Intelligent troubleshooting method, system and equipment based on text recognition and medium

Abstract

The invention relates to an intelligent troubleshooting method and system based on text recognition, equipment and a medium. The method comprises the following steps: acquiring an equipment fault description text in an industrial system; based on the equipment fault description text, adopting a pre-trained word vector model to extract keywords, and forming sentence vectors; and performing similarity comparison on the sentence vector and a pre-generated fault standardized description sentence vector, and outputting a plurality of fault standardized description sentence vectors with the maximum similarity value to complete a troubleshooting process. Compared with the prior art, the method has the advantages of quickly positioning the equipment fault reason, reducing the troubleshooting time, improving the operation and maintenance efficiency and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of troubleshooting, and in particular to an intelligent troubleshooting method, system, device and medium based on text recognition. Background Art

[0002] With the continuous advancement of information technology, equipment maintenance personnel in the industrial field have begun to use information-based handheld terminal devices to diagnose and handle equipment failures. These terminal devices are usually equipped with an interactive electronic technical manual (IETM) system to assist in fault diagnosis.

[0003] However, traditional IETM systems have certain limitations in the fault diagnosis process. They mainly rely on the operation and maintenance personnel to input precise fault description statements, which are matched with the preset fault library to determine the cause of the fault. This method lacks the ability to understand and process natural language, and cannot achieve fuzzy matching based on semantic understanding, resulting in difficulty in quickly and accurately locating the root cause of the fault when faced with complex or non-standardized fault descriptions. This not only affects the efficiency of fault handling, but may also delay maintenance work, increase downtime, and have a negative impact on production efficiency and economic benefits.

[0004] Therefore, in order to improve the accuracy and response speed of fault diagnosis, it is necessary to develop and apply more advanced technologies to achieve intelligent understanding and analysis of fault descriptions, so as to quickly locate and repair equipment faults. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent fault troubleshooting method, system, device and medium based on text recognition to improve the fault location speed.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] An intelligent troubleshooting method based on text recognition includes the following steps:

[0008] Get equipment fault description text in industrial systems;

[0009] Based on the equipment fault description text, a pre-trained word vector model is used to extract keywords and form a sentence vector;

[0010] The sentence vector is compared with the pre-generated standardized fault description sentence vector for similarity, and a plurality of standardized fault description sentence vectors with the largest similarity values ​​are output to complete the fault troubleshooting process.

[0011] Furthermore, the word vector model includes a word segmenter and a pre-trained language sub-model based on Transformer, and the execution steps of the word vector model include:

[0012] Based on the equipment fault description text, combined with the industrial field text dictionary, a word segmenter is used to perform word segmentation and extract keywords;

[0013] Based on the extracted keywords, a Transformer-based pre-trained language sub-model is used to perform encoding and fusion processing in sequence to obtain a sentence vector.

[0014] Furthermore, the word segmenter adopts Jieba word segmenter, and the word segmentation result includes word segmentation phrases and part-of-speech tags.

[0015] Furthermore, the jieba word segmenter integrates the posseg.lcut function and the extract_tags function, and the execution steps of the jieba word segmenter include:

[0016] The posseg.lcut function is used to segment the equipment fault description text according to the industrial field text dictionary to obtain a segmentation result;

[0017] According to the word segmentation results, the extract_tags function is used to extract keywords.

[0018] Furthermore, the word segmentation process of the word segmenter also includes:

[0019] Filter out noise words and stop words.

[0020] Furthermore, the Transformer-based pre-trained language sub-model includes an encoder and a fusion module including a self-attention mechanism, and the execution steps of the Transformer-based pre-trained language sub-model include:

[0021] Based on the keywords, an encoder including a self-attention mechanism is used for encoding, and the attention mechanism is used to calculate the context representation of each keyword to obtain a keyword vector;

[0022] Based on the keyword vector, a fusion module is used to perform fusion to obtain a sentence vector.

[0023] Furthermore, in the similarity comparison process, cosine similarity is used for comparison, and the calculation expression of the cosine similarity is:

[0024] cosθ=(A·B) / (‖A‖×‖B‖)

[0025] Where θ is the angle between vector A and vector B, A and B are the sentence vector and the pre-generated fault standardized description sentence vector, respectively, and ‖·‖ is the modulus.

[0026] The present invention also provides an intelligent troubleshooting system based on text recognition, comprising:

[0027] Text acquisition unit: used to obtain equipment fault description text in industrial systems;

[0028] A sentence vector forming unit: used to extract keywords based on the device fault description text using a pre-trained word vector model and form a sentence vector;

[0029] Similarity comparison unit: used to compare the similarity between the sentence vector and the pre-generated fault standardized description sentence vector, output multiple fault standardized description sentence vectors with the largest similarity values, and complete the fault troubleshooting process.

[0030] The present invention also provides an electronic device, comprising: one or more processors; a memory; and one or more programs stored in the memory, wherein the one or more programs include instructions for executing the intelligent troubleshooting method based on text recognition as described above.

[0031] The present invention also provides a computer-readable storage medium, comprising one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the intelligent troubleshooting method based on text recognition as described above.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] (1) In order to solve the problem that the prior art requires operation and maintenance personnel to input accurate fault descriptions, the present invention adopts a word vector model for natural language processing. By extracting keywords from the fault description text to form a sentence vector, no accurate fault description is required. Then, a similarity comparison is performed with the pre-generated fault standardized description sentence vector to achieve accurate matching. The returned multiple fault standardized description sentence vectors with the largest similarity values ​​enable the operation and maintenance personnel to quickly locate the cause of the equipment failure.

[0034] (2) The present invention uses a Transformer-based pre-trained language sub-model to process keywords, in which the attention mechanism is used to focus on the key information in the text, so that the keyword vector can more accurately capture the semantic features in the text. Then, through keyword vector fusion, the obtained sentence vector can comprehensively and accurately represent the semantic information of the fault description.

[0035] (3) The present invention has fast and accurate fault location, and can improve operation and maintenance efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION

[0037] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0038] Example 1

[0039] This embodiment provides an intelligent troubleshooting method based on text recognition, such as Figure 1 As shown, the method comprises the following steps:

[0040] Step 1: The operation and maintenance personnel enter the equipment fault description text in the industrial system.

[0041] In this example, the equipment fault description text written in Chinese by the operation and maintenance personnel is used to perform text semantic recognition using a pre-trained word vector model to form a sentence vector. The word vector model includes a word segmenter and a pre-trained language sub-model based on Transformer.

[0042] Step 2: Segment the text based on the industrial text dictionary and word segmenter and extract keywords.

[0043] Based on the industrial text dictionary, the posseg.lcut function in the jieba word segmenter is used to implement the word segmentation phrases and part-of-speech tags of Chinese text to support subsequent text analysis, natural language processing or information retrieval tasks. Then, the extract_tags function in the jieba word segmenter is used to extract keywords, including black screen, LED, no video, asynchronous frequent fill light, no automatic broadcast, etc.

[0044] In order to avoid being affected by noise words and stop words in the text, during the word segmentation process, noise words and stop words in the text are filtered out according to the noise word library and the stop word library.

[0045] Step 3: Use the Transformer-based pre-trained language sub-model to convert the extracted keywords into sentence vectors. This process relies on two key components in the Transformer architecture: the encoder with self-attention mechanism and the fusion module.

[0046] 1. Encoder with self-attention mechanism

[0047] 1.1 Self-Attention Mechanism

[0048] The self-attention mechanism is at the heart of the Transformer architecture. It allows the model to focus on other elements in the sequence while processing each element in the sequence. This mechanism works by calculating the relevance score between each element (in this case, the keyword) and other elements, thereby generating a contextually relevant representation.

[0049] 1.2 Working Principle of Encoder

[0050] In the encoder, each keyword is first embedded into a fixed-size vector space to obtain an initial keyword vector. These initial vectors are then fed into the self-attention layer, where each vector is compared with all other vectors in the sequence to calculate the correlation between them.

[0051] Relevance calculations usually involve three vectors: query vector (Q), key vector (K), and value vector (V). These vectors are obtained by multiplying the initial vector by different weight matrices. The output of the self-attention layer is obtained by weighted summing all value vectors, with the weights determined by the dot product between the query vector and the key vector.

[0052] This process allows the encoder to capture the relationship between each keyword and other keywords in the sequence, thereby generating a richer contextual representation.

[0053] 2. Fusion Module

[0054] 2.1 Deep Neural Network Architecture

[0055] The fusion module is a deep neural network that receives the keyword vector from the encoder as input and outputs a sentence vector. This network usually contains multiple hidden layers, each of which uses a nonlinear activation function to increase the expressive power of the model.

[0056] 2.2 Fusion Strategy

[0057] In the fusion module, the keyword vectors are further processed to generate sentence vectors. This can be achieved through various strategies, such as:

[0058] Average pooling: Calculate the average of all keyword vectors to get a sentence vector. This method is simple and effective, but it may not capture the complex relationship between keywords.

[0059] Weighted sum: The keywords are weighted and summed according to their importance. The weights can be obtained through training or determined based on some heuristic rules.

[0060] Attention Mechanism: The attention mechanism is used again to determine the contribution of each keyword to the final sentence vector. This approach allows the model to dynamically adjust the weight of each keyword to reflect their importance in the sentence.

[0061] 2.3 Generation of sentence vectors

[0062] Finally, the output of the fusion module is a sentence vector that fully and accurately represents the semantic information of the fault description. This sentence vector can be used for various downstream tasks such as text classification, sentiment analysis, or information retrieval.

[0063] Step 4: Compare the sentence vector with the pre-generated fault standardized description sentence vector by cosine similarity, and return the three fault standardized descriptions with the largest cosine similarity values.

[0064] The specific formula of cosine similarity is:

[0065] cosθ=(A·B) / (‖A‖×‖B‖)

[0066] Where: A and B are two vectors, namely, the sentence vector and the pre-generated standardized fault description sentence vector; A·B represents the dot product of A and B, and the calculation formula is A·B=Σ(a i ×b i ), which is the sum of the products of the corresponding elements of vector A and vector B. ‖A‖ and ‖B‖ represent the modulus (length) of A and B respectively, that is, the Euclidean norm of the vector, and the calculation formula is ‖A‖=sqrt(Σ(a i ^2)), which is the square root of the sum of the squares of the elements of vector A.

[0067] The value range of cosine similarity is between -1 and 1. When the directions of the two vectors are exactly the same (the angle is 0°), the cosine similarity is 1; when the directions of the two vectors are completely opposite (the angle is 180°), the cosine similarity is -1; when the two vectors are orthogonal (the angle is 90° or 270°), the cosine similarity is 0. Therefore, the closer the value of cosine similarity is to 1, the more consistent the directions of the two vectors are, and the more similar they are; conversely, the closer the value is to -1, the more opposite the directions of the two vectors are, and the less similar they are.

[0068] Step 5: According to the returned standardized description, the cause of the failure and the repair method are returned, and the maintenance personnel will gradually troubleshoot the failure based on the returned cause and method.

[0069] The present invention is based on intelligent recognition of text semantics, which helps operation and maintenance personnel to quickly locate the cause of equipment failure, shorten the troubleshooting time, and achieve the goal of reducing costs and increasing efficiency.

[0070] In the embodiment of the present invention, the operation and maintenance personnel can input a fuzzy description of the equipment failure, extract keywords and generate word vectors based on the Jieba word segmentation and Transformer model, further combine the word vectors into sentence vectors, and calculate the cosine similarity with the sentence vectors of the standardized fault description in the historical knowledge base. The most likely fault description is displayed, and the cause and maintenance method of the fault description are displayed, so that the operation and maintenance personnel can quickly locate the fault problem and achieve the purpose of reducing costs and increasing efficiency.

[0071] If the above 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 invention, 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, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0072] Example 2

[0073] This embodiment provides an intelligent troubleshooting system based on text recognition, the system comprising:

[0074] Text acquisition unit: used to obtain equipment fault description text in industrial systems;

[0075] A sentence vector forming unit: used to extract keywords based on the device fault description text using a pre-trained word vector model and form a sentence vector;

[0076] Similarity comparison unit: used to compare the similarity between the sentence vector and the pre-generated fault standardized description sentence vector, output multiple fault standardized description sentence vectors with the largest similarity values, and complete the fault troubleshooting process.

[0077] The rest is the same as in Example 1.

[0078] Example 3

[0079] This embodiment provides an electronic device, including: one or more processors; a memory; and one or more programs stored in the memory, wherein the one or more programs include instructions for executing the intelligent troubleshooting method based on text recognition as described in Example 1.

[0080] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0081] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0082] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0083] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0084] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0085] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An intelligent troubleshooting method based on text recognition, characterized in that: The following steps are involved: Get equipment fault description text in industrial systems; Based on the equipment fault description text, a pre-trained word vector model is used to extract keywords and form a sentence vector; The sentence vector is compared with the pre-generated standardized fault description sentence vector for similarity, and a plurality of standardized fault description sentence vectors with the largest similarity values ​​are output to complete the fault troubleshooting process.

2. According to claim 1, an intelligent troubleshooting method based on text recognition is characterized in that: The word vector model includes a word segmenter and a pre-trained language sub-model based on Transformer. The execution steps of the word vector model include: Based on the equipment fault description text, combined with the industrial field text dictionary, a word segmenter is used to perform word segmentation and extract keywords; Based on the extracted keywords, a Transformer-based pre-trained language sub-model is used to perform encoding and fusion processing in sequence to obtain a sentence vector.

3. The intelligent troubleshooting method based on text recognition according to claim 2 is characterized in that: The word segmenter adopts the jieba word segmenter, and the word segmentation result includes a word segmentation phrase and a part-of-speech tag.

4. The intelligent troubleshooting method based on text recognition according to claim 3 is characterized in that: The jieba word segmenter integrates the posseg.lcut function and the extract_tags function. The execution steps of the jieba word segmenter include: The posseg.lcut function is used to segment the equipment fault description text according to the industrial field text dictionary to obtain a segmentation result; According to the word segmentation results, the extract_tags function is used to extract keywords.

5. The intelligent troubleshooting method based on text recognition according to claim 2 is characterized in that: The word segmentation process of the word segmenter also includes: Filter out noise words and stop words.

6. The intelligent troubleshooting method based on text recognition according to claim 2 is characterized in that: The Transformer-based pre-trained language sub-model includes an encoder and a fusion module including a self-attention mechanism, and the execution steps of the Transformer-based pre-trained language sub-model include: Based on the keywords, an encoder including a self-attention mechanism is used for encoding, and the attention mechanism is used to calculate the context representation of each keyword to obtain a keyword vector; Based on the keyword vector, a fusion module is used to perform fusion to obtain a sentence vector.

7. The intelligent troubleshooting method based on text recognition according to claim 1 is characterized in that: In the similarity comparison process, cosine similarity is used for comparison, and the calculation expression of the cosine similarity is: cosθ=(A·B) / (‖A‖×‖B‖) Where θ is the angle between vector A and vector B, A and B are the sentence vector and the pre-generated fault standardized description sentence vector, respectively, and ‖·‖ is the modulus.

8. An intelligent troubleshooting system based on text recognition, characterized in that: include: Text acquisition unit: used to obtain equipment fault description text in industrial systems; A sentence vector forming unit: used to extract keywords based on the device fault description text using a pre-trained word vector model and form a sentence vector; Similarity comparison unit: used to compare the similarity between the sentence vector and the pre-generated fault standardized description sentence vector, output multiple fault standardized description sentence vectors with the largest similarity values, and complete the fault troubleshooting process.

9. An electronic device, characterized in that: include: one or more processors; Memory; and One or more programs stored in a memory, wherein the one or more programs include instructions for executing the intelligent troubleshooting method based on text recognition as claimed in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: It includes one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the intelligent troubleshooting method based on text recognition as described in any one of claims 1-7.

Images