Locomotive oil consumption optimization method and system based on edge perception and data mining

Abstract

The invention relates to the technical field of diesel locomotives, in particular to a locomotive fuel consumption optimization method and system based on edge perception and data mining, and the method comprises the following steps: obtaining locomotive motion state perception data and locomotive operation big data; preprocessing the motion state sensing data and the locomotive operation big data, and performing multi-dimensional association rule mining on the preprocessed data to generate an association rule; based on the association rule, mining a multi-dimensional locomotive fuel consumption relation model to obtain a multi-dimensional locomotive fuel consumption relation model; and based on the multi-dimensional locomotive fuel consumption relation model, carrying out model clustering analysis under different road sections and different operation states to obtain a train control scheme which has the lowest fuel consumption in the whole road section of the locomotive and meets train dispatching requirements. Based on edge perception and data mining, high-precision association of locomotive operation data and oil consumption data can be effectively realized, an optimal multi-dimensional locomotive oil consumption relation model is mined, and a train control scheme can be objectively generated without depending on artificial qualitative analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of diesel locomotives, and in particular to a locomotive fuel consumption optimization method and system based on edge sensing and data mining. Background Art

[0002] Since the fuel consumption cost of diesel locomotives in railway transportation is relatively high and its expenditure cost accounts for a relatively high proportion of the total transportation cost, scientific and systematic monitoring and analysis of locomotive fuel consumption and proposing corresponding reasonable fuel-saving methods are of great significance to achieving the national energy conservation and emission reduction goals and the railway transportation companies' energy conservation and emission reduction, safe operation, and maximization of efficiency and benefits.

[0003] At present, the research on methods for optimizing fuel consumption of diesel locomotives mainly includes the following two categories:

[0004] One type adopts qualitative analysis methods, which analyzes the relevant factors affecting locomotive fuel consumption, specifies measures or methods to reduce locomotive fuel consumption and improve locomotive operating efficiency, mainly including improving locomotive operation mode, improving the comprehensive quality of operators, improving fuel quality and increasing locomotive power.

[0005] The other is to quantitatively analyze the main influencing factors of fuel consumption, construct a mathematical model of fuel consumption and influencing factors, rank and analyze the influencing factors, and finally summarize the fuel consumption optimization plan suitable for different vehicle models and routes.

[0006] However, the above method mainly relies on manually segmenting statistical data and building models for calculation and analysis. The data used is limited to the operating data of a locomotive on a certain line. It lacks a large amount of refined actual operating data and cannot accurately, objectively and truly establish a fuel consumption model for the actual operation of the train. At the same time, the optimization plan formed is not suitable for railway transportation lines with long lines, poor operating environment or complex and changeable lines. In addition, the proportion of qualitative analysis components is relatively high, and the scientificity and objectivity of the conclusions formed need to be further verified in practical applications. Summary of the invention

[0007] The purpose of the present invention is to provide a locomotive fuel consumption optimization method and system based on edge sensing and data mining, which can efficiently and accurately establish a relationship model between fuel consumption and locomotive operating status, and output a train control solution with the lowest fuel consumption in the entire section of the locomotive and meeting the train dispatching requirements, so as to solve the problems pointed out in the background technology.

[0008] The embodiment of the present invention is implemented by the following technical solution: A locomotive fuel consumption optimization method based on edge sensing and data mining, comprising the following steps:

[0009] Obtain locomotive motion status perception data and locomotive operation big data to build a basic database;

[0010] Preprocessing the motion state perception data and locomotive operation big data, and performing multi-dimensional association rule mining on the preprocessed data to generate association rules and build an operation big database;

[0011] Based on the association rules, a multi-dimensional locomotive fuel consumption relationship model is mined to obtain a multi-dimensional locomotive fuel consumption relationship model;

[0012] Based on the multi-dimensional locomotive fuel consumption relationship model, model clustering analysis is carried out in different sections and under different operating conditions to obtain a train control solution with the lowest locomotive fuel consumption in the entire section and that meets the train dispatching needs.

[0013] According to a preferred embodiment, the locomotive motion state perception data includes motion state data of locomotives of different types and on different lines in all areas and all time periods.

[0014] According to a preferred embodiment, the locomotive operation big data includes one or more of weather environment, line conditions, locomotive entry and exit time periods, shunting time, line section fuel consumption, near real-time fuel consumption, locomotive power, locomotive thermal status and maintenance status.

[0015] According to a preferred embodiment, the preprocessing includes one or more of data cleaning, data noise reduction, data transformation, smoothing aggregation, data generalization, data normalization and normalization processing.

[0016] According to a preferred embodiment, an association rule mining method is used to perform multi-dimensional association mining on the pre-processed data, specifically including:

[0017] A1. For locomotives of the same type, a standardized data structure of locomotive motion model and fuel consumption under the same load conditions is constructed;

[0018] A2. Divide the route and associate the locomotive data in sections according to the spatial position of each section;

[0019] A3. Carry out normalized correlation of fuel consumption data under the same terrain environment and different motion states for the same load condition and the same road section;

[0020] A4. Carry out normalized correlation of fuel consumption data under different terrain environments and different motion states for the same load conditions and data of different road sections;

[0021] A5. Repeat steps A2 to A4 to complete the normalized association of the motion state, terrain environment and fuel consumption data under different load conditions.

[0022] According to a preferred implementation, based on the association rules, a multi-dimensional locomotive fuel consumption relationship model is mined, specifically including:

[0023] Determine the basic mining unit, take the locomotive operating status data as the basis, and use all relevant fuel consumption data in each section of the operating line as input to carry out locomotive fuel consumption relationship model mining under different operating states, different motion states, and various stages of motion states.

[0024] According to a preferred implementation, model clustering analysis is performed under different sections and different operating conditions based on the multi-dimensional locomotive fuel consumption relationship model, specifically including:

[0025] Based on the multi-dimensional locomotive fuel consumption relationship model, cluster analysis is carried out under different motion states to obtain an integrated model;

[0026] Based on the integrated model of the same section, the classification aggregation of the locomotive's full section operating status under different motion parameters is constructed to establish a digital twin model;

[0027] The train dispatching requirements are obtained, and the digital twin model is used to perform cluster analysis of locomotive motion states.

[0028] According to a preferred implementation, before conducting model clustering analysis under different sections and different operating conditions based on the multi-dimensional locomotive fuel consumption relationship model, the method also includes: iteratively optimizing the multi-dimensional locomotive fuel consumption relationship model according to locomotive real-time data to obtain an optimized multi-dimensional locomotive fuel consumption relationship model.

[0029] The present invention also provides a locomotive fuel consumption optimization system based on edge sensing and data mining, comprising:

[0030] Data acquisition module, used to obtain locomotive motion state perception data and locomotive operation big data, and build a basic database;

[0031] A preprocessing module is used to preprocess the motion state perception data and the locomotive operation big data, and perform multi-dimensional association rule mining on the preprocessed data to generate association rules and build an operation big database;

[0032] A data mining module, used for mining a multi-dimensional locomotive fuel consumption relationship model based on the association rules to obtain a multi-dimensional locomotive fuel consumption relationship model;

[0033] The cluster analysis module is used to carry out model cluster analysis under different sections and different operating conditions based on the multi-dimensional locomotive fuel consumption relationship model, so as to obtain a train control solution with the lowest fuel consumption in the entire section of the locomotive and meeting the train dispatching requirements.

[0034] The technical solution of a locomotive fuel consumption optimization method and system based on edge sensing and data mining provided by an embodiment of the present invention has at least the following advantages and beneficial effects:

[0035] (1) By acquiring locomotive motion state perception data in real time, the granularity and reliability of the basic data on locomotive operation status used for fuel consumption optimization can be greatly improved, and the actual operation status of the locomotive can be better reflected, so that the final fuel consumption optimization train control plan is more scientific and objective; (2) Based on edge perception and data mining, it can effectively realize the high-precision correlation between locomotive operation data and fuel consumption data, and mine the optimal multi-dimensional locomotive fuel consumption relationship model, which can achieve objective generation of train control plans without relying on manual qualitative analysis. It can also provide support for various aspects such as locomotive intelligent operation and railway safety risk monitoring. It has good practical application value and will greatly improve the production efficiency and economic benefits of railway transportation; (3) This method is applicable to any railway transportation line, including lines with long lines, poor operating environment or complex and changeable lines. Compared with traditional methods, it not only improves the reliability and accuracy of the results, but also has good universality. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A schematic flow chart of a locomotive fuel consumption optimization method based on edge sensing and data mining provided in Example 1 of the present invention;

[0037] Figure 2 A detailed implementation flow chart of the locomotive fuel consumption optimization method provided in Example 1 of the present invention;

[0038] Figure 3 This is an example diagram of locomotive motion state perception data provided by Embodiment 1 of the present invention;

[0039] Figure 4 This is a schematic diagram of the multi-dimensional locomotive fuel consumption relationship model architecture provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0041] Example 1

[0042] Figure 1 This is a flow chart of a first preferred embodiment of a locomotive fuel consumption optimization method based on edge sensing and data mining according to an embodiment of the present invention. Figure 2 The following is a detailed implementation flow chart of the locomotive fuel consumption optimization method. Figure 1 As shown, the locomotive fuel consumption optimization method includes steps S1 to S4.

[0043] Step S1, obtain locomotive motion state perception data through the locomotive motion state intelligent perception terminal, collect relevant locomotive operation big data, and build a basic database; it should be noted that the locomotive motion state intelligent perception terminal can use a device that can complete the real-time collection, storage and data feedback of the locomotive motion state perception data, which will not be described in detail here.

[0044] Among them, in one implementation of this embodiment, if Figure 2 As shown, the locomotive motion state perception data obtained includes the motion state data of locomotives of different types and different lines in the whole area and all time periods, such as locomotive spatial position, speed, three-axis speed, acceleration, three-axis acceleration, time, braking time, braking position, train control events, GNSS position information before loss of lock, etc., and finally realizes the acquisition of motion state perception data of locomotives of different types and different lines in the whole area and all time periods, see Figure 3 As shown in the figure; after the locomotive motion state intelligent sensing terminal completes the collection of locomotive motion state sensing data, the Beidou short message or 4G / 5G transmission method is used to transmit the data back to the data monitoring center to form the basic data of the locomotive operation status, which can greatly improve the basic data of the locomotive operation status used for fuel consumption optimization and the reliability of the data, and better reflect the actual operation of the locomotive, so that the subsequent fuel consumption optimization train control plan is more scientific and objective.

[0045] The locomotive operation big data includes weather environment (temperature, precipitation, wind speed, wind direction, etc.), line conditions (elevation, slope, slope length, speed limit, etc.), locomotive entry and exit time periods, shunting time, line section fuel consumption, locomotive power, locomotive thermal status, recent fuel consumption data and maintenance status, etc.

[0046] Step S2: pre-process the motion state perception data and locomotive operation big data, and perform multi-dimensional association rule mining on the pre-processed data to generate association rules, realize multi-dimensional association data planning and processing of fuel consumption, spatial position, and motion state, and build an operation big database.

[0047] In one implementation of this embodiment, the preprocessing includes data cleaning, data noise reduction, data transformation, smooth aggregation, data generalization, data normalization and normalization processing, etc.; among which, smooth aggregation, data generalization and data normalization are used to convert data into a form suitable for data mining.

[0048] In addition, preprocessing also includes constructing prior labels for motion state perception data. Taking the interval station as an example, a priori labels are constructed for the terrain of the interval, and the railway spatial data between the interval stations is divided as follows:

[0049] (1) Flat sections, including the starting and ending point location information, are sections with less impact on fuel consumption. The flat sections referred to here are not completely level and may also include lines with a small slope;

[0050] (2) The uphill section, including the starting point and the end point location information, has a certain impact on fuel consumption and has a significant difference in fuel consumption from the flat section;

[0051] (3) Turning sections, including the starting and ending point location information, including the turning angle and turning length, which can be distributed in flat sections, uphill sections, and downhill sections, and are independently marked with typical speed limit factors;

[0052] (4) The downhill section, including the starting point, end point, and slope angle, has a certain impact on fuel consumption, and has a significant difference in fuel consumption from the flat section.

[0053] Combining the above four types of interval terrain divisions, accurate fuel consumption learning label data can be provided for each section of locomotive operation status perception data.

[0054] like Figure 4 As shown, the association rule mining method is used to perform multi-dimensional association mining on the preprocessed data, specifically including:

[0055] A1. For locomotives of the same type, a standardized data structure of locomotive motion model and fuel consumption under the same load conditions is constructed;

[0056] A2. Divide the line into several flat sections, uphill sections, turning sections, downhill sections and deceleration sections, and associate the locomotive data by sections according to the spatial position of each section;

[0057] A3. Carry out normalized correlation of fuel consumption data under the same terrain environment and different motion states for the same load condition and the same road section;

[0058] A4. Carry out normalized correlation of fuel consumption data under different terrain environments and different motion states for the same load conditions and data of different road sections;

[0059] A5. Repeat steps A2 to A4 to complete the normalized association of the motion state, terrain environment and fuel consumption data under different load conditions.

[0060] According to the above steps A1 to A5, multi-dimensional correlation data planning and processing of fuel consumption, spatial position and motion status are completed for different locomotives.

[0061] Step S3: Based on the association rules, a multi-dimensional locomotive fuel consumption relationship model is mined to obtain a multi-dimensional locomotive fuel consumption relationship model.

[0062] In one implementation of this embodiment, 500 meters is used as the basic mining unit of the running section, and the normalized running state data of the locomotive is used as the basis, and all relevant fuel consumption data in each section of the running line (flat section, uphill section, downhill section, turning section, etc.) is used as input to carry out the mining of the locomotive fuel consumption relationship model under different running states, different motion states (uniform speed state, uniform acceleration state, uniform deceleration state, etc.), and each stage of motion state (uniform speed state, initial speed-acceleration-final speed under uniform acceleration state, initial speed-acceleration-final speed under uniform deceleration state). Among them, in this embodiment, deep learning is used to mine the locomotive fuel consumption relationship model under uniform speed state-fuel consumption, initial speed-acceleration-final speed-fuel consumption under uniform acceleration state, and initial speed-acceleration-final speed-fuel consumption under uniform deceleration state, so as to realize the establishment of a refined fuel consumption model of the real road conditions of the actual running fuel consumption data of the locomotive.

[0063] Furthermore, the multi-dimensional locomotive fuel consumption relationship model is iteratively optimized according to the locomotive real-time data to obtain an optimized multi-dimensional locomotive fuel consumption relationship model, thereby obtaining an optimal model.

[0064] Step S4: Based on the multi-dimensional locomotive fuel consumption relationship model, model clustering analysis is performed under different sections and different operating conditions to obtain a train control solution with the lowest locomotive fuel consumption in the entire section and meeting the train dispatching requirements.

[0065] In one implementation of this embodiment, cluster analysis is performed under different motion states based on the multi-dimensional locomotive fuel consumption relationship model to obtain a comprehensive fuel consumption-speed-acceleration integration model;

[0066] Based on the integrated model of the same section, the classification aggregation of the locomotive's operating status in the entire section under different motion parameters (uniform speed, uniform acceleration, uniform deceleration, etc.) is constructed, dynamic monitoring and fitting evaluation with the actual fuel consumption are carried out, and a digital twin model of the locomotive's current motion parameters, line parameters and fuel consumption parameters is established;

[0067] The train dispatching requirements (departure time, departure location, load, departure station, arrival station, etc.) are obtained, and the digital twin model is used to perform cluster analysis of the locomotive motion state, complete the planning of different locomotive motion states in a specific time interval, complete the optimization management of fuel consumption, and finally output the train control plan with the lowest fuel consumption in the entire section of the locomotive and meeting the train dispatching requirements. The train control plan includes detailed operation plans such as segmented acceleration, speed, deceleration, etc., which can fully cover the motion state information required for the normal operation of the locomotive for driving use. Therefore, it can be applied to any railway transportation line, including those with long lines, poor operating environment or complex and changeable lines. Compared with traditional methods, it not only improves the reliability and accuracy of the results, but also has good universality.

[0068] In summary, based on edge perception and data mining, it is possible to effectively realize high-precision correlation between locomotive operation data and fuel consumption data, and to mine the optimal multi-dimensional locomotive fuel consumption relationship model. It is possible to objectively generate train control plans without relying on manual qualitative analysis, and it can also provide support for various aspects such as locomotive intelligent operation and railway safety risk monitoring. It has good practical application value and will greatly improve the production efficiency and economic benefits of railway transportation.

[0069] Example 2

[0070] This embodiment, based on the technical solution provided in Embodiment 1, provides a locomotive fuel consumption optimization system based on edge sensing and data mining, including:

[0071] Data acquisition module, used to obtain locomotive motion state perception data and locomotive operation big data, and build a basic database;

[0072] A preprocessing module is used to preprocess the motion state perception data and the locomotive operation big data, and perform multi-dimensional association rule mining on the preprocessed data to generate association rules and build an operation big database;

[0073] A data mining module, used for mining a multi-dimensional locomotive fuel consumption relationship model based on the association rules to obtain a multi-dimensional locomotive fuel consumption relationship model;

[0074] The cluster analysis module is used to carry out model cluster analysis under different sections and different operating conditions based on the multi-dimensional locomotive fuel consumption relationship model, so as to obtain a train control solution with the lowest fuel consumption in the entire section of the locomotive and meeting the train dispatching requirements.

[0075] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A locomotive fuel consumption optimization method based on edge sensing and data mining, characterized in that: The following steps are involved: Obtain locomotive motion status perception data and locomotive operation big data to build a basic database; Preprocessing the motion state perception data and locomotive operation big data, and performing multi-dimensional association rule mining on the preprocessed data to generate association rules and build an operation big database; Based on the association rules, a multi-dimensional locomotive fuel consumption relationship model is mined to obtain a multi-dimensional locomotive fuel consumption relationship model; Based on the multi-dimensional locomotive fuel consumption relationship model, model clustering analysis is carried out in different sections and under different operating conditions to obtain a train control solution with the lowest locomotive fuel consumption in the entire section and that meets the train dispatching needs.

2. The locomotive fuel consumption optimization method based on edge sensing and data mining according to claim 1, characterized in that: The locomotive motion state perception data includes motion state data of locomotives of different types and lines in all areas and all time periods.

3. The locomotive fuel consumption optimization method based on edge sensing and data mining according to claim 1, characterized in that: The locomotive operation big data includes one or more of weather environment, line conditions, locomotive entry and exit time periods, shunting time, line section fuel consumption, near real-time fuel consumption, locomotive power, locomotive thermal status and maintenance status.

4. The locomotive fuel consumption optimization method based on edge sensing and data mining according to claim 1, characterized in that: The preprocessing includes one or more of data cleaning, data noise reduction, data transformation, smoothing aggregation, data generalization, data normalization and normalization processing.

5. The locomotive fuel consumption optimization method based on edge sensing and data mining according to claim 1, characterized in that: The association rule mining method is used to conduct multi-dimensional association mining on the pre-processed data, including: A1. For locomotives of the same type, a standardized data structure of locomotive motion model and fuel consumption under the same load conditions is constructed; A2. Divide the route and associate the locomotive data in sections according to the spatial position of each section; A3. Carry out normalized correlation of fuel consumption data under the same terrain environment and different motion states for the same load condition and the same road section; A4. Carry out normalized correlation of fuel consumption data under different terrain environments and different motion states for the same load conditions and data of different road sections; A5. Repeat steps A2 to A4 to complete the normalized association of the motion state, terrain environment and fuel consumption data under different load conditions.

6. The locomotive fuel consumption optimization method based on edge sensing and data mining according to claim 1, characterized in that: Based on the association rules, a multi-dimensional locomotive fuel consumption relationship model is mined, specifically including: Determine the basic mining unit, take the locomotive operating status data as the basis, and use all relevant fuel consumption data in each section of the operating line as input to carry out locomotive fuel consumption relationship model mining under different operating states, different motion states, and various stages of motion states.

7. The locomotive fuel consumption optimization method based on edge sensing and data mining according to claim 1, characterized in that: Based on the multi-dimensional locomotive fuel consumption relationship model, model clustering analysis is carried out under different sections and different operating conditions, specifically including: Based on the multi-dimensional locomotive fuel consumption relationship model, cluster analysis is carried out under different motion states to obtain an integrated model; Based on the integrated model of the same section, the classification aggregation of the locomotive's full section operating status under different motion parameters is constructed to establish a digital twin model; The train dispatching requirements are obtained, and the digital twin model is used to perform cluster analysis of locomotive motion states.

8. The locomotive fuel consumption optimization method based on edge sensing and data mining according to any one of claims 1 to 7, characterized in that: Before carrying out model clustering analysis under different sections and different operating conditions based on the multidimensional locomotive fuel consumption relationship model, the method also includes: iteratively optimizing the multidimensional locomotive fuel consumption relationship model according to locomotive real-time data to obtain an optimized multidimensional locomotive fuel consumption relationship model.

9. A locomotive fuel consumption optimization system based on edge sensing and data mining, characterized in that: include: Data acquisition module, used to obtain locomotive motion state perception data and locomotive operation big data, and build a basic database; A preprocessing module is used to preprocess the motion state perception data and the locomotive operation big data, and perform multi-dimensional association rule mining on the preprocessed data to generate association rules and build an operation big database; A data mining module, used for mining a multi-dimensional locomotive fuel consumption relationship model based on the association rules to obtain a multi-dimensional locomotive fuel consumption relationship model; The cluster analysis module is used to carry out model cluster analysis under different sections and different operating conditions based on the multi-dimensional locomotive fuel consumption relationship model, so as to obtain a train control solution with the lowest fuel consumption in the entire section of the locomotive and meeting the train dispatching requirements.

Images