Vehicle energy consumption monitoring method and system based on multi-source data

The method and system integrate internal vehicle and external environmental data to accurately monitor and optimize energy consumption, addressing the limitations of existing technologies by simulating and analyzing energy consumption under specific conditions.

CN118859813BActive Publication Date: 2024-12-03ANHUI HEFEITER AUTOMOBILE CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411342354.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-25
Publication Date
2024-12-03
Estimated Expiration
2044-09-25

AI Technical Summary

Technical Problem

Existing vehicle energy consumption monitoring methods focus solely on internal vehicle measurements, ignoring external environmental factors, leading to inaccurate energy efficiency judgments and ineffective optimization strategies.

Method used

A vehicle energy consumption monitoring method and system that utilizes multi-source data, including internal vehicle measurements and external environmental factors, using time window filtering, equal-dimensional recursive grey algorithms, and energy consumption structure trees to simulate and analyze energy consumption under specific driving conditions.

Benefits of technology

Provides accurate energy consumption management by integrating internal and external data, enabling real-time anomaly detection and responsive optimization strategies that adapt to actual driving conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118859813B_ABST
    Figure CN118859813B_ABST
Patent Text Reader

Abstract

The present invention discloses a vehicle energy consumption monitoring method and system based on multi-source data, which relates to the field of vehicle energy consumption monitoring. The vehicle energy consumption monitoring method based on multi-source data includes: obtaining initial vehicle measurement data sent by a sensor module installed on a vehicle, and optimizing the initial vehicle measurement data using a time window filtering technique to obtain a vehicle internal measurement data set; identifying energy consumption influencing factors of the vehicle's external environment and generating a vehicle energy consumption external influence sequence; and generating a number of vehicle energy consumption monitoring scenarios based on the vehicle internal measurement data set and the vehicle energy consumption external influence sequence. The present invention combines the vehicle internal measurement data and the external influencing factor sequence, and uses neighborhood mutual information and autoencoder technology for feature fusion to construct an energy consumption monitoring scenario that accurately matches actual driving conditions, so that energy consumption anomalies can be quickly identified and responded to between associated scenarios.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of vehicle energy consumption monitoring, and in particular to a vehicle energy consumption monitoring method and system based on multi-source data. Background Art

[0002] Vehicle energy consumption refers to the energy consumed by a vehicle during operation, and is usually used to measure the efficiency of fuel or electricity use. By monitoring and managing vehicle energy consumption, the impact on the environment can be reduced, and low-energy vehicles help reduce greenhouse gas emissions and the generation of other pollutants, which is crucial to addressing global climate change and protecting the environment. At the same time, energy consumption monitoring helps identify excessive energy consumption, so that corresponding adjustments or maintenance can be made to reduce fuel or electricity use and reduce operating costs. Therefore, vehicle energy consumption monitoring can not only improve the efficiency of energy use, but also help environmental protection and cost control. It is an indispensable part of vehicle management.

[0003] However, the vehicle energy consumption monitoring methods in the existing technology often only focus on internal vehicle measurements and ignore external environmental factors, making it impossible to fully evaluate the actual energy consumption of the vehicle. At the same time, there may be a misjudgment of the vehicle's energy efficiency based solely on internal data, resulting in the energy-saving measures formulated later not being accurate or effective, thereby reducing the optimization of vehicle maintenance and operation strategies.

[0004] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a vehicle energy consumption monitoring method and system based on multi-source data, which has the advantages of simulating and analyzing energy consumption performance under specific driving conditions and providing more accurate energy consumption management, thereby solving the problem that the prior art cannot simulate and analyze energy consumption performance under specific driving conditions.

[0006] In order to achieve the above simulation and analysis of energy consumption performance under specific driving conditions and provide the advantage of more accurate energy consumption management, the specific technical solutions adopted by the present invention are as follows:

[0007] According to one aspect of the present invention, a vehicle energy consumption monitoring method based on multi-source data is provided, and the vehicle energy consumption monitoring method based on multi-source data includes:

[0008] S1. Obtaining initial vehicle measurement data sent by a sensor module installed on a vehicle, and optimizing the initial vehicle measurement data using a time window filtering technique to obtain a vehicle internal measurement data set;

[0009] S2, identifying the energy consumption influencing factors of the vehicle's external environment and generating a vehicle energy consumption external influencing sequence, and segmenting the vehicle energy consumption external influencing sequence using an equal-dimensional recursive grey algorithm;

[0010] S3. Generate several vehicle energy consumption monitoring scenarios based on the vehicle internal measurement data set and the vehicle energy consumption external influence sequence, and construct an energy consumption structure tree branch to find energy consumption anomalies in each energy consumption monitoring scenario.

[0011] Furthermore, the initial vehicle measurement data sent by the sensor module installed on the vehicle is obtained, and the initial vehicle measurement data is optimized using the time window filtering technology to obtain the vehicle internal measurement data set including:

[0012] S11, configuring a sensor module, wherein the sensor module at least includes a speed sensor, an engine speed sensor, and a fuel level sensor;

[0013] S12, respectively obtaining the transmission hop counts of the speed sensor, the engine speed sensor, and the fuel level sensor, calculating the distributed distances between the sensors based on the transmission hop counts, and locating the sensors;

[0014] S13, using a speed sensor, an engine speed sensor and a fuel level sensor to collect vehicle speed characteristics, engine speed characteristics and fuel consumption characteristics;

[0015] S14, mapping the vehicle speed characteristics, engine speed characteristics and fuel consumption characteristics onto a unified time axis, and integrating the characteristics of each vehicle to obtain a vehicle internal measurement data set.

[0016] Furthermore, respectively obtaining the transmission hop counts of the speed sensor, the engine speed sensor, and the fuel level sensor, calculating the distributed distances between the sensors based on the transmission hop counts, and locating the sensors includes:

[0017] S121, using each sensor as a collection node in the vehicle internal network, and each collection node regularly transmits initial vehicle measurement data and the number of transmission hops according to a preset period;

[0018] S122, taking the data receiving end as a beacon node, and sending the transmission hop count of the beacon node to each collection node, and establishing a location information table after the collection node receives the transmission hop count of the beacon node;

[0019] S123, calculating the distributed distance and average distance per hop between the collection node and the beacon node according to the number of transmission hops recorded in the location information table and updating the location information table;

[0020] S124. Each collection node obtains the number of transmission hops of other collection nodes according to the updated location information table, uses the positioning accuracy of the collection node as a weighting factor, and uses the three-sided positioning method to perform distributed positioning of each sensor.

[0021] Furthermore, mapping the vehicle speed characteristics, engine speed characteristics and fuel consumption characteristics onto a unified time axis includes:

[0022] S141, determining an asymmetry ratio between a collection node and a beacon node, and establishing clock models of vehicle speed characteristics, engine speed characteristics, and fuel consumption characteristics based on the determination result of the asymmetry ratio;

[0023] S142, using the Laida criterion to identify the jump abnormal values ​​of each clock model, and combining the autoregressive integrated moving average model to eliminate the time noise jitter generated by each clock model to obtain a stable time series;

[0024] S143. Predictively compensate each clock model based on a stationary time series, and use the compensated clock model to map vehicle speed characteristics, engine speed characteristics, and fuel consumption characteristics onto a unified time axis.

[0025] Furthermore, based on the vehicle internal measurement data set and the vehicle energy consumption external influence sequence, several vehicle energy consumption monitoring scenarios are generated, and an energy consumption structure tree branch is constructed to find energy consumption anomalies in each energy consumption monitoring scenario, including:

[0026] S31, classifying the vehicle internal measurement data set and the vehicle energy consumption external influence sequence according to preset conditions, and constructing several vehicle energy consumption monitoring scenarios using feature fusion technology;

[0027] S32, analyzing the vehicle energy consumption characteristics according to the energy consumption influencing factors of each external environment in each vehicle energy consumption monitoring scenario, and generating energy consumption abnormality monitoring rules corresponding to the vehicle energy consumption monitoring scenario;

[0028] S33, constructing an energy consumption structure tree by using the vehicle energy consumption monitoring scenario as a tree node, and integrating the energy consumption abnormality monitoring rules into the energy consumption structure tree to obtain a multi-scenario rule linkage mechanism of the energy consumption structure tree;

[0029] S34, based on the multi-scenario rule linkage mechanism and through the energy consumption structure tree branching, searching for abnormal energy consumption scenarios in the vehicle energy consumption monitoring scenarios, and establishing a real-time energy consumption monitoring model to identify abnormal points in the abnormal energy consumption scenarios;

[0030] S35. Use adaptive technology to adjust the energy consumption anomaly monitoring rules, and continuously update the multi-scenario rule linkage mechanism of the energy consumption structure tree based on the adjusted energy consumption anomaly monitoring rules.

[0031] Furthermore, the vehicle internal measurement data set and the vehicle energy consumption external influence sequence are classified according to preset conditions, and several vehicle energy consumption monitoring scenarios are constructed using feature fusion technology, including:

[0032] S311, introducing classification intervals from the vehicle internal measurement data set and the vehicle energy consumption external influence sequence respectively, and setting the classification intervals as the sample neighborhood space;

[0033] S312, calculating the neighborhood mutual information between the vehicle internal measurement data set and the vehicle energy consumption external influence sequence, and sorting the segmented vehicle energy consumption external influence subsequences according to the size of the neighborhood mutual information, and fusing the vehicle energy consumption external influence subsequences based on the classification interval to generate several new subsequences;

[0034] S313: using an automatic encoder to fuse the features of each vehicle in the vehicle internal measurement data set with the new subsequence to obtain several vehicle energy consumption monitoring scenarios.

[0035] Furthermore, the vehicle energy consumption monitoring scenarios include urban driving scenarios, highway driving scenarios, and mountain driving scenarios;

[0036] Among them, the urban driving scenario refers to the driving environment on urban roads, which has the characteristics of frequent stopping and starting and low speed limit;

[0037] Highway driving scene refers to the driving environment on highways or expressways, which has the characteristics of high and constant speed of vehicles;

[0038] Mountain driving scenarios refer to driving environments in mountainous areas or on sloped roads, which are characterized by frequent changes in slopes, road widths, and road conditions.

[0039] Furthermore, the energy consumption structure tree is constructed by taking the vehicle energy consumption monitoring scenario as a tree node, and the energy consumption abnormality monitoring rules are integrated into the energy consumption structure tree to obtain a multi-scenario rule linkage mechanism of the energy consumption structure tree including:

[0040] S331, representing a branch of a structure tree for each vehicle energy consumption monitoring scenario, extracting vehicle features and new subsequences as branch points and generating an energy consumption structure tree;

[0041] S332, analyzing the energy consumption characteristic correlation indexes between branches, establishing energy consumption anomaly monitoring rules to express the intimacy between branches, and establishing an intimacy matrix based on the intimacy between branches to realize the trunk expression of the energy consumption structure tree;

[0042] S333, taking the intimacy between branches as input, introducing the path planning algorithm to obtain the optimal solution of the local shortest path between the trunk and each branch, and constructing a hierarchical matrix of the trunk and branches to realize the hierarchical structure of the energy consumption structure;

[0043] S334. Integrate the energy consumption anomaly monitoring rules into the energy consumption structure tree to obtain a multi-scenario rule linkage mechanism that conforms to the hierarchy of the energy consumption structure tree.

[0044] Furthermore, the calculation formula of neighborhood mutual information is:

[0045] ;

[0046] In the formula, A represents a vehicle interior measurement dataset; B represents the sequence of external influences on vehicle energy consumption; NMI μ ( A , B ) represents the neighborhood mutual information between the vehicle internal measurement dataset and the external influence sequence of vehicle energy consumption; μ A ( x i ) indicates vehicle characteristics x i Neighborhood on a dataset of vehicle interior measurements; μ B ( x i ) indicates vehicle characteristics x i Neighborhood on the sequence of external influences on vehicle energy consumption; Indicates vehicle characteristics x i In the sample neighborhood space The neighborhood formed on n Represents the number of vehicle features; log represents the logarithmic function.

[0047] According to another aspect of the present invention, a vehicle energy consumption monitoring system based on multi-source data is also provided, the vehicle energy consumption monitoring system based on multi-source data includes a vehicle internal measurement module, a vehicle external environment analysis module and a vehicle energy consumption monitoring module;

[0048] The vehicle interior measurement module is used to obtain the initial vehicle measurement data sent by the sensor module installed on the vehicle, and optimize the initial vehicle measurement data using the time window filtering technology to obtain the vehicle interior measurement data set;

[0049] The vehicle external environment analysis module is used to identify the energy consumption influencing factors of the vehicle external environment and generate a vehicle energy consumption external influence sequence, and segment the vehicle energy consumption external influence sequence using an equal-dimensional recursive grey algorithm;

[0050] The vehicle energy consumption monitoring module is used to generate several vehicle energy consumption monitoring scenarios based on the vehicle's internal measurement data set and the vehicle's energy consumption external influence sequence, and to construct an energy consumption structure tree branch to find energy consumption anomalies in each energy consumption monitoring scenario.

[0051] Compared with the prior art, the present invention provides a vehicle energy consumption monitoring method and system based on multi-source data, which has the following beneficial effects:

[0052] (1) The present invention can comprehensively monitor the key performance parameters of the vehicle through the speed sensor, engine speed sensor and fuel level sensor, provide the necessary input data for energy consumption analysis, and accurately locate the relative position of each sensor through the interaction between the acquisition node and the beacon node, combined with the calculation of the transmission hop count and the distributed distance, which plays an important role in the optimization of data transmission and fault detection in the vehicle's internal network. At the same time, by establishing a clock model and time series analysis technology, time deviation and noise are effectively processed to ensure the temporal consistency and accuracy of data obtained from different sensors, providing data support for real-time vehicle energy consumption monitoring and optimization.

[0053] (2) The present invention combines the internal measurement data of the vehicle and the sequence of external influencing factors, and uses neighborhood mutual information and autoencoder technology to perform feature fusion, thereby constructing an energy consumption monitoring scenario that accurately matches the actual driving conditions, and organizing different energy consumption monitoring scenarios into a structure tree format, thereby enhancing the organization of the system and the logic of data processing, and realizing effective links and interactions between energy consumption monitoring scenarios, so that energy consumption anomalies can be quickly identified and responded to between related scenarios.

[0054] (3) The present invention generates energy consumption monitoring scenarios through internal and external data, which can simulate and analyze energy consumption performance under specific driving conditions, provide more accurate energy consumption management, and construct an energy consumption structure tree, allowing the system to process and analyze data in an organized manner, thereby improving the efficiency of data retrieval and processing. In addition, the branch search in the energy consumption structure tree enables rapid identification and response to abnormal energy consumption scenarios, supports real-time energy consumption monitoring and exception handling, and ensures that the monitoring strategy keeps pace with the actual situation, thereby enhancing the adaptability and flexibility of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] Figure 1 is a flow chart of a vehicle energy consumption monitoring method based on multi-source data according to an embodiment of the present invention;

[0057] Figure 2 It is a principle block diagram of a vehicle energy consumption monitoring system based on multi-source data according to an embodiment of the present invention.

[0058] In the figure:

[0059] 1. Vehicle internal measurement module; 2. Vehicle external environment analysis module; 3. Vehicle energy consumption monitoring module. DETAILED DESCRIPTION

[0060] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments. They can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in this field should be able to understand other possible implementation methods and advantages of the present invention.

[0061] According to an embodiment of the present invention, a vehicle energy consumption monitoring method and system based on multi-source data are provided.

[0062] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to the vehicle energy consumption monitoring method based on multi-source data according to an embodiment of the present invention, the vehicle energy consumption monitoring method based on multi-source data includes:

[0063] S1. Acquire initial vehicle measurement data sent by a sensor module installed on a vehicle, and optimize the initial vehicle measurement data using a time window filtering technique to obtain a vehicle internal measurement data set.

[0064] The initial vehicle measurement data sent by the sensor module installed on the vehicle is obtained, and the initial vehicle measurement data is optimized by using the time window filtering technology to obtain the vehicle internal measurement data set including:

[0065] S11. Configure a sensor module, and the sensor module at least includes a speed sensor, an engine speed sensor, and a fuel level sensor.

[0066] It should be noted that the sensor module includes at least a speed sensor, an engine speed sensor and a fuel level sensor. The sensor module is used to comprehensively monitor the key performance parameters of the vehicle, and then analyze and optimize the energy efficiency of the vehicle, including:

[0067] Speed ​​sensors are used to measure the instantaneous speed of a vehicle and provide vehicle speed data, which is essential for analyzing driving patterns, calculating driving efficiency, and predicting fuel consumption.

[0068] The engine speed sensor is used to measure the engine speed. The speed data helps analyze the engine operating efficiency, monitor the engine load and evaluate the engine performance status.

[0069] Fuel level sensors are used to measure the amount of fuel in the fuel tank, providing real-time fuel consumption information, which is very important for calculating fuel economy and monitoring the health of the fuel system.

[0070] In addition to the above-mentioned speed sensors, engine speed sensors and fuel level sensors, there are many other measuring devices that are closely related to vehicle energy consumption, such as temperature sensors, pressure sensors and torque sensors, etc., which will not be elaborated in detail here.

[0071] S12, respectively obtaining the transmission hop counts of the speed sensor, the engine speed sensor, and the fuel level sensor, calculating the distributed distances between the sensors based on the transmission hop counts, and locating the sensors.

[0072] The transmission hop counts of the speed sensor, the engine speed sensor, and the fuel level sensor are respectively obtained, and the distributed distances between the sensors are calculated based on the transmission hop counts and the positioning between the sensors is performed, including:

[0073] S121, using each sensor as a collection node in the vehicle internal network, and each collection node regularly transmits initial vehicle measurement data and the number of transmission hops according to a preset period;

[0074] S122, taking the data receiving end as a beacon node, and sending the transmission hop count of the beacon node to each collection node, and establishing a location information table after the collection node receives the transmission hop count of the beacon node;

[0075] S123, calculating the distributed distance and average distance per hop between the collection node and the beacon node according to the number of transmission hops recorded in the location information table and updating the location information table;

[0076] S124. Each collection node obtains the number of transmission hops of other collection nodes according to the updated location information table, uses the positioning accuracy of the collection node as a weighting factor, and uses the three-sided positioning method to perform distributed positioning of each sensor.

[0077] It should be noted that each collection node obtains the number of transmission hops of other collection nodes according to the updated location information table, takes the positioning accuracy of the collection node as a weighting factor, and uses the trilateral positioning method to perform distributed positioning of each sensor, including:

[0078] Each acquisition node regularly transmits its initial vehicle measurement data and the number of transmission hops according to a preset period. The number of transmission hops reflects the number of signal hops from one node to another and is used to estimate the relative distance between nodes.

[0079] The data receiving end is used as a beacon node to transmit its transmission hop count to all collection nodes. After each collection node receives the transmission hop count from the beacon node, it records and updates its own location information table. The location information table contains the transmission hop count between the beacon node and other collection nodes. This information will be used for subsequent positioning calculations.

[0080] According to the number of transmission hops in the location information table, the estimated distance between the collection node and the beacon node is calculated using the average per-hop distance method, and three or more reference nodes (nodes with known locations, at least three, including the beacon node) are selected.

[0081] Specifically, use the calculated distance as the radius, draw circles centered on these reference nodes, and determine the intersection of these circles. Ideally, the three circles should intersect at one point, which is the location of the node to be located, and the location of the unknown node is accurately determined by geometric methods. In the three-sided positioning calculation, different weights are given according to the positioning accuracy of each acquisition node (which may be determined by factors such as signal strength and historical positioning accuracy).

[0082] Each sensor node (speed sensor, engine speed sensor, fuel level sensor) obtains the hop count information of at least three other target sensors according to its own routing information table.

[0083] Using the hop count information and the estimated average distance per hop, calculate the physical distance between each pair of sensor nodes. For example, if the hop count between the speed sensor and the engine speed sensor is 2 and the average distance per hop is 5 meters, then the estimated distance between the two sensors is 2×5=10 meters.

[0084] It should be noted that the distributed positioning using the trilateral positioning method includes:

[0085] When an unknown node (such as a speed sensor) obtains the distance information of at least three target sensors (such as an engine speed sensor, a fuel level sensor, and a beacon node), the three-sided positioning method can be used to estimate its relative position in the vehicle. After each sensor calculates the relative position, it is combined with the existing sensor layout in the vehicle (such as known installation positions and vehicle geometry) to further correct the position estimation result and improve the positioning accuracy.

[0086] S13, using a speed sensor, an engine speed sensor and a fuel level sensor to collect vehicle speed characteristics, engine speed characteristics and fuel consumption characteristics;

[0087] S14, mapping the vehicle speed characteristics, engine speed characteristics and fuel consumption characteristics onto a unified time axis, and integrating the characteristics of each vehicle to obtain a vehicle internal measurement data set.

[0088] Among them, mapping the vehicle speed characteristics, engine speed characteristics and fuel consumption characteristics onto a unified time axis includes:

[0089] S141, determining the asymmetry ratio between the collection node and the beacon node, and establishing clock models of vehicle speed characteristics, engine speed characteristics, and fuel consumption characteristics based on the determination result of the asymmetry ratio.

[0090] It should be noted that the asymmetry ratio between the collection node and the beacon node is determined, and based on the determination result of the asymmetry ratio, the clock models of the vehicle speed characteristic, the engine speed characteristic and the fuel consumption characteristic are respectively established, including:

[0091] The vehicle internal measurement data set is analyzed to evaluate the asymmetry ratio between the data sources (i.e., the time deviation of data transmission). For example, engine speed data and vehicle speed data may have different transmission times due to different sensor locations and network delays.

[0092] A basic clock model is established according to the collection and transmission characteristics of each characteristic data. The basic clock model synchronizes various data based on a unified time reference. Assuming that the engine speed sensor data is collected every 0.1 seconds and the speed sensor data is collected every 0.05 seconds, the clock model will adjust these data streams to ensure time consistency.

[0093] According to the asymmetry ratio evaluation results, the clock model is adjusted so that all feature data can reflect the state at the same time point. t engine is the raw timestamp of the engine speed, t speed is the raw timestamp of the velocity, and the asymmetry ratio is given by α The adjusted timestamp can be expressed as:

[0094] t engine ′= t engine + α ×( t speed t engine );

[0095] If it is found that the speed data arrives 0.02 seconds later than the engine speed data, and the asymmetry ratio is 0.5, the timestamp of the engine data will be adjusted to be 0.01 seconds earlier than the original.

[0096] Use real-time data to verify the accuracy of the clock model, adjust the asymmetry ratio and other related parameters based on actual performance, evaluate the effectiveness of the model by comparing the data before and after adjustment with the actual vehicle operation, and make necessary adjustments to improve the accuracy of time synchronization.

[0097] S142, using the Laida criterion to identify the jump abnormal values ​​of each clock model, and combining the autoregressive integrated moving average model to eliminate the time noise jitter generated by each clock model to obtain a stable time series.

[0098] It should be noted that the Laida criterion is used to identify the jump abnormal values ​​of each clock model, and the autoregressive integrated moving average model is combined to eliminate the time noise jitter generated by each clock model to obtain a stable time series including:

[0099] The Laida criterion is used to detect jump anomalies in time series. The Laida criterion is a statistical test method used to detect whether there is autocorrelation in the time series, that is, whether the data points are correlated or there is a periodic pattern. This method is often used to detect outliers or jumps in time series data. It mainly includes:

[0100] Select the delay order. The delay order is the number of time intervals considered in the test. It has an important influence on the sensitivity of the results. The delay order is selected by observing the autocorrelation function graph of the data. Usually, the delay number is selected after the autocorrelation function graph first approaches zero.

[0101] Execute using relevant functions in statistical software or programming languages ​​and calculate the Q statistic at each delay to determine the overall autocorrelation of the series;

[0102] Analyzing the test results, a smaller Q statistic (usually less than 0.05) indicates that the sequence has significant autocorrelation, which means the existence of periodicity or outliers. If the Q statistic is very low, it indicates that there may be abnormal jumps in the time series, and the source and nature of these jumps need to be further analyzed.

[0103] Build an ARIMA model (Autoregressive Integrated Moving Average model), which combines autoregressive (AR), differencing (I), and moving average (MA) components to simulate and forecast time series, using historical data to estimate the parameters of the ARIMA model;

[0104] The difference method is used to reduce the non-stationarity (such as trend and seasonality) in the time series, and the smoothing characteristics of the ARIMA model are used to remove noise. After the time series is differentiated once, the ARIMA model is applied for fitting to eliminate residual random fluctuations.

[0105] According to the output of the Laida criterion and the ARIMA model, outliers in the time series are identified and processed, and then a smoother and more stable time series is reconstructed. For example, outliers greater than 3 standard deviations are found in the residuals identified from the ARIMA model. These points may represent abnormal data jumps and should be regarded as anomalies and need to be eliminated or corrected.

[0106] S143. Predictively compensate each clock model based on a stationary time series, and use the compensated clock model to map vehicle speed characteristics, engine speed characteristics, and fuel consumption characteristics onto a unified time axis.

[0107] S2. Identify the energy consumption influencing factors of the vehicle's external environment and generate a vehicle energy consumption external influencing sequence, and use an equal-dimensional recursive grey algorithm to segment the vehicle energy consumption external influencing sequence.

[0108] It should be noted that identifying the energy consumption influencing factors of the vehicle's external environment and generating a vehicle energy consumption external influence sequence, and using the equal-dimensional recursive grey algorithm to segment the vehicle energy consumption external influence sequence includes:

[0109] Identify and collect external environmental factors that may affect vehicle energy consumption. These factors may include temperature, humidity, wind speed, road conditions, traffic flow, etc., and collect relevant data through vehicle sensors and external data sources (such as weather station data, traffic information systems, etc.).

[0110] Combine the processed data of various environmental factors to construct a time series that reflects the external influence of vehicle energy consumption. For example, create a time series data table, where each row represents a time point and the columns include the average temperature, humidity, wind speed, road condition index, and traffic flow index at that time point, for example:

[0111] Data Collection

[0112] Target data: temperature, humidity, wind speed, road conditions, and traffic flow.

[0113] Air temperature and humidity: Uses the vehicle's built-in weather sensors to record data every hour.

[0114] Wind Speed: Download hourly wind speed data from the nearest weather station.

[0115] Traffic conditions: Obtain real-time traffic information, such as traffic congestion index, through the vehicle's GPS system and online traffic information system.

[0116] Traffic flow: Obtain hourly traffic flow data from the city traffic management system.

[0117] Time series construction

[0118] Time series structure: Create a comprehensive time series data table with each row corresponding to a time point (for example, every hour) and each column corresponding to an environmental factor.

[0119] Data example: Timestamp: 12:00; Temperature: 22°C; Humidity: 60%; Wind speed: 5m / s; Road condition index: 0.8 (1 means smooth traffic, 0 means severe congestion); Traffic flow: 1500 vehicles / hour.

[0120] Data Merge

[0121] All pre-processed environmental factor data are merged into a complete time series data table according to timestamp alignment, and database management tools or data processing software, such as SQL or Python's Pandas library, are used to merge and integrate the data.

[0122] The equal-dimensional complement grey algorithm is used to segment the external influence sequence to identify and predict the trend of energy consumption. The equal-dimensional complement grey algorithm is a method in grey system theory, which is used to process and predict data sequences, especially when the data volume is small or the information is incomplete. This method can reveal the implicit relationships and patterns in the data through grey generation and grey correlation analysis. Its main steps include:

[0123] Perform an accumulation operation on the original data to convert the original time series Convert to cumulative generation sequence This is the basic step in building a grey model, which aims to amplify the development trend of the data. When each element is the original time series before k Elements ( i Starting from 1, increment to k , used to indicate the position of the current element being accumulated in the original time series):

[0124] ;

[0125] The grey differential equation is established based on the cumulative generation sequence, namely the grey model GM (1,1) model. This model describes the inherent law of the data generation process. The grey differential equation is usually in the form of:

[0126] ;

[0127] In the formula, a and b are the model parameters obtained by data fitting, is the original data sequence, then Generates a sequence for one accumulation.

[0128] Estimate model parameters using the least squares method a and b , and use the known cumulative generation sequence data to calculate the optimal a and b The grey differential equation is solved to obtain the analytical solution of the model.

[0129] Use the segmented model to analyze and predict the external impacts on vehicle energy consumption to identify key energy consumption drivers. For example, based on the model, predict the energy consumption impact trend in the next week and determine which environmental factors have the greatest impact on energy consumption.

[0130] S3. Generate several vehicle energy consumption monitoring scenarios based on the vehicle internal measurement data set and the vehicle energy consumption external influence sequence, and construct an energy consumption structure tree branch to find energy consumption anomalies in each energy consumption monitoring scenario.

[0131] Among them, based on the vehicle internal measurement data set and the vehicle energy consumption external influence sequence, several vehicle energy consumption monitoring scenarios are generated, and the energy consumption structure tree branches are constructed to find energy consumption anomalies in each energy consumption monitoring scenario, including:

[0132] S31. Classify the vehicle internal measurement data set and the vehicle energy consumption external influence sequence according to preset conditions, and use feature fusion technology to construct several vehicle energy consumption monitoring scenarios.

[0133] Among them, the vehicle internal measurement data set and the vehicle energy consumption external influence sequence are classified according to preset conditions, and several vehicle energy consumption monitoring scenarios are constructed using feature fusion technology, including:

[0134] S311, introducing classification intervals from the vehicle internal measurement data set and the vehicle energy consumption external influence sequence respectively, and setting the classification intervals as the sample neighborhood space;

[0135] S312, calculating the neighborhood mutual information between the vehicle internal measurement data set and the vehicle energy consumption external influence sequence, and sorting the segmented vehicle energy consumption external influence subsequences according to the size of the neighborhood mutual information, and fusing the vehicle energy consumption external influence subsequences based on the classification interval to generate several new subsequences.

[0136] It should be noted that neighborhood mutual information is a statistical method for measuring nonlinear dependencies between two time series. This method can not only detect linear relationships, but also capture more complex nonlinear dependencies, which may be overlooked in traditional correlation analysis. MIN evaluates the mutual dependence between two variables by calculating the degree of information sharing between them, and is particularly suitable for processing dynamic data in complex systems, such as internal measurement data of vehicles and external environmental factors.

[0137] Among them, the calculation formula of neighborhood mutual information is:

[0138] ;

[0139] In the formula, A represents a vehicle interior measurement dataset; B represents the sequence of external influences on vehicle energy consumption; NMI μ (A , B ) represents the neighborhood mutual information between the vehicle internal measurement dataset and the external influence sequence of vehicle energy consumption; μ A ( x i ) indicates vehicle characteristics x i Neighborhood on a dataset of vehicle interior measurements; μ B ( x i ) indicates vehicle characteristics x i Neighborhood on the sequence of external influences on vehicle energy consumption; Indicates vehicle characteristics x i In the sample neighborhood space The neighborhood formed on n represents the number of vehicle features; log represents the logarithmic function, which is used to calculate two variables (the vehicle internal measurement data set A and the external influence sequence of vehicle energy consumption B ) between .

[0140] S313: using an automatic encoder to fuse the features of each vehicle in the vehicle internal measurement data set with the new subsequence to obtain several vehicle energy consumption monitoring scenarios.

[0141] Among them, vehicle energy consumption monitoring scenarios include urban driving scenarios, highway driving scenarios, and mountain driving scenarios;

[0142] Urban driving scenarios refer to driving environments on urban roads, which are characterized by frequent stops and starts and low speed limits;

[0143] Highway driving scene refers to the driving environment on highways or expressways, which has the characteristics of high and constant speed of vehicles;

[0144] Mountain driving scenarios refer to driving environments in mountainous areas or on sloped roads, which are characterized by frequent changes in slopes, road widths, and road conditions.

[0145] It should be noted that in order to classify the vehicle internal measurement data set and the segmented vehicle energy consumption external influence subsequence according to preset conditions, and use feature fusion technology to construct several vehicle energy consumption monitoring scenarios, the following is further analyzed in conjunction with specific embodiments:

[0146] 1. Data collection and preprocessing

[0147] Vehicle interior measurement dataset:

[0148] Speed ​​data: The speed of the vehicle, which may be instantaneous speed or average speed.

[0149] Speed ​​data: Engine speed, reflecting the working status of the engine.

[0150] Fuel consumption data: including instantaneous fuel consumption, cumulative fuel consumption, etc.

[0151] Braking data: braking intensity, braking frequency, etc.

[0152] Idle data: idling time, fuel consumption at idle speed.

[0153] External influencing factors:

[0154] Temperature: Ambient outside temperature can affect engine efficiency and fuel consumption.

[0155] Slope: The slope of the road affects vehicle load and fuel consumption.

[0156] Traffic conditions: such as congestion level, average vehicle speed, etc.

[0157] Preprocessing steps:

[0158] Missing value processing: fill or delete missing values ​​that may exist in the measurement data.

[0159] Outlier filtering: Use statistical methods (such as triple standard deviation method, IQR method) to identify and process outliers in the data.

[0160] Time series synchronization: Align data from different sensors by timestamp to ensure that each data item can be analyzed synchronously.

[0161] 2. Data segmentation

[0162] The vehicle internal measurement data is segmented by time, mileage or specific driving conditions (such as driving mode). For example:

[0163] Time dimension: The vehicle speed, rotation speed, fuel consumption and other data can be divided into sub-sequences of multiple time periods on an hourly or daily basis.

[0164] Segment dimension: Split data based on geographic location or specific road segments, such as uphill segments, flat segments, and downhill segments.

[0165] Driving status: data segmentation based on the vehicle’s driving status (idling, acceleration, deceleration, constant speed driving, etc.).

[0166] External influence on the segmentation of subsequences:

[0167] According to different external environmental conditions, the influencing factors (temperature, slope, traffic conditions) are divided into multiple subsequences in time and space dimensions.

[0168] For example, corresponding external influence subsequences are generated in high temperature, low temperature, congestion, and smooth traffic scenarios.

[0169] 3. Condition classification

[0170] Classify the segmented data according to preset conditions. The conditions can be based on:

[0171] Environmental conditions: such as temperature range (high temperature, medium temperature, low temperature), road slope (uphill, downhill, flat road).

[0172] Driving conditions: such as highway scenarios, city driving scenarios, and mountain driving scenarios.

[0173] Traffic conditions: such as congested, semi-congested, and unobstructed.

[0174] Classification example:

[0175] Category 1: Speed ​​> 60 km / h and ambient temperature > 25°C → highway summer driving scenario.

[0176] Category 2: The vehicle brakes frequently and the slope is >10% → complex road scenes in mountainous areas.

[0177] Category 3: Idle time > 10 minutes and outside temperature < 5°C → cold start high idle fuel consumption scenario.

[0178] 4. Feature extraction and fusion

[0179] Feature extraction: Extract features related to vehicle energy consumption from internal data and external influencing factors.

[0180] Speed ​​characteristics: acceleration, deceleration, constant speed duration, maximum speed, etc.

[0181] Speed ​​characteristics: average speed, maximum speed, idle speed, etc.

[0182] Fuel consumption characteristics: instantaneous fuel consumption, fuel consumption per unit time, cumulative fuel consumption, etc.

[0183] Braking characteristics: braking intensity, number of brakes, and braking frequency.

[0184] Idle characteristics: idling time, idling fuel consumption, etc.

[0185] External environmental characteristics: ambient temperature, road slope, traffic flow, etc.

[0186] Feature fusion: Combine internal data with external influencing factors to perform feature fusion and form new energy consumption monitoring scene features. The following methods can be used:

[0187] Weighted fusion: Weights are assigned to different features according to their importance, and internal and external features are weighted and merged. For example, on a road with a large slope, the fuel consumption and acceleration features of the vehicle have greater weights.

[0188] Principal Component Analysis (PCA): Perform dimensionality reduction on internal and external features, project the original high-dimensional features into low-dimensional space, and extract the most important energy consumption features.

[0189] Time series fusion: Fusion of data with time series characteristics (such as speed, rotation speed) to form time-related energy consumption characteristics.

[0190] 5. Build energy consumption monitoring scenarios

[0191] Using the fused features, we build multiple vehicle energy consumption monitoring scenarios:

[0192] Urban driving scenarios: low-speed driving, frequent braking, and high idling times, combined with high external temperatures or congestion.

[0193] Highway driving scenario: constant speed, less acceleration, combined with flat road surface and good traffic conditions outside.

[0194] Mountain driving scenarios: sudden acceleration, sudden braking, frequent climbing and descending, and complex external environment (steep slope roads, low temperatures, etc.).

[0195] Each scenario will contain a different combination of internal measurement data and external influencing factors, forming a different energy consumption pattern.

[0196] 6. Application scenario verification

[0197] Scenario visualization: Visualize the constructed energy consumption monitoring scenario through charts or other means to show the changes in energy consumption characteristics under different scenarios.

[0198] Scenario verification: Verify the vehicle's energy consumption performance in different scenarios through historical data playback or simulation to ensure that the constructed scenario accurately reflects the actual energy consumption.

[0199] 7. Anomaly detection and rule building

[0200] Based on the constructed energy consumption monitoring scenarios, energy consumption anomaly detection rules are generated. According to the energy consumption characteristics of each scenario, anomaly standards are defined. For example:

[0201] In highway driving scenarios, high fuel consumption (above a certain threshold) may indicate an anomaly.

[0202] In urban congestion scenarios, long idling time may also be abnormal.

[0203] Through these rules, abnormal energy consumption in different scenarios can be monitored.

[0204] 8. Dynamic scene adjustment

[0205] Dynamically adjust the division rules of monitoring scenes according to the data in actual operation. As external environmental conditions change (such as climate and traffic conditions), the scene classification and feature weights can be dynamically adjusted to ensure the accuracy and timeliness of monitoring.

[0206] S32, analyzing the vehicle energy consumption characteristics according to the energy consumption influencing factors of each external environment in each vehicle energy consumption monitoring scenario, and generating energy consumption abnormality monitoring rules corresponding to the vehicle energy consumption monitoring scenario;

[0207] It should be noted that the energy consumption abnormality monitoring rules include:

[0208] Baseline energy consumption setting: Set the normal energy consumption range based on the vehicle's historical energy consumption data, and set different energy consumption baselines considering different types of vehicles (such as electric vehicles, fuel vehicles) and different driving modes (city, highway).

[0209] Real-time energy consumption monitoring: Energy consumption data, such as battery consumption rate and fuel consumption, are monitored in real time through the vehicle's sensors. Real-time data is compared with the energy consumption baseline to detect any consumption outside the normal range.

[0210] Abnormal trigger conditions: If the energy consumption is continuously monitored to exceed the preset threshold (such as exceeding the baseline by 10%) for a certain period of time (such as 30 minutes), it is judged as abnormal energy consumption, and the time, location, and duration of the abnormal energy consumption are recorded for subsequent analysis.

[0211] S33, constructing an energy consumption structure tree by using vehicle energy consumption monitoring scenarios as tree nodes, and integrating energy consumption anomaly monitoring rules into the energy consumption structure tree to obtain a multi-scenario rule linkage mechanism of the energy consumption structure tree.

[0212] Among them, the energy consumption structure tree is constructed by taking the vehicle energy consumption monitoring scenario as a tree node, and the energy consumption abnormality monitoring rules are integrated into the energy consumption structure tree to obtain the multi-scenario rule linkage mechanism of the energy consumption structure tree, including:

[0213] S331, each vehicle energy consumption monitoring scenario represents a branch of a structure tree, extracts vehicle features and new subsequences as branch points and generates an energy consumption structure tree.

[0214] It should be noted that the hierarchical structure of the energy consumption model tree includes:

[0215] The energy consumption model tree is a hierarchical analysis tool that decomposes the energy consumption monitoring scenarios hierarchically, from the global to the specific, to find out the specific causes of energy consumption in different scenarios.

[0216] Level 1 node: Set the top level of the energy consumption model tree to different scenario categories (such as urban driving, highway driving, etc.).

[0217] Secondary node: In each primary scenario, the data is further decomposed according to the specific data items of the vehicle's internal sensors (such as speed, engine speed, fuel consumption, etc.).

[0218] Level 3 nodes: Based on the level 2 nodes, they can be further decomposed into sub-factors. For example, speed data can be subdivided into acceleration, deceleration, constant speed driving and other states; fuel data can be subdivided into instantaneous fuel consumption, total fuel consumption, etc.

[0219] The steps of building the energy consumption model tree include:

[0220] Decomposition starts from the scenario nodes: Establish top-level nodes based on the generated energy consumption monitoring scenarios (city, highway, mountainous area, etc.).

[0221] Decompose energy consumption factors layer by layer: For each scenario, find different energy consumption factors (speed, rotation speed, fuel consumption, etc.) and further refine them.

[0222] Associating causal relationships: The energy consumption model tree is not only used to decompose data, but also to associate data at different levels through causal relationships. For example, speed changes may cause instantaneous fuel consumption fluctuations, thereby triggering abnormal energy consumption points.

[0223] S332, analyzing the energy consumption characteristic correlation indexes between branches, establishing energy consumption anomaly monitoring rules to express the intimacy between branches, and establishing an intimacy matrix based on the intimacy between branches to realize the trunk expression of the energy consumption structure tree;

[0224] S333. Taking the intimacy between branches as input, a path planning algorithm is introduced to obtain the optimal solution of the local shortest path between the trunk and each branch, and a hierarchical matrix of the trunk and branches is constructed to realize the hierarchical structure of the energy consumption structure.

[0225] It should be noted that the intimacy between branches is used as input, and the path planning algorithm is introduced to obtain the optimal solution of the local shortest path between the trunk and each branch, and a hierarchical matrix of the trunk and branches is constructed to realize the hierarchical structure of the energy consumption structure, including:

[0226] The shortest path from the trunk to each branch is calculated using Dijkstra's algorithm (an algorithm used to find the shortest path from a single starting point to all other vertices in a graph), considering closeness as the path cost.

[0227] Define a hierarchical matrix in which each row represents a branch, each column represents the connection strength and path length of the branch with the trunk and other branches, and fill the hierarchical matrix with the shortest paths from the trunk to each branch generated by the path planning algorithm. Each element reflects the cost or length of the shortest path to the branch.

[0228] Analyze the calculated paths and hierarchical structures to determine whether they effectively reflect the intimacy and structural priorities, and adjust the path planning parameters or intimacy definitions based on the analysis results to optimize the hierarchical structure.

[0229] S334. Integrate the energy consumption anomaly monitoring rules into the energy consumption structure tree to obtain a multi-scenario rule linkage mechanism that conforms to the hierarchy of the energy consumption structure tree.

[0230] It should be noted that the multi-scenario rule linkage mechanism includes:

[0231] Linkage strategy: When abnormal energy consumption is detected in a certain scenario, it automatically triggers the inspection and adjustment of other related scenarios or levels. For example, if abnormal energy consumption is found in a high-speed driving scenario, the system can check the power system settings or aerodynamic optimization measures related to high-speed driving.

[0232] Response actions: Depending on the type and severity of the anomaly detected, the system can implement a range of response measures, such as adjusting driving modes, notifying the driver for inspection or maintenance, or even automatically adjusting the vehicle's performance settings.

[0233] S34. Based on the multi-scenario rule linkage mechanism and through the energy consumption structure tree branching, find the energy consumption abnormal scenes in the vehicle energy consumption monitoring scenes, and establish a real-time energy consumption monitoring model to identify abnormal points in the energy consumption abnormal scenes.

[0234] S35. Use adaptive technology to adjust the energy consumption anomaly monitoring rules, and continuously update the multi-scenario rule linkage mechanism of the energy consumption structure tree based on the adjusted energy consumption anomaly monitoring rules.

[0235] like Figure 2 As shown, according to another embodiment of the present invention, a vehicle energy consumption monitoring system based on multi-source data is also provided, and the vehicle energy consumption monitoring system based on multi-source data includes a vehicle internal measurement module 1, a vehicle external environment analysis module 2 and a vehicle energy consumption monitoring module 3;

[0236] The vehicle interior measurement module 1 is used to obtain the initial vehicle measurement data sent by the sensor module installed on the vehicle, and optimize the initial vehicle measurement data using the time window filtering technology to obtain the vehicle interior measurement data set;

[0237] The vehicle external environment analysis module 2 is used to identify the energy consumption influencing factors of the vehicle external environment and generate a vehicle energy consumption external influence sequence, and segment the vehicle energy consumption external influence sequence using an equal-dimensional recursive grey algorithm;

[0238] The vehicle energy consumption monitoring module 3 is used to generate a number of vehicle energy consumption monitoring scenarios based on the vehicle internal measurement data set and the vehicle energy consumption external influence sequence, and to construct an energy consumption structure tree branch to find energy consumption anomalies in each energy consumption monitoring scenario.

[0239] In summary, with the help of the above technical solutions of the present invention, the present invention can comprehensively monitor the key performance parameters of the vehicle through speed sensors, engine speed sensors and fuel level sensors, provide necessary input data for energy consumption analysis, and accurately locate the relative positions of each sensor through the interaction of acquisition nodes and beacon nodes, combined with the calculation of transmission hops and distributed distances, which plays an important role in optimizing data transmission and fault detection in the internal network of the vehicle. At the same time, by establishing a clock model and time series analysis technology, time deviation and noise are effectively processed to ensure the consistency and accuracy of data obtained from different sensors in time, and provide data support for real-time vehicle energy consumption monitoring and optimization. The present invention combines the internal measurement data of the vehicle and the sequence of external influencing factors, and uses neighborhood mutual information and automatic encoder technology for feature fusion to construct an energy consumption monitoring scene that accurately matches the actual driving conditions, and organizes different energy consumption monitoring scenes into a structure tree form, which enhances the organization of the system and the logic of data processing, and realizes the effective link and interaction between energy consumption monitoring scenes, so that energy consumption anomalies can be quickly identified and responded to between related scenes. The present invention generates energy consumption monitoring scenarios through internal and external data, can simulate and analyze energy consumption performance under specific driving conditions, provide more accurate energy consumption management, and construct an energy consumption structure tree allowing the system to process and analyze data in an organized manner, thereby improving the efficiency of data retrieval and processing. The branch search in the energy consumption structure tree enables rapid identification and response to abnormal energy consumption scenarios, supports real-time energy consumption monitoring and exception handling, ensures that the monitoring strategy keeps pace with the actual situation, and thereby enhances the adaptability and flexibility of the system.

[0240] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A vehicle energy consumption monitoring method based on multi-source data, It is characterized in that The vehicle energy consumption monitoring method based on multi-source data includes: S1. Obtaining initial vehicle measurement data sent by a sensor module installed on a vehicle, and optimizing the initial vehicle measurement data using a time window filtering technique to obtain a vehicle internal measurement data set; S2, identifying the energy consumption influencing factors of the vehicle's external environment and generating a vehicle energy consumption external influencing sequence, and segmenting the vehicle energy consumption external influencing sequence using an equal-dimensional recursive grey algorithm; S3, generating several vehicle energy consumption monitoring scenarios based on the vehicle internal measurement data set and the vehicle energy consumption external influence sequence, and constructing an energy consumption structure tree branch to find energy consumption anomalies in each energy consumption monitoring scenario; The generating of a plurality of vehicle energy consumption monitoring scenarios based on the vehicle internal measurement data set and the vehicle energy consumption external influence sequence, and constructing an energy consumption structure tree branch to search for energy consumption abnormal points in each energy consumption monitoring scenario includes: S31, classifying the vehicle internal measurement data set and the vehicle energy consumption external influence sequence according to preset conditions, and constructing several vehicle energy consumption monitoring scenarios using feature fusion technology; S32, analyzing the vehicle energy consumption characteristics according to the energy consumption influencing factors of each external environment in each vehicle energy consumption monitoring scenario, and generating energy consumption abnormality monitoring rules corresponding to the vehicle energy consumption monitoring scenario; S33, constructing an energy consumption structure tree by using the vehicle energy consumption monitoring scenario as a tree node, and integrating the energy consumption abnormality monitoring rules into the energy consumption structure tree to obtain a multi-scenario rule linkage mechanism of the energy consumption structure tree; S34, based on the multi-scenario rule linkage mechanism and through the energy consumption structure tree branching, searching for abnormal energy consumption scenarios in the vehicle energy consumption monitoring scenarios, and establishing a real-time energy consumption monitoring model to identify abnormal points in the abnormal energy consumption scenarios; S35, using adaptive technology to adjust the energy consumption anomaly monitoring rules, and continuously updating the multi-scenario rule linkage mechanism of the energy consumption structure tree based on the adjusted energy consumption anomaly monitoring rules; The method of classifying the vehicle internal measurement data set and the vehicle energy consumption external influence sequence according to preset conditions and constructing several vehicle energy consumption monitoring scenarios using feature fusion technology includes: S311, introducing classification intervals from the vehicle internal measurement data set and the vehicle energy consumption external influence sequence respectively, and setting the classification intervals as the sample neighborhood space; S312, calculating the neighborhood mutual information between the vehicle internal measurement data set and the vehicle energy consumption external influence sequence, and sorting the segmented vehicle energy consumption external influence subsequences according to the size of the neighborhood mutual information, and fusing the vehicle energy consumption external influence subsequences based on the classification interval to generate several new subsequences; S313, using an automatic encoder to fuse the features of each vehicle in the vehicle internal measurement data set with the new subsequence to obtain several vehicle energy consumption monitoring scenarios; The energy consumption structure tree is constructed by taking the vehicle energy consumption monitoring scenario as a tree node, and the energy consumption abnormality monitoring rules are integrated into the energy consumption structure tree to obtain a multi-scenario rule linkage mechanism of the energy consumption structure tree, including: S331, representing a branch of a structure tree for each vehicle energy consumption monitoring scenario, extracting vehicle features and new subsequences as branch points and generating an energy consumption structure tree; S332, analyzing the energy consumption characteristic correlation indexes between branches, establishing energy consumption anomaly monitoring rules to express the intimacy between branches, and establishing an intimacy matrix based on the intimacy between branches to realize the trunk expression of the energy consumption structure tree; S333, taking the intimacy between branches as input, introducing the path planning algorithm to obtain the optimal solution of the local shortest path between the trunk and each branch, and constructing a hierarchical matrix of the trunk and branches to realize the hierarchical structure of the energy consumption structure; S334. Integrate the energy consumption anomaly monitoring rules into the energy consumption structure tree to obtain a multi-scenario rule linkage mechanism that conforms to the hierarchy of the energy consumption structure tree.

2. A vehicle energy consumption monitoring method based on multi-source data according to claim 1, It is characterized in that The method of obtaining the initial vehicle measurement data sent by the sensor module installed on the vehicle and optimizing the initial vehicle measurement data using the time window filtering technology to obtain the vehicle internal measurement data set includes: S11, configuring a sensor module, wherein the sensor module includes a speed sensor, an engine speed sensor, and a fuel level sensor; S12, respectively obtaining the transmission hop counts of the speed sensor, the engine speed sensor, and the fuel level sensor, calculating the distributed distances between the sensors based on the transmission hop counts, and locating the sensors; S13, using a speed sensor, an engine speed sensor and a fuel level sensor to collect vehicle speed characteristics, engine speed characteristics and fuel consumption characteristics; S14, mapping the vehicle speed characteristics, engine speed characteristics and fuel consumption characteristics onto a unified time axis, and integrating the characteristics of each vehicle to obtain a vehicle internal measurement data set.

3. A vehicle energy consumption monitoring method based on multi-source data according to claim 2, It is characterized in that The steps of respectively obtaining the transmission hop counts of the speed sensor, the engine speed sensor, and the fuel level sensor, calculating the distributed distances between the sensors based on the transmission hop counts, and locating the distances between the sensors include: S121, using each sensor as a collection node in the vehicle internal network, and each collection node regularly transmits initial vehicle measurement data and the number of transmission hops according to a preset period; S122, taking the data receiving end as a beacon node, and sending the transmission hop count of the beacon node to each collection node, and establishing a location information table after the collection node receives the transmission hop count of the beacon node; S123, calculating the distributed distance and average distance per hop between the collection node and the beacon node according to the number of transmission hops recorded in the location information table and updating the location information table; S124. Each collection node obtains the number of transmission hops of other collection nodes according to the updated location information table, uses the positioning accuracy of the collection node as a weighting factor, and uses the three-sided positioning method to perform distributed positioning of each sensor.

4. A vehicle energy consumption monitoring method based on multi-source data according to claim 3, It is characterized in that The mapping of the vehicle speed characteristics, the engine speed characteristics and the fuel consumption characteristics onto a unified time axis includes: S141, determining an asymmetry ratio between a collection node and a beacon node, and establishing clock models of vehicle speed characteristics, engine speed characteristics, and fuel consumption characteristics based on the determination result of the asymmetry ratio; S142, using the Laida criterion to identify the jump abnormal values ​​of each clock model, and combining the autoregressive integrated moving average model to eliminate the time noise jitter generated by each clock model to obtain a stable time series; S143. Predictively compensate each clock model based on a stationary time series, and use the compensated clock model to map vehicle speed characteristics, engine speed characteristics, and fuel consumption characteristics onto a unified time axis.

5. A vehicle energy consumption monitoring method based on multi-source data according to claim 4, It is characterized in that The vehicle energy consumption monitoring scenarios include urban driving scenarios, highway driving scenarios and mountain driving scenarios; Among them, the urban driving scenario refers to the driving environment on urban roads, which has the characteristics of frequent stopping and starting and low speed limit; Highway driving scene refers to the driving environment on highways or expressways, which has the characteristics of high and constant speed of vehicles; Mountain driving scenarios refer to driving environments in mountainous areas or on sloped roads, which are characterized by frequent changes in slopes, road widths, and road conditions.

6. A vehicle energy consumption monitoring method based on multi-source data according to claim 5, It is characterized in that The calculation formula of the neighborhood mutual information is: ; In the formula, A represents a vehicle interior measurement dataset; B represents the sequence of external influences on vehicle energy consumption; NMI μ ( A , B ) represents the neighborhood mutual information between the vehicle internal measurement dataset and the external influence sequence of vehicle energy consumption; μ A ( x i ) indicates vehicle characteristics x i Neighborhood on a dataset of vehicle interior measurements; μ B ( x i ) indicates vehicle characteristics x i Neighborhood on the sequence of external influences on vehicle energy consumption; Indicates vehicle characteristics x i In the sample neighborhood space The neighborhood formed on n Indicates the number of vehicle features; log represents the logarithmic function.

7. A vehicle energy consumption monitoring system based on multi-source data, used to implement the vehicle energy consumption monitoring method based on multi-source data according to any one of claims 1 to 6, It is characterized in that The vehicle energy consumption monitoring system based on multi-source data includes a vehicle internal measurement module, a vehicle external environment analysis module and a vehicle energy consumption monitoring module; The vehicle interior measurement module is used to obtain initial vehicle measurement data sent by a sensor module installed on the vehicle, and optimize the initial vehicle measurement data using a time window filtering technology to obtain a vehicle interior measurement data set; The vehicle external environment analysis module is used to identify the energy consumption influencing factors of the vehicle external environment and generate a vehicle energy consumption external influence sequence, and segment the vehicle energy consumption external influence sequence using an equal-dimensional recursive grey algorithm; The vehicle energy consumption monitoring module is used to generate a number of vehicle energy consumption monitoring scenarios based on the vehicle internal measurement data set and the vehicle energy consumption external influence sequence, and to construct an energy consumption structure tree branch to find energy consumption anomalies in each energy consumption monitoring scenario.

Citation Information

Cited By

  • Abnormal monitoring and early warning method for motion control state of road train

    CN122065207A