A Data Fusion-Based Method and System for Optimizing Energy Consumption of Railway Freight Equipment

Abstract

This invention relates to the field of energy consumption optimization technology, and in particular to a method and system for optimizing the energy consumption of railway freight equipment based on data fusion. The method includes the following steps: acquiring the operating status data and track environment data of the target railway freight equipment; preprocessing the operating status data and track environment data to generate a standardized freight dataset; constructing a freight equipment operating condition analysis window based on the time series characteristics of the standardized freight dataset; extracting data segments related to the energy consumption characteristics in the standardized freight dataset from the freight equipment operating condition analysis window, as a set of energy consumption analysis units; mapping the input parameters of the energy consumption analysis unit set to the standardized freight dataset. This invention achieves closed-loop optimization from data perception to control decision-making through multi-source data fusion, time-series windowed analysis, and feature-driven optimization mapping mechanisms.

Description

Technical Field

[0001] This invention relates to the field of energy consumption optimization technology, and in particular to a method and system for optimizing energy consumption of railway freight equipment based on data fusion. Background Technology

[0002] In the railway freight transport system, the energy consumption level of freight equipment directly affects transportation costs and operational efficiency. With the continuous growth of heavy-haul railways and long-distance transportation demands, train operating conditions are becoming increasingly complex and dynamic. Especially when different track gradients, curve radii, and speed limits change, the load fluctuation of the traction system is significant, leading to a continuous increase in the difficulty of energy consumption control. In existing technologies, energy consumption is usually estimated and optimized based on single operating status data or empirical models, which is difficult to fully reflect the coupling relationship between the operating environment and equipment status. Although existing data acquisition systems can acquire train operating status data or track environment data, there are generally problems such as time asynchrony, inconsistent formats, and uneven data quality among multi-source data, making data fusion processing difficult to carry out effectively. Traditional energy consumption analysis methods mostly adopt global statistics or static interval analysis methods, lacking the ability to classify the fine-grained operating conditions of the train operation process, and failing to accurately depict the energy consumption characteristics and changes in different operating stages, resulting in poor adaptability of optimization results. Summary of the Invention

[0003] Therefore, it is necessary to provide a data fusion-based method and system for optimizing the energy consumption of railway freight equipment to solve at least one of the aforementioned technical problems.

[0004] To achieve the above objectives, a method for optimizing the energy consumption of railway freight equipment based on data fusion includes the following steps: Step S1: Obtain the operating status data and line environment data of the target railway freight equipment, preprocess the operating status data and line environment data, and generate a standardized freight dataset; Step S2: Based on the time series characteristics of the standardized freight dataset, construct a window for analyzing the operating conditions of freight equipment; Step S3: In the freight equipment operation condition analysis window, extract the data segments that are associated with the energy consumption characteristics in the standardized freight dataset, and use them as the energy consumption analysis unit set; Step S4: Map the input parameters of the energy consumption analysis unit set to the standardized freight dataset, call the preset energy consumption optimization strategy to perform calculations, and generate energy consumption optimization control parameters for railway freight equipment.

[0005] This invention also provides a data fusion-based railway freight equipment energy consumption optimization system for executing the data fusion-based railway freight equipment energy consumption optimization method described above. The data fusion-based railway freight equipment energy consumption optimization system includes: The data acquisition module is used to acquire the operating status data and line environment data of the target railway freight equipment, preprocess the operating status data and line environment data, and generate a standardized freight dataset. The operating condition analysis module is used to construct an operating condition analysis window for freight equipment based on the time series characteristics of standardized freight datasets. The feature extraction module is used to extract data segments that are associated with energy consumption characteristics in the standardized freight data set from the freight equipment operation condition analysis window, and use them as a set of energy consumption analysis units. The optimization decision module maps the input parameters of the energy consumption analysis unit set to the standardized freight dataset, calls the preset energy consumption optimization strategy for calculation, and generates energy consumption optimization control parameters for railway freight equipment.

[0006] This invention achieves a systematic and refined improvement in the energy consumption optimization and control of railway freight equipment by constructing an integrated closed-loop processing mechanism that combines data acquisition, time series modeling, working condition window parsing, feature unit extraction strategy mapping and optimization. At the data level, by integrating operational status data and line environment data, and performing time synchronization, anomaly removal, missing data completion, and normalization, the inconsistencies in time scale, dimensional system, and data quality among multi-source heterogeneous data are effectively eliminated. This generates a highly consistent and highly available standardized freight dataset, providing a stable data foundation for subsequent modeling.

[0007] At the time series modeling level, by introducing a sliding window-based operating condition analysis mechanism, continuous operating data is divided into overlapping operating condition analysis windows according to time series characteristics. This allows the originally discrete operating process to be structured into computable local spatiotemporal units, thereby enabling segmented characterization of energy consumption behavior at different operating stages and effectively enhancing the ability to express complex operating state changes.

[0008] At the feature extraction level, by extracting key energy consumption features such as unit energy consumption, traction load and line gradient within the operating condition window, and further forming window-level statistical feature units, the system can extract representative energy consumption behavior patterns from local operating sections, thereby avoiding information loss caused by global averaging analysis and improving the sensitivity and discriminativeness of energy consumption feature identification.

[0009] At the optimization decision-making level, by mapping the energy consumption analysis unit set to a standardized freight dataset and calling a preset energy consumption optimization strategy to solve for parameters, the coordinated optimization of control variables such as traction control, operating speed, and energy allocation is achieved. This enables the dynamic generation of optimal control parameters for different operating conditions. Through multi-source data fusion, time-series windowed analysis, and feature-driven optimization mapping mechanisms, a closed-loop optimization from data perception to control decision-making is realized. Compared with traditional energy consumption control methods based on static or single-dimensional data, this approach has stronger adaptability, higher precision, and better energy-saving optimization effects. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating the steps of a data fusion-based method for optimizing the energy consumption of railway freight equipment. Figure 2 A schematic diagram of a railway freight equipment energy consumption optimization system based on data fusion; Figure 3 A schematic diagram of an energy-efficient architecture for railway freight equipment; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0011] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0012] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0013] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0014] To achieve the above objectives, please refer to Figures 1 to 3 A method for optimizing energy consumption of railway freight equipment based on data fusion includes the following steps: All specific values ​​involved in this embodiment are exemplary parameters used to clearly illustrate the technical operation process and are not the only limitation of the present invention.

[0015] Step S1: Obtain the operating status data and line environment data of the target railway freight equipment, preprocess the operating status data and line environment data, and generate a standardized freight dataset; Step S2: Based on the time series characteristics of the standardized freight dataset, construct a window for analyzing the operating conditions of freight equipment; Step S3: In the freight equipment operation condition analysis window, extract the data segments that are associated with the energy consumption characteristics in the standardized freight dataset, and use them as the energy consumption analysis unit set; Step S4: Map the input parameters of the energy consumption analysis unit set to the standardized freight dataset, call the preset energy consumption optimization strategy to perform calculations, and generate energy consumption optimization control parameters for railway freight equipment.

[0016] In one embodiment, a data fusion-based method for optimizing the energy consumption of railway freight equipment is applied to a railway freight equipment energy consumption management platform. This platform is communicatively connected to onboard sensors, a track monitoring system, and an energy consumption optimization control terminal. Onboard sensors collect operational status data of the target railway freight equipment, including traction current, traction voltage, operating speed, traction load, braking status, motor temperature, and energy consumption per unit time. The track monitoring system collects track environment data corresponding to the target railway freight equipment, including track gradient, curve radius, track condition, section speed limit information, and ambient temperature information. The railway freight equipment energy consumption management platform performs time synchronization, format standardization, outlier removal, missing value imputation, normalization, and feature encoding on the operational status data and track environment data to generate a standardized freight dataset.

[0017] Furthermore, the railway freight equipment energy consumption management platform constructs operational condition analysis windows for freight equipment based on the time index and sampling frequency of the standardized freight dataset. During the construction process, continuous time-series slicing operations are performed on the standardized freight dataset according to the preset sliding window length and window overlap rate, forming multiple operational condition analysis windows with overlapping areas. Subsequently, data segments corresponding to unit energy consumption, traction load, and track gradient are extracted from each operational condition analysis window, and mean calculation, variance calculation, and threshold filtering are performed on each data segment to generate an energy consumption analysis unit set. The railway freight equipment energy consumption management platform maps the input parameters in the energy consumption analysis unit set to the standardized freight dataset and calls the preset energy consumption optimization strategy model to optimize traction control parameters, speed control parameters, and power distribution parameters, thereby generating energy consumption optimization control parameters for railway freight equipment to guide the energy-saving operation control of the target railway freight equipment.

[0018] In another embodiment, a data fusion-based method for optimizing the energy consumption of railway freight equipment is applied to an intelligent energy consumption optimization system in a railway freight dispatching center. The intelligent energy consumption optimization system includes a data access server, a condition analysis server, and an optimization control server. The data access server receives operational status data from the target railway freight equipment and line environment data from the line monitoring system, and associates and stores data from different sources based on a unified timestamp to form an original freight data set. Subsequently, the condition analysis server performs time axis alignment and dimension unification processing on the data from different sources in the original freight data set, generating synchronous heterogeneous data; it also identifies abnormal data points in the synchronous heterogeneous data, performs linear interpolation to complete missing data, and performs smoothing processing on outlier data to generate cleaned freight data. Afterwards, continuous variables in the cleaned freight data are normalized, and discrete variables are converted using one-hot encoding to form a standardized freight dataset.

[0019] Furthermore, the operating condition analysis server, based on the time series length, sampling frequency, and window sliding step size parameters of the standardized freight dataset, delineates multiple candidate window boundary points on the time series and generates corresponding operating condition analysis windows based on these boundary points. Subsequently, statistical features are extracted from the unit energy consumption, traction load, and track gradient in each operating condition analysis window to obtain window feature statistics. These statistics are then compared with preset energy consumption feature thresholds to select windows that meet the energy consumption analysis conditions, forming an energy consumption analysis unit set. The optimization control server, based on the feature parameters in the energy consumption analysis unit set, invokes preset energy consumption optimization strategies to optimize the traction power output, speed regulation strategy, and braking energy recovery strategy of the target railway freight equipment under different operating sections. This generates corresponding railway freight equipment energy consumption optimization control parameters, which are then sent to the control terminal of the target railway freight equipment for execution.

[0020] Please refer to [link / reference needed] for further information. Figure 3 The process involves data acquisition, standardization transformation, outlier correction, and cleaning to generate a standardized freight dataset. Operating parameters are extracted based on time-series features, time windows are divided, and similarity matching is performed to construct an operational condition analysis window for freight equipment. The window data is then sliced, features are extracted, and influencing factors are correlated to generate a set of energy consumption analysis units. Based on a multi-objective optimization model, the solution is verified, and energy consumption optimization control parameters for railway freight equipment are output, achieving the energy consumption optimization objective driven by the entire process.

[0021] Preferably, step S1 includes the following steps: Step S11: Obtain the operating status data and track environment data of the target railway freight equipment; Step S12: Perform time synchronization and format unification on the operating status data and line environment data to generate synchronized heterogeneous data; Step S13: Remove outliers and fill in missing values ​​in the synchronous heterogeneous data to generate cleaned freight data; Step S14: Normalize and encode the cleaned freight data to generate a standardized freight dataset.

[0022] In one embodiment, operational status data is acquired through onboard sensors installed on the target railway freight equipment. This operational status data includes traction current, traction voltage, operating speed, traction load, motor temperature, braking status, and energy consumption per unit time. Simultaneously, track environmental data is acquired through a track monitoring system. This track environmental data includes track gradient, curve radius, track condition, section speed limit information, and ambient temperature information. Subsequently, the operational status data and track environmental data are associated and stored according to the acquisition timestamp to form the original freight data set.

[0023] The system reads the time stamps from the original freight data set and performs time axis alignment processing on data from different sources to align the operational status data and the route environment data to a unified sampling time point. Subsequently, it performs a unified conversion on the field structure and units of data from different sources, converting data of different formats into a unified data format and data of different units into a unified unit standard. After that, it merges the time-synchronized data with the format-unified data to generate synchronized heterogeneous data.

[0024] Statistical distribution analysis is performed on numerical fields in synchronous heterogeneous data. Outlier data points are identified and marked based on the mean, standard deviation, and preset deviation range. Then, linear interpolation is used to complete the missing data corresponding to the outlier data points, generating interpolated and completed data. After that, smoothing is performed on the outlier data in the interpolated and completed data to reduce the impact of local abnormal fluctuations on the subsequent analysis results, thereby generating cleaned freight data.

[0025] The continuous variables in the cleaned freight data are normalized to their maximum and minimum values ​​to map different continuous variables to a unified numerical range. At the same time, the discrete variables in the cleaned freight data are transformed by one-hot encoding to form corresponding discrete feature encoding vectors. Subsequently, the normalized continuous variable data and the encoded discrete variable data are concatenated in chronological order to generate a standardized freight dataset for subsequent analysis of freight equipment operating conditions and energy consumption optimization calculations.

[0026] Preferably, step S11 includes the following steps: Step S111: Collect operational status data of the target railway freight equipment using onboard sensors; Step S112: Collect track environment data of the target railway freight equipment through the track monitoring system; Step S113: Associate and store the operating status data and the line environment data according to the collection timestamp to generate the original freight data set.

[0027] In one embodiment, onboard sensors installed on the target railway freight equipment continuously collect operational status data. These onboard sensors include current sensors, voltage sensors, speed sensors, temperature sensors, and braking status detection sensors. The operational status data includes traction current, traction voltage, operating speed, traction load, motor temperature, braking status, and energy consumption per unit time. Each onboard sensor samples the status parameters of the target railway freight equipment in real time according to a preset sampling period and sends the collected operational status data to a data access terminal for caching and recording.

[0028] The track monitoring system collects track environmental data for the target railway freight equipment's corresponding operating line. This system includes gradient monitoring, track condition monitoring, curvature monitoring, and environmental monitoring units. The track environmental data includes track gradient, curve radius, track smoothness, section speed limits, ambient temperature, and climate conditions. Based on the target railway freight equipment's current location, the track monitoring system synchronously collects track environmental parameters for the corresponding section and sends the collected data to a data access terminal.

[0029] The system reads the timestamps corresponding to the operational status data and the line environment data, and matches the time correspondence between data from different sources. Then, it performs association mapping processing on the operational status data and the line environment data according to a unified time axis, so that the operational status parameters and the line environment parameters within the same time interval form a correspondence. After that, the time-associated data is stored uniformly according to a preset data structure to generate the original freight data set for subsequent time synchronization processing and data cleaning processing.

[0030] Preferably, step S12 includes the following steps: Step S121: Read the time stamps from the original freight data set, align the timelines of data from different sources, and generate time-synchronized data; Step S122: Perform a unified conversion on the field structure and units of data from different sources to generate data with a unified format; Step S123: Merge the time-synchronized data and the format-unified data to generate synchronized heterogeneous data.

[0031] In one embodiment, the time stamps in the original freight data set are read, and the sampling time information corresponding to the operation status data and the line environment data is extracted. Since the sampling frequency and upload cycle of data from different sources are different, time axis alignment processing is performed on data from different sources according to a preset unified time interval. For data with a sampling frequency higher than the preset time interval, downsampling processing is performed according to a unified time node. For data with a sampling frequency lower than the preset time interval, interpolation compensation processing is performed according to adjacent time nodes, so that data from different sources correspond to a unified time series and time synchronization data is generated.

[0032] The field structures of data from different sources are uniformly converted, mapping the field names of data output from different devices to a unified field identifier. At the same time, the units of measurement in the data from different sources are uniformly converted, with speed units uniformly converted to kilometers per hour, traction current units uniformly converted to amperes, and temperature units uniformly converted to degrees Celsius. Subsequently, the data types of data from different sources are standardized, so that numerical fields, status fields, and character fields meet the requirements of a unified data format, thereby generating data with a unified format.

[0033] The time-synchronized data and the data with the unified format are matched and merged according to a unified time index. At each time node, the traction parameters, operation parameters and energy consumption parameters in the operation status data are associated and combined with the slope parameters, curvature parameters and environmental parameters in the line environment data to form a multi-dimensional data record for the corresponding time segment. Then, the data records corresponding to each time node are continuously arranged and uniformly stored according to the preset data structure to generate synchronous heterogeneous data for subsequent outlier removal and missing value filling.

[0034] Preferably, step S13 includes the following steps: Step S131: Perform statistical distribution analysis on the numerical fields in the synchronous heterogeneous data, identify and mark outlier data points, and thus generate outlier marked data; Step S132: Use linear interpolation to fill in missing values ​​in the outlier-marked data to generate interpolated and completed data; Step S133: Smooth outliers in the interpolated data to generate cleaned freight data.

[0035] In one embodiment, statistical distribution analysis is performed on each numerical field in the synchronous heterogeneous data. Specifically, the mean, variance, and standard deviation of each field are calculated, and deviation of the data distribution is detected by combining the preset confidence interval range. When a data point exceeds the preset statistical threshold range, it is determined as an abnormal data point and is marked to form abnormal marked data, so as to distinguish the different processing paths of normal data and abnormal data in subsequent processing.

[0036] Based on the anomaly-marked data, linear interpolation is performed on the data sequence containing missing values. Specifically, a linear change relationship model is constructed based on the valid data points adjacent to the missing data points, and the missing values ​​are numerically estimated and filled according to the time interval ratio, thereby achieving the continuous repair of the data sequence, generating interpolated and completed data, and ensuring that the data maintains continuous consistency in the time dimension.

[0037] Outliers in the interpolated and completed data undergo further smoothing. Outliers are identified by their local fluctuation amplitude and differences from neighboring data. A sliding window averaging method or a local weighted smoothing method is used to correct outliers so that their values ​​gradually approximate the distribution trend of neighboring data. This reduces the impact of local mutations on the overall data structure and ultimately generates cleaned freight data to ensure the stability and consistency of subsequent standardization processing and energy consumption analysis.

[0038] Preferably, step S14 includes the following steps: Step S141: Perform maximum and minimum value normalization on the continuous variables in the cleaned freight data to generate normalized numerical data; Step S142: Perform one-hot encoding transformation on the discrete variables in the cleaned freight data to generate coded feature data; Step S143: Concatenate the normalized numerical data and the coded feature data in chronological order to generate a standardized freight dataset.

[0039] In one embodiment, the continuous variables in the cleaned freight data are normalized to their maximum and minimum values. The continuous variables include numerical parameters such as traction current, traction voltage, operating speed, traction load, motor temperature, line gradient, and energy consumption per unit time. By extracting the maximum and minimum values ​​of each continuous variable in historical data or the current data set, and performing linear scaling on each data point according to a preset normalization mapping relationship, the continuous variables of different dimensions are uniformly mapped to the same numerical range, thereby generating normalized numerical data to eliminate scale differences between different physical quantities.

[0040] One-hot encoding transformation is performed on discrete variables in the cleaned freight data. The discrete variables include non-continuous fields such as operating mode status, braking status, line grade information, section speed limit category, and environmental status category. By constructing a category set corresponding to the discrete variables, an independent binary vector mapping is performed on each category. The original discrete values ​​are converted into a feature vector representation in which only the value is 1 in the corresponding category position and 0 in the other positions, thereby generating coded feature data to realize the numerical expression of discrete information.

[0041] Normalized numerical data and coded feature data are aligned and concatenated according to a unified time index. Specifically, based on the time series, the normalization results of continuous variables and the coded results of discrete variables at the same timestamp are horizontally combined so that each time node corresponds to a complete multi-dimensional feature vector. Then, all feature vectors are arranged and stored in a structured manner according to the chronological order, thereby generating a standardized freight dataset, which provides a unified input data foundation for subsequent operation condition window construction and energy consumption optimization analysis.

[0042] Preferably, step S2 includes the following steps: Step S21: Read the time index and sampling frequency from the standardized freight dataset, calculate the total data length, and thus generate the basic parameters of the time series. Step S22: Based on the basic parameters of the time series and combined with the preset sliding window length ratio, define the initial segmentation interval on the data series to generate candidate window boundary points; Step S23: Based on the candidate window boundary points, perform a cyclic truncation operation on the standardized freight dataset to generate an initial analysis window sequence with overlap rate; Step S24: Perform boundary integrity verification on the initial analysis window sequence to generate the final freight equipment operating condition analysis window.

[0043] In one embodiment, the time index and sampling frequency in the standardized freight dataset are read, and the time index is parsed continuously to determine the time coverage of the data sequence. At the same time, the time interval between adjacent data points is calculated based on the sampling frequency, and the total number of data points is determined in combination with the data sequence length, thereby generating the basic parameters of the time series used to describe the time distribution structure of the data.

[0044] Based on the basic parameters of the time series and combined with the preset sliding window length ratio coefficient, the standardized freight dataset is divided into intervals. Specifically, based on the total length of the data, the number of data points covered by a single window is determined according to the sliding window length ratio, and multiple initial segmentation interval boundary points are delineated at equal intervals on the time series, thereby generating candidate window boundary points with continuous coverage relationship.

[0045] Based on the candidate window boundary points, a cyclical truncation operation is performed on the standardized freight dataset. Specifically, the starting point of the window is gradually shifted in the time series direction using a preset window sliding step size as the recursive reference, and corresponding data segments are truncated according to the window length. At the same time, a preset overlap rate parameter is introduced to make some data overlap between adjacent windows, thereby generating an initial analysis window sequence with temporal continuity and overlap characteristics.

[0046] The initial analysis window sequence is subjected to boundary integrity verification. Specifically, for each analysis window, it is verified whether its start time and end time completely cover the standardized freight data points within the corresponding time period, and whether there are any data missing or structural incompleteness caused by boundary truncation. Windows that do not meet the integrity requirements are removed or corrected, and finally a freight equipment operation condition analysis window with complete structure and consistent time is generated for subsequent energy consumption feature extraction and analysis.

[0047] Preferably, step S23 includes the following steps: Step S231: Call the preset window sliding step size parameter and overlap coefficient, and combine them with the sampling frequency of the standardized freight dataset to calculate the physical time interval and data point offset between adjacent windows, and generate a window displacement quantification index. Step S232: Using the first candidate window boundary point as the reference origin, iteratively recursively calculate the window displacement quantization index on the time series, and sequentially mark the start and end time cutoffs of each sub-window to construct the window spatiotemporal index matrix. Step S233: Traverse the window spatiotemporal index matrix, perform time-series slicing operation from the standardized freight dataset according to each set of start and end time segments, extract continuous numerical vector segments, and form a multi-dimensional slice data cube; Step S234: Perform dimensional integrity verification on each data block in the multidimensional sliced ​​data cube, truncate and discard residual data that does not meet the preset window dimensions, and generate an initial analysis window sequence.

[0048] In one embodiment, preset window sliding step size parameters and overlap coefficients are invoked, and combined with the sampling frequency of the standardized freight dataset, to perform quantitative analysis on the coverage relationship between adjacent analysis windows in the time series; by converting the window sliding step size and the sampling frequency, the physical time interval between adjacent windows in the time dimension is determined, and the offset between windows at the data point level is further calculated, thereby generating a window displacement quantification index to describe the window movement pattern.

[0049] Using the first candidate window boundary point as the baseline origin for time series analysis, recursive calculations are performed in the time series direction based on the window displacement quantification index. Specifically, the window start position is gradually advanced according to the sliding step size parameter, and the termination time segment corresponding to each window is determined simultaneously, so that each sub-window forms an ordered arrangement relationship on the time axis. By structurally recording the start and end time segments of all windows, a window spatiotemporal index matrix is ​​constructed to achieve unified expression and index management of windows in the time dimension.

[0050] Iterate through each set of start and end time segments in the spatiotemporal index matrix of the window, and perform time-series slicing operations from the standardized freight dataset according to the corresponding time range; extract continuous multidimensional numerical vectors within the corresponding time period during the slicing process, and combine and map the operation status data and line environment data within the window range to form a multidimensional slice data cube with time continuity, which is used to characterize the joint operation condition characteristics within each window.

[0051] The dimensional integrity of each data block in the multidimensional slice data cube is verified to check whether the data in each window meets the preset feature dimension requirements and time coverage integrity requirements. For window data blocks with missing data, inconsistent dimensions, or incomplete structure, truncation or discard processing is performed to avoid abnormal windows from interfering with subsequent analysis results. Finally, an initial analysis window sequence with consistent structure and complete dimensions is generated for subsequent operating condition identification and energy consumption analysis.

[0052] Preferably, step S3 includes: Step S31: Select unit energy consumption, traction load and line gradient as energy consumption features from the standardized freight dataset; Step S32: In the freight equipment operation condition analysis window, calculate the mean and variance of key energy consumption characteristics window by window to generate window feature statistics; Step S33: Compare the window feature statistics with the preset energy consumption feature threshold, filter out the windows that meet the conditions, and generate a set of energy consumption analysis units.

[0053] In one embodiment, key feature variables for energy consumption characterization are selected from a standardized freight dataset, specifically unit energy consumption, traction load, and track gradient as core energy consumption features. Unit energy consumption characterizes the energy consumption level per unit of transportation, traction load characterizes the stress and power demand state of the train traction system, and track gradient characterizes the intensity of the influence of the external operating environment on traction energy consumption. The above feature variables are uniformly extracted and structured from the standardized freight dataset to form the basic feature set for subsequent analysis.

[0054] Within the freight equipment operation condition analysis window, statistical calculations are performed on the data subset corresponding to each window. Specifically, for three types of energy consumption characteristics—unit energy consumption, traction load, and line gradient—the mean and variance are calculated within the corresponding time window to reflect the concentration trend and fluctuation of energy consumption levels within that window. Through parallel calculations of multiple windows, window characteristic statistics corresponding to each operation condition analysis window are generated to characterize the distribution of energy consumption characteristics under different operation conditions.

[0055] The window characteristic statistics corresponding to each operating condition analysis window are compared and analyzed with the pre-set energy consumption characteristic thresholds. The energy consumption characteristic thresholds are used to limit the fluctuation range of unit energy consumption, traction load and line gradient within a reasonable operating range. When the characteristic statistics of a certain window simultaneously meet the unit energy consumption threshold constraint, traction load threshold constraint and line gradient constraint, the window is determined to meet the energy consumption analysis conditions, and the window is selected and included in the target set. Finally, an energy consumption analysis unit set is generated to serve as the input basis for subsequent energy consumption optimization calculation and control parameter generation.

[0056] This invention also provides a data fusion-based railway freight equipment energy consumption optimization system for executing the data fusion-based railway freight equipment energy consumption optimization method described above. The data fusion-based railway freight equipment energy consumption optimization system includes: The data acquisition module 101 is used to acquire the operating status data and line environment data of the target railway freight equipment, preprocess the operating status data and line environment data, and generate a standardized freight dataset. The operating condition analysis module 102 is used to construct an operating condition analysis window for freight equipment based on the time series characteristics of the standardized freight dataset. Feature extraction module 103 is used to extract data segments associated with energy consumption characteristics in the standardized freight data set in the freight equipment operation condition analysis window, and use them as a set of energy consumption analysis units. The optimization decision module 104 is used to map the input parameters of the energy consumption analysis unit set to the standardized freight dataset, call the preset energy consumption optimization strategy to perform calculations, and generate energy consumption optimization control parameters for railway freight equipment.

[0057] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for optimizing energy consumption of railway freight equipment based on data fusion, characterized in that, Includes the following steps: Step S1: Obtain the operating status data and line environment data of the target railway freight equipment, preprocess the operating status data and line environment data, and generate a standardized freight dataset; Step S2: Based on the time series characteristics of the standardized freight dataset, construct a window for analyzing the operating conditions of freight equipment; Step S3: In the freight equipment operation condition analysis window, extract the data segments that are associated with the energy consumption characteristics in the standardized freight dataset, and use them as the energy consumption analysis unit set; Step S4: Map the input parameters of the energy consumption analysis unit set to the standardized freight dataset, call the preset energy consumption optimization strategy to perform calculations, and generate energy consumption optimization control parameters for railway freight equipment.

2. The energy consumption optimization method for railway freight equipment based on data fusion according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Obtain the operating status data and track environment data of the target railway freight equipment; Step S12: Perform time synchronization and format unification on the operating status data and line environment data to generate synchronized heterogeneous data; Step S13: Remove outliers and fill in missing values ​​in the synchronous heterogeneous data to generate cleaned freight data; Step S14: Normalize and encode the cleaned freight data to generate a standardized freight dataset.

3. The energy consumption optimization method for railway freight equipment based on data fusion according to claim 2, characterized in that, Step S11 includes the following steps: Step S111: Collect operational status data of the target railway freight equipment using onboard sensors; Step S112: Collect track environment data of the target railway freight equipment through the track monitoring system; Step S113: Associate and store the operating status data and the line environment data according to the collection timestamp to generate the original freight data set.

4. The energy consumption optimization method for railway freight equipment based on data fusion according to claim 2, characterized in that, Step S12 includes the following steps: Step S121: Read the time stamps from the original freight data set, align the timelines of data from different sources, and generate time-synchronized data; Step S122: Perform a unified conversion on the field structure and units of data from different sources to generate data with a unified format; Step S123: Merge the time-synchronized data and the format-unified data to generate synchronized heterogeneous data.

5. The energy consumption optimization method for railway freight equipment based on data fusion according to claim 2, characterized in that, Step S13 includes the following steps: Step S131: Perform statistical distribution analysis on the numerical fields in the synchronous heterogeneous data, identify and mark outlier data points, and thus generate outlier marked data; Step S132: Use linear interpolation to fill in missing values ​​in the outlier-marked data to generate interpolated and completed data; Step S133: Smooth outliers in the interpolated data to generate cleaned freight data.

6. The energy consumption optimization method for railway freight equipment based on data fusion according to claim 2, characterized in that, Step S14 includes the following steps: Step S141: Perform maximum and minimum value normalization on the continuous variables in the cleaned freight data to generate normalized numerical data; Step S142: Perform one-hot encoding transformation on the discrete variables in the cleaned freight data to generate coded feature data; Step S143: Concatenate the normalized numerical data and the coded feature data in chronological order to generate a standardized freight dataset.

7. The energy consumption optimization method for railway freight equipment based on data fusion according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: Read the time index and sampling frequency from the standardized freight dataset, calculate the total data length, and thus generate the basic parameters of the time series. Step S22: Based on the basic parameters of the time series and combined with the preset sliding window length ratio, define the initial segmentation interval on the data series to generate candidate window boundary points; Step S23: Based on the candidate window boundary points, perform a cyclic truncation operation on the standardized freight dataset to generate an initial analysis window sequence with overlap rate; Step S24: Perform boundary integrity verification on the initial analysis window sequence to generate the final freight equipment operating condition analysis window.

8. The energy consumption optimization method for railway freight equipment based on data fusion according to claim 7, characterized in that, Step S23 includes the following steps: Step S231: Call the preset window sliding step size parameter and overlap coefficient, and combine them with the sampling frequency of the standardized freight dataset to calculate the physical time interval and data point offset between adjacent windows, and generate a window displacement quantification index. Step S232: Using the first candidate window boundary point as the reference origin, iteratively recursively calculate the window displacement quantization index on the time series, and sequentially mark the start and end time cutoffs of each sub-window to construct the window spatiotemporal index matrix. Step S233: Traverse the window spatiotemporal index matrix, perform time-series slicing operation from the standardized freight dataset according to each set of start and end time segments, extract continuous numerical vector segments, and form a multi-dimensional slice data cube; Step S234: Perform dimensional integrity verification on each data block in the multidimensional sliced ​​data cube, truncate and discard residual data that does not meet the preset window dimensions, and generate an initial analysis window sequence.

9. The energy consumption optimization method for railway freight equipment based on data fusion according to claim 1, characterized in that, Step S3 includes: Step S31: Select unit energy consumption, traction load and line gradient as energy consumption features from the standardized freight dataset; Step S32: In the freight equipment operation condition analysis window, calculate the mean and variance of key energy consumption characteristics window by window to generate window feature statistics; Step S33: Compare the window feature statistics with the preset energy consumption feature threshold, filter out the windows that meet the conditions, and generate a set of energy consumption analysis units.

10. A railway freight equipment energy consumption optimization system based on data fusion, characterized in that, The system is used to implement the data fusion-based energy consumption optimization method for railway freight equipment as described in claim 1, and includes: The data acquisition module is used to acquire the operating status data and line environment data of the target railway freight equipment, preprocess the operating status data and line environment data, and generate a standardized freight dataset. The operating condition analysis module is used to construct an operating condition analysis window for freight equipment based on the time series characteristics of standardized freight datasets. The feature extraction module is used to extract data segments that are associated with energy consumption characteristics in the standardized freight data set from the freight equipment operation condition analysis window, and use them as a set of energy consumption analysis units. The optimization decision module maps the input parameters of the energy consumption analysis unit set to the standardized freight dataset, calls the preset energy consumption optimization strategy for calculation, and generates energy consumption optimization control parameters for railway freight equipment.

Images