Line equipment unit maintenance decision-making method based on clustering analysis

Abstract

The invention relates to a line equipment unit maintenance decision-making method based on clustering analysis, which is characterized in that when a daily maintenance plan of railway line equipment is formulated, the line equipment unit maintenance decision-making method based on clustering analysis is adopted on the basis of effectively utilizing the existing detection data to evaluate the comprehensive state of the line equipment unit. The maintenance capability of field personnel, tools and tools and the like is fully utilized, a next-month or next-week daily maintenance plan is automatically generated, the influence of subjective factors of personnel is reduced, the pertinence of line equipment maintenance is improved, the line equipment quality is improved, and good economic and practical values are achieved.

Description

Technical Field

[0001] The invention relates to the field of railway line equipment maintenance and repair, and in particular to a line equipment unit maintenance decision method based on cluster analysis. Background Art

[0002] With the improvement of my country's railway transportation capacity, higher requirements have been put forward for the maintenance and repair level of railway line equipment. To this end, my country has proposed a maintenance model for track unit repair, which divides the line equipment into units of a certain length. According to the status evaluation results of the unit equipment, the line equipment unit status is disqualified, qualified, good or divided into Level I, Level II, Level III, and then the maintenance plan is arranged according to the evaluation results. If the units with poor status are arranged based solely on the evaluation results, it is easy to cause waste of personnel or surplus personnel. In addition, the distance between each maintenance unit is far, and the line operation points are scattered, which brings certain hidden dangers to traffic, personal safety and other protections. The site mainly relies on experienced personnel in the workshop or work area to formulate daily maintenance plans for the next month and next week based on experience, lacking a scientific and standardized plan-making method.

[0003] Technical problems and defects in the formulation of existing railway routine maintenance plans

[0004] 1. Rely on manual experience:

[0005] The current daily railway maintenance plan formulation mainly relies on human experience and judgment. In this case, the scientificity and rationality of the maintenance plan largely depend on the experience and expertise of the maintenance personnel. However, due to the different professional backgrounds and experiences of each maintenance personnel, the maintenance plan may be subjective and diverse, making it difficult to achieve standardization and normalization.

[0006] 2. Excessive maintenance and insufficient maintenance coexist:

[0007] Due to the lack of scientific evaluation and monitoring methods, the daily maintenance plan of railways is prone to both over-maintenance and under-maintenance. On the one hand, some sections may waste resources due to over-frequent maintenance; on the other hand, some sections may be in poor condition but no maintenance is scheduled, which not only affects the service life of railway equipment, but also may leave safety hazards.

[0008] 3. Difficulty in obtaining data:

[0009] When formulating maintenance plans, due to the lack of effective monitoring and detection methods, maintenance personnel find it difficult to obtain the real-time status and fault information of equipment, which leads to a lack of basis for maintenance decisions and increases the blindness and uncertainty of maintenance.

[0010] defect

[0011] 1. Inefficiency:

[0012] Maintenance plans based on manual experience are inefficient and difficult to meet the needs of railway operations. At the same time, over-maintenance and under-maintenance not only affect the service life of railway equipment, but may also lead to delays in train operations and other safety issues.

[0013] 2. Safety hazards:

[0014] Due to the lack of scientific and reasonable maintenance plans, railway equipment may be in poor condition without maintenance, which will leave safety hazards and affect the lives of passengers.

[0015] 3. Waste of resources:

[0016] Over-maintenance and under-maintenance will lead to waste of resources and increase the cost of railway operation. At the same time, due to unreasonable maintenance plans, the workload of maintenance personnel may be unbalanced, affecting the overall efficiency of maintenance work.

[0017] In summary, the existing method of formulating daily railway maintenance plans has obvious defects and problems, and new technologies and methods need to be introduced to improve the scientificity and rationality of maintenance plans and ensure the stable operation of railway equipment and the safety of passengers. Summary of the invention

[0018] The purpose of the present invention is to provide a line equipment unit maintenance decision method based on cluster analysis, and its technical solution is: based on the comprehensive status evaluation of the line equipment unit section, firstly, a preliminary maintenance unit is proposed according to the threshold value of the comprehensive status evaluation index, and then according to the maintenance and repair capacity of the workshop or work area and the interference analysis with the mechanized maintenance sections such as large machine tamping and large machine screening, the preliminary maintenance units are clustered by using the cluster analysis method, and the originally scattered maintenance units are optimized and merged to avoid a large number of on-site maintenance points, reduce the pressure of on-site protection, and reduce the safety hazards of personnel.

[0019] Specific technical solutions:

[0020] A line equipment unit maintenance decision method based on cluster analysis comprises the following steps:

[0021] Step 1: Preliminarily screen out units exceeding the maintenance threshold according to the comprehensive status evaluation results of line equipment units to form preliminary maintenance units;

[0022] Step 2: According to the mechanized maintenance work plan of large machine tamping, large machine screening and rail replacement in the next month or next week, conduct a coherence analysis with the preliminary maintenance units, eliminate the mechanized maintenance units, and select the maintenance units;

[0023] Step 3: According to the on-site maintenance and repair operation capacity, skylight settings, etc., on the basis of the comprehensive status evaluation of the line equipment unit, set the iteration step length ε, consider the differences of different levels of lines, and make the line equipment maintenance unit equal to the actual maintenance and repair capacity Thus, an optimized maintenance unit is obtained;

[0024] L 能 ——Monthly maintenance and repair capabilities.

[0025] ——Critical maintenance value of the jth category line.

[0026] j——The number of over-limit sections in the same grade of lines.

[0027] j represents the line grade, with values ​​ranging from 1 to 3, representing heavy vehicle line, light vehicle line, station line and arrival and departure track respectively;

[0028] Step 4: Considering the on-site construction transfer, protection capabilities, and on-site personnel safety, the distance between maintenance and repair units should not be too long or too scattered. The DBSCAN cluster analysis method is used to form continuous segmented operation units.

[0029] In cluster analysis, the clustering effect evaluation index GF is used to evaluate the clustering effect and to solve the optimal clustering parameters. The larger the GF, the better the clustering effect. The clustering effect evaluation index GF consists of the intra-class similarity function and the out-of-class similarity function:

[0030]

[0031] Intra-class similarity

[0032] Out-of-class similarity

[0033] In the formula,

[0034] r stands for G i The number of points contained in ;

[0035] x j Represents G i The jth point in , 1≤j≤r;

[0036] represents the average value of the points contained in Gi;

[0037] h represents the number of clusters;

[0038] α represents the weight adjustment parameter of Stra(G) and Ster(G), which is determined empirically and is between 0 and 1.

[0039] Furthermore, in step one, the comprehensive status evaluation result of the line equipment unit is obtained through real-time monitoring and data analysis, including but not limited to vibration monitoring, temperature monitoring, wear degree monitoring, etc.

[0040] Furthermore, in step 2, the coherence analysis includes comparing the mechanized maintenance work plan with the location, type and time information of the preliminary maintenance unit to ensure that there is no conflict or duplication between the mechanized maintenance work and the manual maintenance work.

[0041] Furthermore, in step three, the setting of the iteration step is determined based on historical maintenance data and current maintenance resource status to balance maintenance efficiency and resource utilization.

[0042] Furthermore, in step 4, the parameters of the DBSCAN clustering analysis method (such as the minimum sample number MinPts and the neighborhood radius Eps) are determined through experiments or optimization algorithms to obtain the best clustering effect.

[0043] Furthermore, the method further comprises step five: adjusting clustering parameters according to the result of the clustering effect evaluation index GF, and re-performing clustering analysis until a clustering effect that meets the requirements is obtained.

[0044] Furthermore, in step 4, the intra-class similarity function of the clustering effect evaluation index GF is calculated based on measurement methods such as Euclidean distance, Manhattan distance or cosine similarity.

[0045] Furthermore, it also includes step six: according to the formed continuous segmented operation units, a detailed maintenance plan and resource allocation plan are formulated, including the scheduling of maintenance personnel, the preparation of maintenance tools, and the arrangement of maintenance time.

[0046] The present invention adopts the above technical solution, which has the following advantages:

[0047] Based on the comprehensive status assessment of the line equipment unit section, the preliminary maintenance unit is first proposed according to the threshold of the comprehensive status assessment index. Then, according to the maintenance and repair capacity of the workshop or work area and the interference analysis with mechanized maintenance sections such as large-scale tamping and large-scale screening, the preliminary maintenance units are clustered using the cluster analysis method. The originally scattered maintenance units are optimized and merged to avoid a large number of on-site maintenance points, reduce the pressure of on-site protection, and reduce personnel safety hazards.

[0048] Specifically, the DBSCAN clustering analysis method is used to form continuous segmented operation units. The clustering effect evaluation index GF is used to evaluate the clustering effect during cluster analysis to solve the optimal clustering parameters. The larger the GF, the better the clustering effect. The clustering effect evaluation index GF consists of the intra-class similarity function and the out-of-class similarity function. In other words, different from the existing qualitative processing method based on experience, the analysis and evaluation are quantified to achieve the accurate formulation of the daily maintenance plan for railway line equipment.

[0049] When formulating the daily maintenance plan for railway line equipment, a line equipment unit maintenance decision-making method based on cluster analysis is adopted. On the basis of effectively using the existing detection data to evaluate the comprehensive status of the line equipment unit, the maintenance capabilities of on-site personnel, tools and other tools are fully utilized to automatically generate daily maintenance plans for the next month or week. This reduces the influence of personnel's subjective factors, improves the targeted maintenance of line equipment, improves the quality of line equipment, and has good economic and practical value. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a unit maintenance decision flow chart;

[0051] Figure 2 It is a maintenance unit flow chart formed based on maintenance capabilities;

[0052] Figure 3 This is a diagram of maintenance units (red, relatively dispersed) formed based on maintenance capabilities;

[0053] Figure 4 This is the segmented maintenance unit (red) diagram obtained by cluster analysis. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0055] A line equipment unit maintenance decision method based on cluster analysis comprises the following steps:

[0056] Step 1: Preliminarily screen out units exceeding the maintenance threshold according to the comprehensive status evaluation results of line equipment units to form preliminary maintenance units;

[0057] Step 2: According to the mechanized maintenance work plan of large machine tamping, large machine screening and rail replacement in the next month or next week, conduct a coherence analysis with the preliminary maintenance units, eliminate the mechanized maintenance units, and select the maintenance units;

[0058] Step 3: According to the on-site maintenance and repair operation capacity, skylight settings, etc., on the basis of the comprehensive status evaluation of the line equipment unit, set the iteration step length ε, consider the differences of different levels of lines, and make the line equipment maintenance unit equal to the actual maintenance and repair capacity Thus, an optimized maintenance unit is obtained;

[0059] L 能 ——Monthly maintenance and repair capabilities.

[0060] ——Critical maintenance value of the jth category line.

[0061] j——The number of over-limit sections in the same grade of lines.

[0062] j represents the line grade, with values ​​ranging from 1 to 3, representing heavy vehicle line, light vehicle line, station line and arrival and departure track respectively;

[0063] Step 4: Considering the on-site construction transfer, protection capabilities, and on-site personnel safety, the distance between maintenance and repair units should not be too long or too scattered. The DBSCAN cluster analysis method is used to form continuous segmented operation units.

[0064] In cluster analysis, the clustering effect evaluation index GF is used to evaluate the clustering effect and to solve the optimal clustering parameters. The larger the GF, the better the clustering effect. The clustering effect evaluation index GF consists of the intra-class similarity function and the out-of-class similarity function:

[0065]

[0066] Intra-class similarity

[0067] Out-of-class similarity

[0068] In the formula,

[0069] r stands for G i The number of points contained in ;

[0070] x j Represents G i The jth point in , 1≤j≤r;

[0071] represents the average value of the points contained in Gi;

[0072] h represents the number of clusters;

[0073] α represents the weight adjustment parameter of Stra(G) and Ster(G), which is determined empirically and is between 0 and 1.

[0074] In step one, the comprehensive status evaluation result of the line equipment unit is obtained through real-time monitoring and data analysis, including but not limited to vibration monitoring, temperature monitoring, wear degree monitoring, etc.

[0075] In step 2, the coherence analysis includes comparing the mechanized maintenance work plan with the location, type and time information of the preliminary maintenance unit to ensure that there is no conflict or duplication between the mechanized maintenance work and the manual maintenance work.

[0076] In step three, the setting of the iteration step is determined based on historical maintenance data and current maintenance resource status to balance maintenance efficiency and resource utilization.

[0077] In step 4, the parameters of the DBSCAN clustering analysis method (such as the minimum sample number MinPts and the neighborhood radius Eps) are determined through experiments or optimization algorithms to obtain the best clustering effect.

[0078] The method also includes step five: adjusting clustering parameters according to the result of the clustering effect evaluation index GF, and re-performing clustering analysis until a clustering effect that meets the requirements is obtained.

[0079] In step 4, the intra-class similarity function of the clustering effect evaluation index GF is calculated based on measurement methods such as Euclidean distance, Manhattan distance or cosine similarity.

[0080] It also includes step six: based on the formed continuous segmented operation units, formulate a detailed maintenance plan and resource allocation plan, including the scheduling of maintenance personnel, preparation of maintenance tools, and arrangement of maintenance time.

[0081] The present invention, based on the comprehensive status evaluation of the line equipment unit section, firstly proposes a preliminary maintenance unit according to the threshold value of the comprehensive status evaluation index, and then clusters the preliminary maintenance units according to the maintenance and repair capacity of the workshop or work area and the interference analysis with the mechanized maintenance sections such as large machine tamping and large machine screening, and optimizes and merges the originally scattered maintenance units to avoid a large number of on-site maintenance points, reduce the pressure of on-site protection, and reduce the safety hazards of personnel.

[0082] Specifically, the DBSCAN clustering analysis method is used to form continuous segmented operation units. The clustering effect evaluation index GF is used to evaluate the clustering effect during cluster analysis to solve the optimal clustering parameters. The larger the GF, the better the clustering effect. The clustering effect evaluation index GF consists of the intra-class similarity function and the out-of-class similarity function. In other words, different from the existing qualitative processing method based on experience, the analysis and evaluation are quantified to achieve the accurate formulation of the daily maintenance plan for railway line equipment.

[0083] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A line equipment unit maintenance decision method based on cluster analysis, characterized in that: The steps include: Step 1: Preliminarily screen out units exceeding the maintenance threshold according to the comprehensive status evaluation results of line equipment units to form preliminary maintenance units; Step 2: According to the mechanized maintenance work plan of large machine tamping, large machine screening and rail replacement in the next month or next week, conduct a coherence analysis with the preliminary maintenance units, eliminate the mechanized maintenance units, and select the maintenance units; Step 3: According to the on-site maintenance and repair operation capacity, skylight settings, etc., on the basis of the comprehensive status evaluation of the line equipment unit, set the iteration step length ε, consider the differences of different levels of lines, and make the line equipment maintenance unit equal to the actual maintenance and repair capacity Thus, an optimized maintenance unit is obtained; L 能 ——Monthly maintenance and repair capabilities; ——critical maintenance value of the jth category line; j——the number of over-limit sections in the same grade of lines; j represents the line grade, with values ​​ranging from 1 to 3, representing heavy vehicle line, light vehicle line, station line and arrival and departure track respectively; Step 4: Considering the on-site construction transfer, protection capabilities, and on-site personnel safety, the distance between maintenance and repair units should not be too long or too scattered. The DBSCAN cluster analysis method is used to form continuous segmented operation units. In cluster analysis, the clustering effect evaluation index GF is used to evaluate the clustering effect and to solve the optimal clustering parameters; the larger the GF, the better the clustering effect; The clustering effect evaluation index GF consists of the intra-class similarity function and the out-of-class similarity function: Intra-class similarity Out-of-class similarity In the formula, r stands for G i The number of points contained in ; x j Represents G i The jth point in , 1≤j≤r; represents the average value of the points contained in Gi; h represents the number of clusters; α represents the weight adjustment parameter of Stra(G) and Ster(G), which is determined empirically and is between 0 and 1.

2. A line equipment unit maintenance decision method based on cluster analysis as claimed in claim 1, characterized in that: In step one, the comprehensive status evaluation result of the line equipment unit is obtained through real-time monitoring and data analysis, including but not limited to vibration monitoring, temperature monitoring, wear degree monitoring, etc.

3. A line equipment unit maintenance decision method based on cluster analysis as claimed in claim 1, characterized in that: In step 2, the coherence analysis includes comparing the mechanized maintenance work plan with the location, type and time information of the preliminary maintenance unit to ensure that there is no conflict or duplication between the mechanized maintenance work and the manual maintenance work.

4. A line equipment unit maintenance decision method based on cluster analysis as claimed in claim 1, characterized in that: In step three, the setting of the iteration step is determined based on historical maintenance data and current maintenance resource status to balance maintenance efficiency and resource utilization.

5. A line equipment unit maintenance decision method based on cluster analysis as claimed in claim 1, characterized in that: In step 4, the parameters of the DBSCAN clustering analysis method are determined through experiments or optimization algorithms to obtain the best clustering effect.

6. A line equipment unit maintenance decision method based on cluster analysis as claimed in claim 1, characterized in that: The method also includes step five: adjusting clustering parameters according to the result of the clustering effect evaluation index GF, and re-performing clustering analysis until a clustering effect that meets the requirements is obtained.

7. A line equipment unit maintenance decision method based on cluster analysis as claimed in claim 1, characterized in that: In step 4, the intra-class similarity function of the clustering effect evaluation index GF is calculated based on measurement methods such as Euclidean distance, Manhattan distance or cosine similarity.

8. A line equipment unit maintenance decision method based on cluster analysis as claimed in claim 6, characterized in that: It also includes step six: based on the formed continuous segmented operation units, formulate a detailed maintenance plan and resource allocation plan, including the scheduling of maintenance personnel, preparation of maintenance tools, and arrangement of maintenance time.

Images