CN-121304080-B - Fault repairing method and system based on big data analysis

CN121304080BCN 121304080 BCN121304080 BCN 121304080BCN-121304080-B

Abstract

The application relates to the technical field of data processing, in particular to a fault repairing method and system based on big data analysis; the method comprises the steps of responding to a new repair request, obtaining state deviation degree based on the difference between a real-time operation parameter and each type of historical operation parameter, converting text information in the new repair request into a real-time fault vector, calculating semantic similarity between the real-time fault vector and each type of fault historical text vector, carrying out weighted fusion on the state deviation degree and the semantic similarity corresponding to any fault type to obtain a comprehensive evaluation index of the fault type, determining the fault type with the highest evaluation index as a diagnosis result of current repair, and carrying out intelligent dispatch for the repair task according to the diagnosis result. The application has the effects of improving the maintenance timeliness and improving the utilization rate of human resources.

Inventors

Li Shaozhao
LI LI
JIANG GUOQING

Assignees

广东的修数字科技有限公司

Dates

Publication Date: 20260508
Application Date: 20251021

Claims (7)

1. The fault repairing method based on big data analysis is characterized by comprising the steps of obtaining historical operation parameters and maintenance records of faults of all types of equipment, and obtaining historical text vectors of all fault types by using a pre-trained natural language processing model based on text information in the maintenance records; The method comprises the steps of receiving a new repair request, acquiring state deviation degree based on the difference between a real-time operation parameter and each type of historical operation parameter, converting text information in the new repair request into a real-time fault vector, calculating semantic similarity between the real-time fault vector and each type of historical text vector, carrying out weighted fusion on the state deviation degree and the semantic similarity corresponding to the fault type for any fault type to obtain a comprehensive evaluation index of the fault type, and determining the fault type with the highest evaluation index as a diagnosis result of current repair; The step of intelligently dispatching the repair task according to the diagnosis result comprises the steps of constructing a dispatching cost function, wherein the dispatching cost function comprises the estimated time cost of reaching a task place by an engineer, the matching cost of the skill of the engineer and the task requirement and the current workload cost of the engineer; The step of acquiring the state deviation degree based on the difference between the real-time operation parameters and the historical operation parameters of all types comprises the steps of enabling the real-time operation parameters and the historical operation parameters to comprise multidimensional data, wherein the multidimensional data in the real-time operation parameters form real-time feature vectors, taking the mean value of all the dimensional data in the historical operation parameters as a historical fault vector, and taking the mahalanobis distance between the real-time feature vectors and the historical fault vector as the state deviation degree; the preset allocation algorithm is a Hungary algorithm; The step of calculating the matching cost of the skill of the engineer and the task requirement comprises the steps of obtaining the grade requirement of the skill required by the task and the actual grade of the skill of the engineer, and accumulating according to the product of the grade difference value of the grade requirement and the actual grade of the skill of the engineer and the preset skill importance coefficient; for any task, the calculation formula of the dispatch cost to dispatch it to any engineer can be expressed as: ; Representing tasks Dispatch to engineers The dispatch cost of (a); As a first weight coefficient, a first set of weights, Is a second weight coefficient; is a third weight coefficient; Is a fourth weight coefficient; the projected time cost for an engineer to arrive at a task site; Representing the matching cost of engineer skill and task requirements; Representing the current workload cost of the engineer; representing spare part acquisition costs.
2. The method of claim 1, wherein the pre-trained natural language processing model is a BERT model.
3. The method for repairing faults based on big data analysis of claim 1, wherein the dispatch cost function further comprises spare part acquisition cost, and the spare part acquisition cost is calculated according to real-time navigation information of the current position of an engineer, the position of a spare part library and a task place.
4. The method for fault repair based on big data analysis of claim 1, further comprising retrieving a standard workflow, a historical maintenance case or a technical drawing related to the diagnosis result from the knowledge base after dispatch and pushing it to a terminal device of the assigned engineer.
5. The fault repairing method based on big data analysis according to claim 1, wherein the weight coefficients of each cost in the form cost function are dynamically adjusted by using a reinforcement learning model, wherein the input of the reinforcement learning model is a state index of a current operation and maintenance system, and the input of the reinforcement learning model is a set of updated weight coefficients.
6. The method for fault reporting based on big data analysis of claim 1, further comprising receiving feedback data submitted by an engineer after completion of maintenance, the feedback data including a final confirmed cause of the fault and an actual maintenance operation.
7. The big data analysis based fault repairing system is characterized by comprising a processor and a memory, wherein the memory stores computer program instructions, and the computer program instructions realize the big data analysis based fault repairing method according to any one of claims 1-6 when executed by the processor.

Description

Fault repairing method and system based on big data analysis Technical Field The application relates to the technical field of data processing, in particular to a fault repairing method and system based on big data analysis. Background In modern industrial and commercial environments, stable operation of industrial equipment is critical. The traditional fault repairing flow is highly dependent on manual processing, namely, a field personnel manually creates a work order after finding an abnormality, and a dispatcher judges the fault and assigns engineers by means of personal experience and text description of the work order. The mode has more inconveniences, for example, due to subjectivity and ambiguity of user description, judgment of a dispatcher is easy to miss, so that the skill of a dispatched engineer is not matched or correct spare parts are not carried, the secondary maintenance rate and the cost are increased, meanwhile, the manual dispatch process is complex, the response is slow, the failure high-frequency period is difficult to cope with, resource allocation cannot be carried out from the global view, and the task overload of part of engineers is often caused while other tasks are idle, so that the manpower resource waste is caused. To address the above problems, the industry has begun to attempt to introduce data analysis techniques. In fault diagnosis, some schemes utilize real-time operating data of the equipment to judge whether the state of the equipment deviates from a normal range through a statistical model such as a mahalanobis distance. After the abnormality of the equipment is confirmed, intelligent dispatch is carried out, and an optimization algorithm such as a Hungary algorithm is generally adopted to solve the task allocation problem, so that an allocation scheme with the lowest cost is sought. In the related art, the method for judging the abnormality of the equipment only pays attention to the numerical value change, and the condition of false detection and missing report exists in the abnormality detection process. Meanwhile, in the dispatch process, only the distance or the idle state of engineers is considered, and the manpower resources cannot be comprehensively and comprehensively distributed. In summary, the fault repairing method in the related art has the problems of untimely maintenance and low utilization rate of human resources in the actual operation process. Disclosure of Invention In order to improve timeliness of equipment maintenance and human resource utilization rate, the application provides a fault reporting and repairing method and system based on big data analysis. In a first aspect, the present application provides a fault repairing method based on big data analysis, which adopts the following technical scheme: The fault repairing method based on big data analysis comprises the steps of obtaining historical operation parameters and maintenance records of various fault types of equipment, obtaining historical text vectors of various fault types by using a pre-trained natural language processing model based on text information in the maintenance records; The method comprises the steps of receiving a new repair request, acquiring state deviation degree based on the difference between a real-time operation parameter and each type of historical operation parameter, converting text information in the new repair request into a real-time fault vector, calculating semantic similarity between the real-time fault vector and each type of historical text vector, carrying out weighted fusion on the state deviation degree and the semantic similarity corresponding to the fault type for any fault type to obtain a comprehensive evaluation index of the fault type, and determining the fault type with the highest evaluation index as a diagnosis result of current repair; The intelligent dispatching step for the repair task according to the diagnosis result comprises the steps of constructing a dispatching cost function, wherein the dispatching cost function comprises the estimated time cost of arriving at a task place by an engineer, the matching cost of the skill of the engineer and the task requirement and the current workload cost of the engineer, and determining an allocation scheme for minimizing the dispatching cost by adopting a preset allocation algorithm. In the application, the state deviation degree is calculated through the quantized data operated by the equipment, and the semantic similarity is calculated through text information submitted by a user. And then, the state deviation degree and the semantic similarity are weighted and fused, so that the double-channel and high-precision comprehensive diagnosis of the fault type is realized, and the misjudgment rate is greatly reduced. Further, a multi-dimensional dispatch cost function is constructed through multiple aspects of time of engineers to task sites, skills mastered by the engineers, workload of the engineer