediction set contains various types of operation and maintenance related features for predicting customer A in the next customer operation and maintenance demand forecast cycle, and then the multiple operation and maintenance related time series sets are divided into training set, validation set and test set according to a specific ratio. The specific division ratio is determined to be 70%:20%:10%. The LSTM model is first repeatedly trained using the training set. During the training process, the validation set is used to timely verify the performance of the model in the training phase. According to the verification results, the model parameters are adjusted in time, and the model structure is optimized to make it develop in a more accurate and stable direction. Finally, the operation and maintenance related feature prediction model for customer A is completed.

[0033] A set of customers with operation and maintenance requirements contains multiple customers, and the similarity of operation and maintenance requirements between any two of these customers is greater than or equal to the set threshold. The similarity of operation and maintenance requirements between any two customers is as follows: two customers are marked as customer M and customer N, and each type of operation and maintenance related feature in the three-dimensional operation and maintenance feature prediction set of customer M is vectorized and converted into a vector M = (M1, M2, ..., M n), each type of operation and maintenance related feature in the three-dimensional operation and maintenance feature prediction set of customer N is vectorized and converted into a vector N = (N1, N2, ..., N n ), using the similarity formula Calculate the similarity of operation and maintenance requirements between customers M and N.

[0034] The customer set sequence generation module generates various types of operation and maintenance related prediction fusion features for each customer set with operation and maintenance requirements, builds a unified operation and maintenance demand prediction model for each customer set with operation and maintenance requirements, and then determines the comprehensive operation and maintenance demand prediction index of each customer set with operation and maintenance requirements to generate a customer set sequence.

[0035] A method for generating a type of operation and maintenance related prediction fusion feature for a set of customers with operation and maintenance requirements: determine all customers included in the set of customers with operation and maintenance requirements, obtain a three-dimensional operation and maintenance feature prediction set for each customer, select a type of operation and maintenance related feature, extract the type of operation and maintenance related features in each three-dimensional operation and maintenance feature prediction set, and fuse all the extracted operation and maintenance related features of the type to generate the type of operation and maintenance related prediction fusion feature.

[0036] Taking the device operating status features as an example, feature fusion processing includes but is not limited to the following methods: Standardization: Device operating status features are usually numerical data that changes over time, such as the device's CPU utilization, memory usage, disk I / O, etc. Before fusion, this data needs to be standardized to eliminate the dimensional effects between different features. A common standardization method is Z-score standardization, with the following formula: Where x is the original eigenvalue, μ is the mean of the feature, and σ is the standard deviation of the feature; Time series fusion: Since the equipment operation status feature is a time series data, it can be fused using the time series analysis method. For example, the sliding average method can be used to smooth the equipment operation status time series of multiple customers, and then perform weighted averaging. Suppose there are n customers' equipment operation status time series {x i1 ,x i2 ,…,x it}(i=1,2,…,n), for each time point t, the fused eigenvalue X t It can be calculated by the following formula: Where wi is the weight of each customer.

[0037] Taking fault and maintenance features as an example, feature fusion processing includes but is not limited to the following methods: Fault frequency statistics: Fault and maintenance features usually include information such as the number of equipment failures, fault types, maintenance time, etc. For information such as the number of failures and fault types, statistical methods can be used for fusion. For example, the number of failures of each customer within a certain period of time is counted, and then the average number of failures of all customers is calculated as the fused fault frequency feature.

[0038] Fault severity assessment: Fault types can be classified and quantified based on their severity. For example, faults can be categorized as minor, moderate, and severe, each assigned a different weight. A fault severity score is then calculated for each customer, and the weighted average of all customer scores is then taken to generate a fused fault severity signature.

[0039] Maintenance time series fusion: For time series data such as maintenance time, we can use time series fusion methods similar to those used for equipment operating status characteristics. For example, we can use the sliding average method to smooth the maintenance time series of multiple customers and then perform a weighted average.

[0040] The construction method of the unified prediction model of operation and maintenance demand for a set of customer sets with operation and maintenance demand is as follows: obtain the various types of operation and maintenance related prediction fusion features corresponding to a set of customer sets with operation and maintenance demand, import the various types of operation and maintenance related prediction fusion features into the general finite element software, and map the various types of operation and maintenance related prediction fusion features into the mapping parameters required for finite element simulation (taking the basic attribute features of the equipment as an example, the basic attribute features of the equipment include the number of CPU cores, fan air volume and other features, and the corresponding mapping parameters include the number of heat sources, thermal conductivity, convection heat transfer coefficient, etc. Taking the equipment operation status features as an example, the equipment operation status features include CPU utilization, memory temperature, fan speed, and the corresponding mapping parameters include power consumption (W), initial Initial temperature (°C), speed (RPM), taking fault and maintenance characteristics as an example, fault and maintenance characteristics include the number of historical hard disk failures, fan models replaced during maintenance, etc. The corresponding mapping parameters include contact thermal resistance (increases 10% after hard disk failure) and fan efficiency attenuation coefficient. The geometric model is constructed in general finite element software. Taking the server chassis as an example, parametric design is first performed: heat sink thickness t, fan position (x, y, z), and CPU layout n are used as adjustable parameters, corresponding to the basic property characteristics of the equipment. Then, the material and boundary conditions are defined. For example, the material of the heat sink is aluminum alloy, the thermal conductivity is defined as 180W / (m·K), and the density is 2700kg / m 3 For example, if the ambient temperature is 25°C, the fan air volume is Q = fan speed × 0.01m 3 / min (linear mapping of operating status characteristics), multi-physics field settings (such as thermal-structural coupling, thermal analysis: calculate CPU power consumption (P = CPU utilization × 200W) as a heat source, simulate temperature distribution), and finally build a unified operation and maintenance demand prediction model for the set of operation and maintenance demand customers through model verification and validation. The unified operation and maintenance demand prediction model can simulate equipment operation.

[0041] The customer set sequence is generated by sorting each operation and maintenance demand customer set in descending order according to the value of the comprehensive operation and maintenance demand prediction index, and generating the customer set sequence based on the sorting.

[0042] The method for determining the comprehensive forecast index of the operation and maintenance demand of the set of operation and maintenance demand customers is as follows: a unified forecast model of the operation and maintenance demand of a set of operation and maintenance demand customers is used to perform T mi The simulation of the duration (its essence is to conduct T mi duration of the run), T mi After the duration is over, the unified operation and maintenance demand prediction model outputs various types of core physical quantities of the equipment (types of core physical quantities include temperature distribution cloud map, airflow velocity vector map, thermal resistance / power consumption curve), and performs feature extraction on various types of core physical quantities (feature extraction methods of temperature distribution cloud map include basic statistical feature extraction, image feature extraction, etc., feature extraction methods of airflow velocity vector map include vector field statistical feature extraction, critical path feature extraction, fluid mechanics parameter feature extraction, etc., feature extraction methods of thermal resistance / power consumption curve include curve morphology features, etc.), extracts various types of core physical features, combines various types of core physical features into a comprehensive core physical feature set in the form of a feature set, imports the comprehensive core physical feature set into the operation and maintenance demand model, and the operation and maintenance demand model derives a comprehensive operation and maintenance demand prediction index.

[0043] The construction method of the operation and maintenance demand model is as follows: build a deep learning model, collect multiple comprehensive core physical feature sets, use the comprehensive core physical feature sets as basic data, and train the built deep learning model. In this process, each comprehensive core physical feature set is given an operation and maintenance demand comprehensive prediction index. The value range of the operation and maintenance demand comprehensive prediction index is set between 1 and 100. The size of the operation and maintenance demand comprehensive prediction index has a clear meaning. The larger the value, the greater the operation and maintenance demand of the comprehensive core physical feature set, which means that the equipment has better thermal stability, heat dissipation efficiency, and load bearing capacity. The higher the comprehensive risk in terms of capabilities, the more urgent the corresponding operation and maintenance needs (such as inspection, maintenance, upgrades, etc.). Then, multiple comprehensive core physical feature sets are divided into training sets, verification sets, and test sets according to a specific ratio. The specific division ratio is determined to be 70%:20%:10%. First, the deep learning model is repeatedly trained using the training set. During the training process, the performance of the model in the training stage is timely verified with the help of the verification set. According to the verification results, the model parameters are adjusted in time, and the model structure is optimized to make it develop in a more accurate and stable direction. Finally, the operation and maintenance demand model is completed.

[0044] The customer operation and maintenance intervention guidance module first determines the simulated operation and maintenance recommendation strategy for the operation and maintenance demand customer set ranked first, obtains the operation and maintenance root cause legacy value of each customer in the operation and maintenance demand customer set, sorts each customer in order from large to small according to the value of the operation and maintenance root cause legacy value, and intervenes in the operation and maintenance management of each customer in turn according to the sorting (intervenes in the actual operation and maintenance management). When all customers in the operation and maintenance demand customer set complete the operation and maintenance management, determine the simulated operation and maintenance recommendation strategy for the operation and maintenance demand customer set ranked next, obtains the operation and maintenance root cause legacy value of each customer in the operation and maintenance demand customer set, sorts each customer in order from large to small according to the value of the operation and maintenance root cause legacy value, and intervenes in the operation and maintenance management of each customer in turn according to the sorting (that is, repeats the processing process of the previous operation and maintenance demand customer set in a cycle).

[0045] The method for determining the simulated operation and maintenance recommendation strategy for the set of customers with operation and maintenance needs is as follows: a unified prediction model of operation and maintenance needs corresponding to the set of customers with operation and maintenance needs is obtained to perform multiple simulated operations and maintenance (an operation and maintenance strategy is automatically generated for each simulated operation and maintenance, and the operation and maintenance processing of the equipment in the unified prediction model of operation and maintenance needs is simulated. The operation and maintenance strategy corresponding to each simulated operation and maintenance is different). The operation and maintenance strategy corresponding to the simulated operation and maintenance with the smallest comprehensive prediction index of operation and maintenance needs is marked as the simulated operation and maintenance recommendation strategy.

[0046] The method for obtaining the residual value of the customer's operation and maintenance root cause is as follows: select a customer, obtain the simulated operation and maintenance recommendation strategy of the customer's operation and maintenance demand customer set corresponding to the customer's previous cycle, extract features of the simulated operation and maintenance recommendation strategy of the current cycle and the simulated operation and maintenance recommendation strategy of the previous cycle respectively, extract the simulated operation and maintenance strategy features of the current cycle and the simulated operation and maintenance strategy features of the previous cycle, match the simulated operation and maintenance strategy features of the current cycle and the simulated operation and maintenance strategy features of the previous cycle into a simulated operation and maintenance feature group, import the simulated operation and maintenance feature group into the operation and maintenance strategy comparison model, and derive an operation and maintenance strategy legacy index AD(s) from the operation and maintenance strategy comparison model. Set the comprehensive prediction threshold index of operation and maintenance demand (based on system settings). When the comprehensive prediction index of operation and maintenance demand of the customer's operation and maintenance demand customer set in the customer's operation and maintenance demand prediction cycle is ≥ the comprehensive prediction threshold index of operation and maintenance demand The corresponding customer operation and maintenance demand forecast cycle is marked as an emergency demand cycle, and the number of emergency demands is increased by one. When the comprehensive operation and maintenance demand forecast index of the operation and maintenance demand customer set in the customer operation and maintenance demand forecast cycle is less than the comprehensive operation and maintenance demand forecast threshold index, no marking is performed. All emergency demand cycles are compared pairwise. When the two compared emergency demand cycles are adjacent customer operation and maintenance demand forecast cycles, the number of emergency demand durations is increased by one. When the two compared emergency demand cycles are not adjacent customer operation and maintenance demand forecast cycles, there is no need to increase the number of emergency demand durations by one. Finally, the number of emergency demands is marked as KS(c), and the number of emergency demand durations is marked as RY(u). The actual possible operation and maintenance demand value EW(a) of the customer is calculated by EW(a)=AD(s)*[KS(c)+RY(u)].

[0047] The construction method of the operation and maintenance strategy comparison model is as follows: build a deep learning model, collect multiple simulated operation and maintenance feature groups, each simulated operation and maintenance feature group contains two simulated operation and maintenance strategy features, and use the simulated operation and maintenance feature group as the basic data to train the built deep learning model. In this process, an operation and maintenance strategy legacy index is assigned to each simulated operation and maintenance feature group. The value range of the operation and maintenance strategy legacy index is set between 1 and 5. The size of the operation and maintenance strategy legacy index has a clear meaning. The larger the value, the closer the two simulated operation and maintenance strategy features in the simulated operation and maintenance feature group are, which means that the strategies used in the two simulated operations are the same. The more consistent, that is, the more consistent the later simulated operation and maintenance is with the previous simulated operation and maintenance, it is very likely that there are residual operation and maintenance problems that have not been fundamentally solved. Then, multiple simulated operation and maintenance feature groups are divided into training set, validation set and test set according to a specific ratio. The specific division ratio is determined to be 70%:15%:15%. First, the deep learning model is repeatedly trained using the training set. During the training process, the validation set is used to timely verify the performance of the model in the training stage. According to the verification results, the model parameters are adjusted in time, and the model structure is optimized to make it develop in a more accurate and stable direction. Finally, the operation and maintenance strategy comparison model is completed.

[0048] The system of the present invention uses feature fusion means and time series models to regularly predict the operation and maintenance related features of each customer of the enterprise, classify customers with similar operation and maintenance needs in a collective manner, and build a unified prediction model for the operation and maintenance needs of various customers. It conducts in-depth operation and maintenance demand simulation analysis on various customer sets, and further improves the efficiency and accuracy of each customer's operation and maintenance demand analysis on the basis of avoiding the waste of operation and maintenance resources caused by the "one-size-fits-all" operation and maintenance model. It intervenes in the operation and maintenance management of various customer sets in the order of operation and maintenance needs, and through in-depth analysis of the operation and maintenance needs of each customer in the customer set, accurately and efficiently locks in the customers who need to be intervened in operation and maintenance management first. This system significantly improves the intelligence level of enterprise operation and maintenance, and is particularly suitable for industries with large equipment clusters and complex business scenarios (such as cloud computing, intelligent manufacturing, and financial technology), providing customers with all-round protection of "accurate prediction-efficient configuration-root cause governance".

[0049] Example 2, refer to Figure 3 , the AI-based enterprise customer operation and maintenance demand forecasting and management method has the following steps:

[0050] Step 1: Regularly determine the company's three-dimensional operation and maintenance feature prediction set for each customer, and further generate multiple sets of customers with operation and maintenance requirements.

[0051] Step 2: Control each set of customers with operation and maintenance requirements to generate various types of operation and maintenance related prediction fusion features, and build a unified prediction model for operation and maintenance requirements for each set of customers with operation and maintenance requirements.

[0052] Step 3: Determine the comprehensive forecast index of operation and maintenance demand for each set of customers with operation and maintenance demand, and generate a customer set sequence.

[0053] Step 4: Determine the simulated operation and maintenance recommendation strategy for the set of customers with the highest operation and maintenance requirements, and obtain the residual value of the operation and maintenance root cause of each customer in the set of customers with the highest operation and maintenance requirements.

[0054] Step 5: Sort each customer by the value of the residual value of the root cause of operation and maintenance from large to small, and intervene in operation and maintenance management for each customer in turn based on the ranking.

[0055] The above formulas are all dimensionless and numerically calculated, and the preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0056] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0057] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0058] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0059] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0060] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0061] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0062] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. AI-based enterprise customer operation and maintenance demand forecasting and management system, characterized by: It includes enterprise customer operation and maintenance forecast demand classification module, customer set sequence generation module, and customer operation and maintenance intervention guidance module; The enterprise customer operation and maintenance forecast demand classification module is used to set the enterprise's customer operation and maintenance demand forecast cycle. Based on the customer operation and maintenance demand forecast cycle, the enterprise regularly determines the three-dimensional operation and maintenance feature forecast set for each customer and further generates multiple sets of operation and maintenance demand customers; The customer set sequence generation module controls each customer set with operation and maintenance requirements to generate various types of operation and maintenance related prediction fusion features, builds a unified operation and maintenance demand prediction model for each customer set with operation and maintenance requirements, and then determines the comprehensive operation and maintenance demand prediction index for each customer set with operation and maintenance requirements to generate a customer set sequence; The customer operation and maintenance intervention guidance module first determines the simulated operation and maintenance recommendation strategy for the first-ranked customer set of operation and maintenance needs, obtains the operation and maintenance root cause legacy value of each customer in the customer set of operation and maintenance needs, sorts each customer in descending order according to the value of the operation and maintenance root cause legacy value, and intervenes in operation and maintenance management for each customer in turn according to the ranking. When all customers in the customer set of operation and maintenance needs have completed operation and maintenance management, the simulated operation and maintenance recommendation strategy for the next-ranked customer set of operation and maintenance needs is determined, and this cycle is repeated.

2. The AI-based enterprise customer operation and maintenance demand forecasting and management system according to claim 1 is characterized in that: The regular determination method of the customer's three-dimensional operation and maintenance feature prediction set is as follows: whenever a customer operation and maintenance demand forecast cycle ends, various types of operation and maintenance related features of the customer are collected, and various types of operation and maintenance related features are combined into an operation and maintenance related feature set in the form of feature sets. The operation and maintenance related feature sets corresponding to the customer's previous c consecutive customer operation and maintenance demand forecast cycles are collected, and c+1 operation and maintenance related feature sets are combined into an operation and maintenance related time series set in the form of time series. The operation and maintenance related time series set is imported into the operation and maintenance related feature prediction model belonging to the customer, and the operation and maintenance related feature prediction model derives a three-dimensional operation and maintenance feature prediction set.

3. The AI-based enterprise customer operation and maintenance demand forecasting and management system according to claim 1 is characterized in that: A method for generating a type of operation and maintenance related prediction fusion feature for a set of customers with operation and maintenance requirements: determine all customers included in the set of customers with operation and maintenance requirements, obtain a three-dimensional operation and maintenance feature prediction set for each customer, select a type of operation and maintenance related feature, extract the type of operation and maintenance related features in each three-dimensional operation and maintenance feature prediction set, and fuse all the extracted operation and maintenance related features of the type to generate the type of operation and maintenance related prediction fusion feature.

4. The AI-based enterprise customer operation and maintenance demand forecasting and management system according to claim 1 is characterized in that: The method for determining the comprehensive forecast index of the operation and maintenance demand of the set of operation and maintenance demand customers is as follows: a unified forecast model of the operation and maintenance demand of a set of operation and maintenance demand customers is used to perform T mi Duration simulation, T mi After the duration ends, the unified operation and maintenance demand prediction model outputs various types of core physical quantities of the equipment, performs feature extraction on each type of core physical quantity, extracts each type of core physical features, combines each type of core physical features in the form of a feature set into a comprehensive core physical feature set, imports the comprehensive core physical feature set into the operation and maintenance demand model, and the operation and maintenance demand model derives a comprehensive operation and maintenance demand prediction index.

5. The AI-based enterprise customer operation and maintenance demand forecasting and management system according to claim 1 is characterized in that: Method for generating a customer set sequence: sort each set of customers with operation and maintenance requirements in descending order according to the value of the comprehensive forecast index of operation and maintenance requirements, and generate a customer set sequence based on the sorting.

6. The AI-based enterprise customer operation and maintenance demand forecasting and management system according to claim 1 is characterized in that: The method for determining the simulated operation and maintenance recommendation strategy for the set of customers with operation and maintenance requirements is as follows: a unified prediction model for operation and maintenance requirements corresponding to the set of customers with operation and maintenance requirements is obtained to perform multiple simulated operations and maintenance, and the operation and maintenance strategy corresponding to the simulated operation and maintenance with the smallest comprehensive prediction index of operation and maintenance requirements is marked as the simulated operation and maintenance recommendation strategy.

7. The AI-based enterprise customer operation and maintenance demand forecasting and management system according to claim 1 is characterized in that: The method for obtaining the residual value of the customer's operation and maintenance root cause is as follows: select a customer, obtain the simulated operation and maintenance recommendation strategy of the customer's operation and maintenance demand customer set corresponding to the previous period, extract features of the simulated operation and maintenance recommendation strategy of the current period and the simulated operation and maintenance recommendation strategy of the previous period respectively, extract the simulated operation and maintenance strategy features of the current period and the simulated operation and maintenance strategy features of the previous period, match the simulated operation and maintenance strategy features of the current period and the simulated operation and maintenance strategy features of the previous period into a simulated operation and maintenance feature group, import the simulated operation and maintenance feature group into the operation and maintenance strategy comparison model, and derive an operation and maintenance strategy legacy index AD(s) from the operation and maintenance strategy comparison model, and set the comprehensive prediction threshold index of operation and maintenance demand When the comprehensive operation and maintenance demand forecast index of the operation and maintenance demand customer set in the customer operation and maintenance demand forecast cycle is ≥ the comprehensive operation and maintenance demand forecast threshold index, the corresponding customer operation and maintenance demand forecast cycle will be marked as an emergency demand cycle, and the number of emergency demands will be increased by one. All emergency demand cycles will be compared pairwise. When the two compared emergency demand cycles are adjacent customer operation and maintenance demand forecast cycles, the number of emergency demand durations will be increased by one. Finally, the number of emergency demands will be marked as KS(c), and the number of emergency demand durations will be marked as RY(u). The actual possible operation and maintenance demand value EW(a) of the customer is calculated by EW(a)=AD(s)*[KS(c)+RY(u)].

8. An AI-based enterprise customer operation and maintenance demand forecasting and management method, applied to the AI-based enterprise customer operation and maintenance demand forecasting and management system according to any one of claims 1 to 7, characterized in that: Here are the steps: Step 1: Regularly determine the enterprise's three-dimensional operation and maintenance feature prediction set for each customer, and further generate multiple sets of customers with operation and maintenance requirements; Step 2: Control each set of customers with operation and maintenance requirements to generate various types of operation and maintenance related prediction fusion features, and build a unified operation and maintenance demand prediction model for each set of customers with operation and maintenance requirements; Step 3: Determine the comprehensive forecast index of operation and maintenance demand for each set of customers with operation and maintenance demand, and generate a customer set sequence; Step 4: Determine the simulated operation and maintenance recommendation strategy for the top-ranked set of customers with operation and maintenance needs, and obtain the residual value of the operation and maintenance root cause for each customer in the set of customers with operation and maintenance needs; Step 5: Sort each customer by the value of the residual value of the root cause of operation and maintenance from large to small, and intervene in operation and maintenance management for each customer in turn based on the ranking.

Images