An intelligent recommended treatment method for guide type locomotive fault rapid entry

Abstract

The application discloses an intelligent recommended treatment method for guide type locomotive fault rapid entry, relates to the field of FXN3C type locomotive fault collection technology, and comprises the following steps: step S1, initializing a locomotive hierarchical fault knowledge base KB in advance, and deploying the locomotive hierarchical fault knowledge base KB to a cloud database, and synchronously caching a lightweight knowledge base copy in a mobile terminal. With the linkage mechanism of local offline caching and cloud asynchronous synchronization, the application can adapt to unstable network use scenes in a maintenance site, guarantees complete storage of fault collection data, and does not cause information loss; based on a similarity matching model constructed according to a locomotive equipment feature vector and a fault feature vector, the application can automatically search an adaptive fault treatment template in the knowledge base and complete pushing, thereby providing professional maintenance reference for maintenance personnel and reducing fault judgment deviation.

Description

Technical Field

[0001] This invention relates to the field of FXN3C locomotive fault acquisition technology, and more specifically, to a guided intelligent recommendation and handling method for rapid entry of locomotive faults. Background Technology

[0002] The FXN3C diesel locomotive is the main traction equipment in domestic locomotive depots. During long-term, high-intensity operation, various systems, including the diesel engine, brakes, electrical system, and running gear, are prone to frequent malfunctions. On-site maintenance personnel need to record fault information in real time and maintain maintenance logs to provide basic data support for locomotive condition assessment, periodic preventative maintenance, and fault tracing analysis. Currently, in the process of digital transformation of locomotive maintenance, many depots still use paper registers and local Excel files for fault recording, while a few depots have introduced general-purpose equipment maintenance mini-programs to assist in information collection. Digital data collection is gradually becoming more widespread. Mobile on-site data collection, with its advantages of portability and instant entry, has become the mainstream development direction for locomotive fault information collection. The industry is also gradually adopting collaborative filtering and feature matching algorithms to assist in recommending maintenance solutions, and building knowledge bases based on historical fault data to assist new maintenance personnel.

[0003] Existing fault registration methods and general digital data entry systems suffer from several unavoidable technical flaws: First, they lack a dedicated hierarchical fault knowledge base and mandatory guided input constraint logic specific to the FXN3C locomotive. Maintenance personnel can freely fill in fault texts without unified coding or classification labels, leading to inconsistent recording standards across different shifts and maintenance personnel, generating heterogeneous fault data that cannot be directly used for automated statistical analysis. Second, the traditional data entry process lacks a step-by-step mandatory guidance mechanism, with an average of five minutes spent filling in complete fault information, resulting in significant downtime. Furthermore, the general-purpose app lacks a local offline caching mechanism, making data easily lost when there is no network connection at the maintenance site. Third, existing... The current maintenance solution matching method relies solely on simple keyword searches, lacking a quantitative matching model based on locomotive feature vectors and fault feature vectors. It also lacks weighted association calculation logic based on the frequency of historical fault occurrences, resulting in low accuracy in matching similar fault handling templates and failing to effectively reduce the probability of misjudgments by new staff. Fourth, the existing knowledge base lacks a self-iterative optimization loop; new fault records cannot be automatically updated with new fault association weights, hindering continuous improvement in recommendation matching effectiveness over long-term use. Fifth, the general cloud synchronization mechanism lacks automatic deduplication logic based on unique fault codes, leading to duplicate records due to concurrent uploads from multiple terminals, significantly increasing the manual workload for backend data processing. These multiple shortcomings combined make it difficult for existing technologies to simultaneously meet the comprehensive business needs of rapid locomotive fault entry, datafication, intelligent assisted maintenance, stable storage under weak network conditions, and autonomous knowledge base optimization.

[0004] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0005] To address the problems in related technologies, this invention proposes a guided intelligent recommendation and handling method for rapid entry of locomotive faults, in order to overcome the aforementioned technical problems existing in the existing related technologies.

[0006] The technical solution of this invention is implemented as follows:

[0007] A guided, intelligent recommendation and handling method for rapid entry of locomotive faults includes the following steps:

[0008] Step S1: Pre-initialize the locomotive hierarchical fault knowledge base KB and deploy the locomotive hierarchical fault knowledge base KB to the cloud database, and synchronously cache a lightweight knowledge base copy on the mobile terminal;

[0009] Step S2: Start the wizard-guided step-by-step entry of node sequences. The process involves sequentially entering the locomotive number, selecting the primary fault category, and entering the fault location and symptoms. The generated fault form is then stored in the mobile offline cache fault form dataset D. local ;

[0010] Step S3: Construct a matching model between locomotive equipment feature vectors and fault feature vectors based on a similarity function, and traverse all fault metadata F in the knowledge base. i Calculate the overall similarity score, and push faults and corresponding fault handling templates based on the overall similarity score;

[0011] Step S4: After the mobile device detects network connectivity, it will cache the fault form dataset D offline. local Batch asynchronous upload to cloud fault ledger dataset D cloud Based on the unique fault code, the ledger is merged and automatically deduplicated; the fault weight matrix W is updated according to the newly added fault records, the cloud locomotive hierarchical fault knowledge base KB is updated to a lightweight copy on the mobile device, and the offline cached data that has been synchronized on the mobile device is cleared.

[0012] Wherein, the locomotive hierarchical fault knowledge base KB is represented as: Where C is the set of first-level fault categories, represented as: In the formula, c1 represents the diesel engine system, c2 represents the auxiliary system, c3 represents the electrical system, c4 represents the braking system, c5 represents the running gear system, c6 represents the vehicle safety protection system, and c7 represents other faults; F represents the complete set of fault metadata. , where n is the total number of fault entries, and the metadata of a single fault entry is... In the formula, A unique code for the fault; This fault is classified as a primary category. ; Code the specific parts of the fault; The text describes the fault symptoms; This is a template text for emergency response. T represents the cumulative number of times this fault has occurred in history; T is the set of fault handling templates used to map each fault. W is the fault weight matrix, used to store the pairwise correlation weights between all faults. .

[0013] Among them, the initial calculation , is represented as:

[0014] ;

[0015] In the formula, Fault F i With fault F j The correlation weight coefficient; For fault F i Historical cumulative number of occurrences; Fault F j Historical cumulative number of occurrences; This represents the total frequency of all faults occurring within the knowledge base. This is the weighting amplification factor for the same model.

[0016] Among them, the startup wizard-style step-by-step entry node sequence This includes the following steps:

[0017] Step S201: Execute input node Z1, indicating that locomotive number L is to be forcibly entered. num During data entry, a locomotive feature vector V is generated in real time. L , represented as ;in, For locomotive number characteristic values; This is the model code of the locomotive, which is fixed as the FXN3C code value; Number the maintenance team; This is the timestamp of the incident.

[0018] Step S202: Execute the data entry node Z2, indicating that the first-level fault category is selected from the drop-down menu. ;

[0019] Step S203: Execute input node Z3, indicating that the detailed fault location L is input. loc Fault symptoms;

[0020] Step S204: Execute input node Z4, which means temporarily saving the current form to the offline cache dataset D. localWhen there is no network at the maintenance site, the form is permanently saved locally on the mobile device; when the network is normal, it directly jumps to step S3.

[0021] The step of pushing faults and corresponding fault handling templates based on comprehensive similarity ranking includes the following steps:

[0022] Step S301: Extract the currently entered vehicle model and fault location parameters to generate the locomotive equipment feature vector V. L Retrieve fault feature vectors from the knowledge base In the formula, The fault is coded as a first-level classification. Code the specific parts of the fault; This refers to the historical frequency of the fault. For fault Fi, there is an array of all associated fault weights.

[0023] Step S302: Traverse all fault metadata F in the knowledge base. i Calculate the overall similarity of each fault. , represented as:

[0024] ;

[0025] In the formula, For the current input fault and knowledge base fault F i The overall similarity; Assign weighting coefficients; Match vehicle models with similarity scores; Match similarity scores to faulty parts; Fault F i Average association weight;

[0026] Step S303: Sort by Sim value in descending order and output the preset fault and corresponding handling template T. i A pop-up window is displayed.

[0027] The process of clearing the offline cache data that has already been synchronized on the mobile device includes the following steps:

[0028] S401, After the mobile device detects the network connectivity status, it will cache the fault form dataset D offline. local Batch upload to cloud fault ledger dataset D cloud The system identifies duplicate fault records by using a unique fault code and automatically removes duplicate ledger data.

[0029] S402, extract the classification and sub-location information of the newly added faults, iteratively update the fault weight matrix W, and automatically increase the association weight between the current fault and faults of the same classification and location for each completed fault entry. The weight of a single update is expressed as follows:

[0030] ;

[0031] In the formula, The updated association weight after adding a new fault record; The original weights before the update; This serves as a correction factor for knowledge base iteration. Total failure frequency of the knowledge base;

[0032] S403, the cloud will distribute and synchronize the updated locomotive hierarchical fault standard knowledge base KB to the lightweight knowledge base copy on the mobile device. After synchronization, the offline cached fault form dataset D on the mobile device will be cleared. local This frees up local storage resources on mobile devices.

[0033] The beneficial effects of this invention are:

[0034] This invention, by establishing a hierarchical and unified fault classification system coupled with mandatory step-by-step data entry logic, can standardize the fault information recording format from the source, eliminate data inconsistencies caused by manual data entry, and form a regular and unified structured fault ledger. Utilizing a linkage mechanism of local offline caching and cloud-based asynchronous synchronization, it can adapt to usage scenarios with unstable networks at maintenance sites, ensuring the complete retention of fault data and preventing information loss. Based on a similarity matching model constructed from locomotive equipment feature vectors and fault feature vectors, it can automatically retrieve suitable fault handling templates from the knowledge base and push them to maintenance personnel, providing professional maintenance references and reducing fault judgment errors.

[0035] Meanwhile, this invention can continuously iterate and update the fault weight matrix based on newly entered fault records, enabling the locomotive hierarchical fault knowledge base to have autonomous optimization capabilities. With continuous use, the accuracy of fault matching recommendations is constantly improved. The standardized fault ledger stored in the cloud can also directly provide a complete and reliable data foundation for the full-cycle fault statistics of locomotives, the analysis of component fault trends, and preventive maintenance work. Overall, it simplifies the entire fault registration process, reduces the manual workload of on-site maintenance information collection and back-end data processing, and is suitable for the routine digital management of diesel locomotive faults in locomotive depots. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1This is a flowchart illustrating an intelligent recommended handling method for rapid recording of locomotive faults according to an embodiment of the present invention. Detailed Implementation

[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0039] According to an embodiment of the present invention, a guided intelligent recommendation and handling method for rapid entry of locomotive faults is provided.

[0040] like Figure 1 As shown, the intelligent recommendation and handling method for rapid entry of locomotive faults according to an embodiment of the present invention includes the following steps:

[0041] Step S1, pre-initialize the locomotive hierarchical fault knowledge base KB, represented as ;

[0042] Where C is the set of 7 major categories of first-level fault classifications, represented as: In the formula, c1 is the diesel engine system, c2 is the auxiliary system, c3 is the electrical system, c4 is the braking system, c5 is the running gear system, c6 is the vehicle safety protection system, and c7 is other faults;

[0043] Where F represents the complete set of fault metadata, , where n is the total number of fault entries, and the metadata of a single fault entry is... In the formula, A unique code for the fault; This fault is classified as a primary category. ; Code the specific parts of the fault; The text describes the fault symptoms; This is a template text for emergency response. This represents the cumulative number of times this fault has occurred throughout history.

[0044] Where T is the set of fault handling templates, used to map each fault handling template. ;

[0045] Where W is the fault weight matrix, used to store the pairwise correlation weights between all faults. ;

[0046] Specifically, it includes the following steps:

[0047] Pre-divide the faults into 7 primary categories C, and construct the metadata F for each fault.i Bind to a dedicated processing template T i ;

[0048] Initialize the fault weight matrix W, and assign initial values ​​to cnt based on historical maintenance data. i Calculate the initial , is represented as:

[0049] ;

[0050] In the formula, For fault F i With fault F j The correlation weight coefficient; For fault F i Historical cumulative number of occurrences; For fault F j Historical cumulative number of occurrences; This represents the total frequency of all faults occurring within the knowledge base. This is a weighting factor for the same model, with a fixed value of α=1.2. It is only used on the FXN3C locomotive to distinguish it from other models.

[0051] Specifically, two types of faults occur simultaneously and frequently within the same locomotive and the same work crew. The higher the value, the stronger the correlation with the fault, and the higher the recommendation priority.

[0052] In this technical solution, during implementation, the locomotive hierarchical fault knowledge base KB is pre-deployed to the cloud development cloud database, and a lightweight copy of the knowledge base is cached locally on the mobile device.

[0053] Step S2: Start the wizard-guided step-by-step entry of node sequences. The process involves sequentially entering the locomotive number, selecting the primary fault category, entering the fault location and symptoms, caching the form locally, and then storing the entered form in the mobile offline cache fault form dataset D. local ;

[0054] Specifically, it includes the following steps:

[0055] Step S201: Execute input node Z1, that is, force input of locomotive number L. num During data entry, a locomotive feature vector V is generated in real time. L , represented as ;in, For locomotive number characteristic values; This is the model code of the locomotive, which is fixed as the FXN3C code value; Number the maintenance team; This is the timestamp of the incident.

[0056] Step S202: Execute the data entry node Z2, that is, select the first-level fault category from the drop-down menu. Free text input is prohibited; without selection, the next node cannot be accessed.

[0057] Step S203: Execute input node Z3, that is, input the detailed fault location L. loc Fault symptoms, system extracts location features;

[0058] Step S204: Execute input node Z4, which temporarily saves the current form to the offline cache dataset D. local ;

[0059] Specifically, during implementation, if there is no network on site, the data will be permanently stored locally on the mobile device; if the network is normal, proceed directly to the next step S3.

[0060] Step S3, based on the similarity function The intelligent matching and processing template includes the following steps:

[0061] Step S301: Extract the currently entered vehicle model and fault location parameters to generate the locomotive equipment feature vector V. L Retrieve fault feature vectors from the knowledge base In the formula, The fault is coded as a first-level classification. Code the specific parts of the fault; This refers to the historical frequency of the fault. For fault Fi, there is an array of all associated fault weights.

[0062] Step S302: Traverse all fault metadata F in the knowledge base. i Calculate the overall similarity of each fault. , is represented as:

[0063] ;

[0064] In the formula, For the current input fault and knowledge base fault F i The overall similarity score ranges from [0,1], with higher scores indicating higher recommendation priority. Assign weighting coefficients; The similarity score is matched to the model. For the FXN3C model, the value is 1 if it is a complete match and 0 if it is not a match. The similarity score is calculated based on the location of the fault. A score of 1 is given if the user-entered location is exactly the same as the location, 0.5 is given if the major category matches, and 0 is given if there is no match. The average association weight of fault Fi is expressed as:

[0065] ;

[0066] Step S303: Sort by Sim value in descending order and output the Top 5 faults and their corresponding handling templates T. i Pop-up display;

[0067] Specifically, a preset recommendation ranking rule is established, which involves calculating all F... i The Sim values ​​are sorted from largest to smallest, and the top 5 faults are pushed to the pop-up window for handling templates. Maintenance personnel can select the recommended templates with one click to fill in the fault handling content, which greatly reduces the time for text input and editing.

[0068] Step S4: The network is restored asynchronously and synchronously, and the locomotive hierarchical fault knowledge base KB is iteratively updated;

[0069] Specifically, it includes the following steps:

[0070] Step S401: After detecting network connectivity, based on the constraint relationship between the offline cache and the cloud-synchronized dataset... The local cached fault form dataset D under weak network conditions local Batch upload cloud fault ledger dataset D cloud Complete the merging and deduplication of ledgers;

[0071] in, Store the fault log dataset in the cloud; For a set of duplicate fault records, categorized by ID i Unique encoding for deduplication.

[0072] Step S402: Extract the classification and location information of the newly added fault, and update the weight matrix W, as follows:

[0073] ;

[0074] In the formula, The updated association weight after adding a new fault record; The original weights before the update; This serves as a correction factor for knowledge base iteration. Total failure frequency of the knowledge base;

[0075] Specifically, each time a fault is entered, the association weight between this fault and faults of the same category and location is automatically increased, leading to continuous improvement in subsequent recommendation accuracy. The locomotive hierarchical fault knowledge base (KB) updated in the cloud is copied to a local copy on the mobile device, completing self-optimization of the knowledge base; the locally cached data that has already been synchronized is cleared. local This frees up storage resources on mobile devices.

[0076] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Those skilled in the art, upon considering the disclosure in the specification and embodiments, will readily conceive of other embodiments of this disclosure. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0077] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A guided, intelligent recommendation and handling method for rapid entry of locomotive faults, characterized in that, Includes the following steps: Step S1: Pre-initialize the locomotive hierarchical fault knowledge base KB and deploy the locomotive hierarchical fault knowledge base KB to the cloud database, and synchronously cache a lightweight knowledge base copy on the mobile terminal; Step S2: Start the wizard-guided step-by-step entry of node sequences. The process involves sequentially entering the locomotive number, selecting the primary fault category, and entering the fault location and symptoms. The generated fault form is then stored in the mobile offline cache fault form dataset D. local ; Step S3: Construct a matching model between locomotive equipment feature vectors and fault feature vectors based on a similarity function, and traverse all fault metadata F in the knowledge base. i Calculate the overall similarity score, and push faults and corresponding fault handling templates based on the overall similarity score; Step S4: After the mobile device detects network connectivity, it will cache the fault form dataset D offline. local Batch asynchronous upload to cloud fault ledger dataset D cloud Based on the unique fault code, the ledger is merged and automatically deduplicated; the fault weight matrix W is updated according to the newly added fault records, the cloud locomotive hierarchical fault knowledge base KB is updated to a lightweight copy on the mobile device, and the offline cached data that has been synchronized on the mobile device is cleared.

2. The intelligent recommendation and handling method for rapid input of locomotive faults according to claim 1, characterized in that, The locomotive hierarchical fault knowledge base KB is represented as follows: Where C is the set of first-level fault categories, represented as: In the formula, c1 represents the diesel engine system, c2 represents the auxiliary system, c3 represents the electrical system, c4 represents the braking system, c5 represents the running gear system, c6 represents the vehicle safety protection system, and c7 represents other faults; F represents the complete set of fault metadata. , where n is the total number of fault entries, and the metadata of a single fault entry is... In the formula, A unique code for the fault; This fault is classified as a primary category. ; Code the specific parts of the fault; The text describes the fault symptoms; This is a template text for emergency response. T represents the cumulative number of times this fault has occurred in history; T is the set of fault handling templates used to map each fault. W is the fault weight matrix, used to store the pairwise correlation weights between all faults. .

3. The intelligent recommendation and handling method for rapid input of locomotive faults according to claim 2, characterized in that, Calculate the initial , represented as: ； In the formula, Fault F i With fault F j The correlation weight coefficient; Fault F i Historical cumulative number of occurrences; Fault F j Historical cumulative number of occurrences; This represents the total frequency of all faults occurring within the knowledge base. This is the weighting amplification factor for the same model.

4. The intelligent recommendation and handling method for rapid input of locomotive faults according to claim 1, characterized in that, The startup wizard-style step-by-step data entry node sequence This includes the following steps: Step S201: Execute input node Z1, indicating that locomotive number L is to be forcibly input. num During data entry, a locomotive feature vector V is generated in real time. L , represented as ;in, For locomotive number characteristic values; This is the model code of the locomotive, which is fixed as the FXN3C code value; Number the maintenance team; This is the timestamp of the incident.

5. Step S202: Execute the data entry node Z2, indicating that the first-level fault category is selected from the drop-down menu. ; Step S203: Execute input node Z3, indicating that the detailed fault location L is input. loc Fault symptoms; Step S204: Execute input node Z4, which means temporarily saving the current form to the offline cache dataset D. local ;in, When there is no network at the maintenance site, the form is permanently saved locally on the mobile device. When the network is normal, it will directly jump to step S3.

6. The intelligent recommendation and handling method for rapid input of locomotive faults according to claim 1, characterized in that, The method of pushing faults and corresponding fault handling templates based on comprehensive similarity ranking includes the following steps: Step S301: Extract the currently entered vehicle model and fault location parameters to generate the locomotive equipment feature vector V. L Retrieve fault feature vectors from the knowledge base In the formula, The fault is coded as a first-level classification. Code the specific parts of the fault; This refers to the historical frequency of the fault. For fault Fi, there is an array of all associated fault weights. Step S302: Traverse all fault metadata F in the knowledge base. i Calculate the overall similarity of each fault. , represented as: ； In the formula, For the current input fault and knowledge base fault F i The overall similarity; Assign weighting coefficients; Match vehicle models with similarity scores; Match similarity scores to faulty parts; Fault F i Average association weight; Step S303: Sort by Sim value in descending order and output the preset fault and corresponding handling template T. i A pop-up window is displayed.

7. The intelligent recommendation and handling method for rapid input of locomotive faults according to claim 1, characterized in that, Clearing the offline cache data that has been synchronized on the mobile device includes the following steps: S401, After the mobile device detects the network connectivity status, it will cache the fault form dataset D offline. local Batch upload to cloud fault ledger dataset D cloud The system identifies duplicate fault records by using a unique fault code and automatically removes duplicate ledger data. S402, extract the classification and sub-location information of the newly added faults, iteratively update the fault weight matrix W, and automatically increase the association weight between the current fault and faults of the same classification and location for each completed fault entry. The weight of a single update is expressed as follows: ； In the formula, The updated association weight after adding a new fault record; The original weights before the update; This serves as a correction factor for knowledge base iteration. Total failure frequency of the knowledge base; S403, the cloud will distribute and synchronize the updated locomotive hierarchical fault standard knowledge base KB to the lightweight knowledge base copy on the mobile device. After synchronization, the offline cached fault form dataset D on the mobile device will be cleared. local This frees up local storage resources on mobile devices.

Images