CN115759347A - Method for quickly predicting travel energy consumption of electric bus based on characteristic data - Google Patents
Method for quickly predicting travel energy consumption of electric bus based on characteristic data Download PDFInfo
- Publication number
- CN115759347A CN115759347A CN202211299932.1A CN202211299932A CN115759347A CN 115759347 A CN115759347 A CN 115759347A CN 202211299932 A CN202211299932 A CN 202211299932A CN 115759347 A CN115759347 A CN 115759347A
- Authority
- CN
- China
- Prior art keywords
- data
- energy consumption
- vehicle
- time
- characteristic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
A method for quickly predicting travel energy consumption of an electric bus based on characteristic data belongs to the field of traffic. The invention provides a vehicle energy consumption prediction value for the charging plan of the electric bus fleet, and predicts the vehicle travel energy consumption based on historical data. Considering the relation between the vehicle energy consumption and the characteristic data, the invention provides a method for quickly calculating the energy consumption of the electric bus based on the characteristic data. The method comprises the steps of firstly collecting historical electric quantity data and travel data of the electric bus, extracting stable key features, searching historical data closest to the features of a new travel, and taking the mean value of the historical energy consumption data as a predicted value of the energy consumption of the new travel. The method fully excavates the energy consumption characteristics of the vehicle travel in the existing electric bus operation data; a new method for rapidly and accurately predicting the energy consumption of the electric bus is provided.
Description
Technical Field
The invention relates to the field of carrying tool application engineering and public transportation, in particular to a historical data-based method for quickly calculating energy consumption of an electric bus.
Background
In recent years, in order to reduce carbon emission and realize 'carbon peak reaching and carbon neutralization', the nation increases the strength of popularizing electric buses. The electric bus as a public transportation mode depending on green energy has the advantages of small pollutant discharge amount, energy conservation, emission reduction, low operation cost and the like. Under the combined promotion of government policy support, battery technology development and social approval, the electric buses are rapidly developed. However, the problem of uncertainty of the power consumption of the electric bus causes the disordered large-scale charging phenomenon of the electric bus and certain pressure on a power distribution network, and further development of the electric bus is hindered. Therefore, it is necessary to research how to quickly and accurately predict the energy consumption of the electric bus.
Because the research on the electric bus energy consumption prediction model is less, the invention mainly refers to an electric vehicle energy consumption prediction method. At present, the energy consumption prediction method of the electric automobile mainly comprises the following steps: and predicting energy consumption based on historical data and constructing a model by utilizing the influence factors to predict the energy consumption. Influence factors considered by the model are comprehensive, influence among the factors is complex, and the related data types are large, so that the related data of the influence factors are difficult to obtain, and the model is difficult to serve in practice. The method can truly reflect the energy consumption change of the vehicle. However, when the external environment changes, the energy consumption prediction precision will also change. Therefore, the method needs to select a more stable characteristic value from the historical data for energy consumption prediction.
Based on the method, the stable vehicle travel energy consumption characteristic parameters are extracted based on the operation historical data of the electric bus, and then the electric bus energy consumption prediction model based on similar characteristic search is constructed, so that the model can quickly and accurately predict the electric bus energy consumption. The difficult points and innovation points of the patent are as follows: fully mining the characteristics of vehicle travel energy consumption in the existing electric bus operation data; a new method for rapidly and accurately predicting the energy consumption of the electric bus is provided.
Disclosure of Invention
The invention provides a vehicle energy consumption prediction value for the charging plan of the electric bus fleet, and predicts the vehicle travel energy consumption based on historical data. Considering the relation between the vehicle energy consumption and the characteristic data, the invention provides a method for quickly calculating the energy consumption of the electric bus based on the characteristic data. The method comprises the steps of firstly collecting historical electric quantity data and historical travel data of the electric bus, extracting stable key features, searching historical data closest to the features of a new travel, and taking the average value of the historical energy consumption data as a new travel energy consumption predicted value.
A method for quickly predicting travel energy consumption of an electric bus based on characteristic data comprises the following steps:
the method comprises the following steps: and fusing the real-time electric quantity data of the electric bus and the road order data of the electric bus according to the timestamp and the vehicle number.
The method for acquiring the real-time electric quantity data of the electric bus comprises the following steps: "vehicle number", "electric quantity", and "time stamp" corresponding to real-time electric quantity; the electric bus road list data comprises: "date", "route name", "vehicle number", "driver name", "traveling direction", "route type", "actual departure time", and "actual arrival time".
The method for fusing the real-time electric quantity data of the electric bus and the road list data of the electric bus comprises the following steps: firstly, converting a time stamp in the real-time electric quantity data of the electric bus into a form of year/month/day, minute/second; then, according to data of 'date', 'actual departure time', 'actual arrival time' and 'vehicle number' in the electric bus route order data, 'electric quantity' in the real-time electric quantity data of the electric bus is matched into the electric bus route order data, and the data of 'departure time electric quantity' and 'arrival time electric quantity' of the vehicle are added. And finally, performing initial data processing, including deleting the data with null values and calculating basic data. Wherein calculating the base data comprises: calculating the running time of the vehicle running the journey, and calculating the 'vehicle running time' of the vehicle according to the existing data, namely subtracting the actual departure time from the actual arrival time of the vehicle; and calculating the 'consumed electric quantity' of the vehicle in the travel, namely subtracting the electric quantity of the arrival time from the electric quantity of the departure time of the vehicle, and finally forming training data.
The training data parameters include: "date", "route name", "vehicle number", "driver name", "traveling direction", "route type", "actual departure time", "departure time electric quantity", "actual arrival time", "arrival time electric quantity", "vehicle operation time", and "consumed electric quantity".
Step two: and extracting characteristic parameters of the departure time.
The departure schedule feature is primarily represented by the vehicle operating time. The vehicle running time is related to the departure time of the vehicle, the road traffic state and other factors, and therefore can be expressed as a departure schedule characteristic. And predicting the running time of the vehicle according to the actual departure time of the electric bus in the training data.
The prediction method comprises the following steps: firstly, constructing a starting time characteristic data set, wherein the data set comprises: the system comprises a departure time and a running time, wherein the departure time is a time period formed by one minute interval from the earliest departure time to the latest departure time of a vehicle in training data, and the running time is data to be acquired. Secondly, arranging the actual departure time in the training data according to an ascending order, grouping the actual departure time according to the same departure time by taking 'minutes' as a minimum time unit, averaging the running time corresponding to the same departure time, taking the running time as the running time of the departure vehicle at the moment, and splicing the obtained running time into a departure time characteristic data set according to the 'departure time'. Finally, some blank data exist in the departure time characteristic data set, so that the running time of two adjacent departure times is filled by adopting an equal difference method, and the running time can be smoothed to form a complete departure time characteristic data set. Finally, each departure time is corresponding to a departure time characteristic.
And meanwhile, inserting the characteristic parameters of the departure time into corresponding training data.
Step three: and extracting vehicle performance characteristic parameters.
The data characterizing the vehicle performance characteristics in the experimental data is "vehicle number". However, the vehicle number cannot intuitively distinguish the difference of the energy consumption of different vehicles, so that the vehicle characteristics of the electric bus are replaced by the average value and the standard deviation of the energy consumption of the electric bus, namely the vehicle characteristics are digitalized, and the vehicle characteristics are determined from the energy consumption perspective. The average value of the vehicle energy consumption represents the characteristics of the electric quantity consumed by different vehicles, and the standard deviation of the vehicle energy consumption represents the stable condition of the electric quantity consumed by different vehicles.
The specific calculation method is as follows: firstly, grouping data with the same vehicle number in training data, solving the average value and standard deviation of 'energy consumption' in each group of data, and taking the value as a characteristic value representing the energy consumption of the vehicle; finally, a vehicle performance characteristic data set is formed, namely, each vehicle number corresponds to a vehicle energy consumption average value and a vehicle energy consumption standard deviation.
And simultaneously, inserting the vehicle performance characteristic parameters into the corresponding training data.
Step four: and extracting the driver attribute parameters.
The data characterizing the driver attribute characteristics in the experimental data is "driver name". The name of the driver cannot intuitively represent the characteristics of the energy consumption of the driver for driving the vehicle, so that the average value and the standard deviation of the energy consumption of the driver are used for replacing the operating characteristics of the driver, namely the characteristics of the driver are digitalized, and the characteristics of the driver are determined from the energy consumption perspective. The average value of the energy consumption of the driver represents the characteristics of the electric quantity consumed by different drivers, and the standard deviation of the energy consumption of the driver represents the stable condition of the electric quantity consumed by different drivers.
The specific calculation method is as follows: firstly, grouping data of the same driver name in training data, solving the average value and standard deviation of 'energy consumption' in each group of data, and taking the value as a characteristic value representing the energy consumption of the driver; finally, a driver attribute feature data set is formed, namely, each driver name corresponds to one driver energy consumption average value and one driver energy consumption standard deviation.
And simultaneously, inserting the driver attribute characteristic parameters into the corresponding training data.
Step five: the raw data of the new trip is processed.
Firstly, according to the nature of the new journey of the electric bus, selecting the training data of the same route, the same driving direction and the same working day, and calculating the training data according to the second step, the third step and the fourth step to obtain a departure time characteristic data set, a vehicle performance characteristic data set and a driver attribute characteristic data set; and then, respectively searching in corresponding data sets according to departure time, vehicle number and driver name in the data to be predicted to obtain a departure time characteristic parameter, a vehicle performance characteristic parameter and a driver attribute characteristic parameter of a new journey, and inserting the characteristic value into the original data to finish initial processing of the predicted data.
Step six: and calculating the Euclidean distance between the prediction data and the training data.
Firstly, the characteristic value data of training data and prediction data are standardized, and the data standardization principle is as follows:
x 'in the formula' i For normalized data values, x i Representing the raw data value, sigma representing the raw data standard deviation, n representing the number of data samples in the data set, x mean Representing the average of data objects x in the raw data.
Then, the principle of the euclidean distance is used:
wherein x is prediction data, C i For the ith training data, x j Representing the j-th characteristic value, C, in the prediction data ij Represents the jth eigenvalue in the ith training data, and m represents the number of eigenvalues. The above equation can calculate the distance between the predicted data eigenvalue and all the training data eigenvalues one by one.
Step seven: the predicted energy consumption for the new trip is given.
Before predicting the energy consumption of the new journey of the electric bus, the optimal data number of the shortest distance between the predicted data and the training data needs to be determined. The specific calculation method is as follows: firstly, arranging the Euclidean distances calculated in the step six in an ascending order, and respectively taking the first k data with the minimum distance; secondly, calculating the average value of the energy consumption values in the k data as the energy consumption of the prediction data, and further calculating the average error of the prediction data; finally, 12 groups of data in different directions of different lines are trained respectively to obtain average errors corresponding to different k values. And selecting the k value with the minimum average error as the optimal k value, and finally selecting k =7.
And selecting the first 7 training data with the shortest distance, and taking the average value of the energy consumption as the predicted energy consumption of the new journey, namely completing the prediction of the journey energy consumption of the electric bus.
Compared with the prior art, the invention has the beneficial effects that:
according to the method, the vehicle travel energy consumption characteristics are fully mined according to historical electric quantity data and travel data of the electric bus, and the electric bus energy consumption prediction model is constructed according to the characteristic parameters, so that the electric bus energy consumption is rapidly and accurately predicted, the influence factors of collecting various data such as road information data and weather meteorological data and extracting the electric bus energy consumption in the traditional method are effectively avoided, and the application of the energy consumption prediction model in reality is facilitated.
Description of the drawings:
fig. 1 is a processing process diagram of fusing real-time electric quantity data and electric bus route order data of an electric bus in the method for rapidly predicting the journey energy consumption of the electric bus based on characteristic data according to the invention;
FIG. 2 is a graph showing the trend and accuracy of the operation time of the electric bus in the method for rapidly predicting the energy consumption of the electric bus journey based on the characteristic data according to the present invention;
FIG. 3 is a flow chart of obtaining characteristic parameters of vehicle performance in the method for rapidly predicting the energy consumption of the electric bus journey based on the characteristic data according to the invention;
FIG. 4 is a flow chart of obtaining driver attribute characteristic parameters in the method for rapidly predicting the energy consumption of the electric bus journey based on the characteristic data according to the invention;
FIG. 5 is a flowchart of the method for rapidly predicting the energy consumption of the journey of the electric bus based on the characteristic data for obtaining the optimal data number of the shortest distance between the predicted data and the training data;
FIG. 6 is a frame diagram of the method for rapidly predicting the energy consumption of the electric bus journey based on the characteristic data.
Detailed Description
The invention is described in detail below with reference to the attached drawing figures:
a method for rapidly predicting travel energy consumption of an electric bus based on characteristic data refers to the accompanying figure 5, and comprises the following steps:
the method comprises the following steps: and fusing the real-time electric quantity data of the electric bus and the road order data of the electric bus according to the timestamp and the vehicle number.
The real-time electric quantity data of the electric bus comprises: "vehicle number", "quantity of electricity", and "timestamp"; the electric bus road list data comprises: "date", "route name", "vehicle number", "driver name", "direction of travel", "route type", "actual departure time", and "actual arrival time".
Referring to the attached figure 1, the real-time electric quantity data of the electric bus and the road order data of the electric bus are fused.
Firstly, converting a time stamp in the real-time electric quantity data of the electric bus into a form of year/month/day, minute/second; then, according to data of 'date', 'actual departure time', 'actual arrival time' and 'vehicle number' in the electric bus route order data, 'electric quantity' in the electric bus real-time electric quantity data is matched into the electric bus route order data, and data of 'departure time electric quantity' and 'arrival time electric quantity' are added. And finally, performing initial data processing, including deleting the data with null values and calculating basic data. Wherein calculating the base data comprises: calculating the running time of the vehicle running the journey, and calculating the 'vehicle running time' of the vehicle according to the existing data, namely subtracting the actual departure time from the actual arrival time of the vehicle; and calculating the 'consumed electric quantity' of the vehicle in the travel, namely subtracting the electric quantity of the arrival time from the electric quantity of the departure time of the vehicle to finally form training data.
Step two: and extracting characteristic parameters of the departure time.
Referring to fig. 2, the operation time trend graph has a starting time on the abscissa and a vehicle operation time on the ordinate, and it can be known from the graph that there is a double-hump relationship between the starting time characteristic of the electric bus and the vehicle operation time. Therefore, the departure schedule feature is expressed in terms of the vehicle operating time. And predicting the vehicle running time according to the actual departure time of the electric bus in the training data.
The prediction method comprises the following steps: firstly, constructing a starting time characteristic data set, wherein the data set comprises: the system comprises a 'departure time' and a 'running time', wherein the departure time 'is a time period formed by the earliest departure time to the latest departure time of a vehicle in training data and separated by one minute, and the running time' is data to be obtained. Secondly, arranging the actual departure time in the training data according to an ascending order, grouping the actual departure time according to the same departure time by taking 'minutes' as a minimum time unit, averaging the running time corresponding to the same departure time, taking the running time as the running time of the departure vehicle at the moment, and splicing the obtained running time into a departure time characteristic data set according to the 'departure time'. Finally, some blank data exist in the departure time characteristic data set, so that the running time of two adjacent departure times is filled by adopting an equal difference method, and the running time can be smoothed to form a complete departure time characteristic data set. Finally, each departure time is formed to have a departure time characteristic corresponding to the departure time, and the prediction accuracy is shown in the attached figure 2.
And simultaneously, inserting the characteristic parameters of the departure time into corresponding training data.
Step three: and extracting vehicle performance characteristic parameters.
The data characterizing the vehicle performance characteristics in the experimental data is the "vehicle number". However, the vehicle number cannot intuitively distinguish the difference of the energy consumption of different vehicles, so that the vehicle characteristics of the electric buses are replaced by the average value and the standard deviation of the energy consumption of the electric buses, namely the vehicle characteristics are digitalized, and the vehicle characteristics are determined from the energy consumption perspective. The average value of the vehicle energy consumption represents the characteristics of the electric quantity consumed by different vehicles, and the standard deviation of the vehicle energy consumption represents the stable condition of the electric quantity consumed by different vehicles.
Referring to fig. 3, a vehicle performance characteristic data set is calculated.
The specific calculation method is as follows: firstly, grouping data with the same vehicle number in training data, solving the average value and standard deviation of 'energy consumption' in each group of data, and taking the value as a characteristic value representing the energy consumption of the vehicle; finally, a vehicle performance characteristic data set is formed, namely, each vehicle number corresponds to a vehicle energy consumption average value and a vehicle energy consumption standard deviation. And simultaneously, inserting the vehicle performance characteristic parameters into the corresponding training data.
Step four: and extracting the driver attribute parameters.
The data characterizing the driver attribute characteristics in the experimental data is "driver name". The driver name cannot intuitively represent the characteristics of the energy consumption of the driver for driving the vehicle, so that the driver operation characteristics are replaced by the average value and the standard deviation of the driver energy consumption, namely the driver characteristics are digitalized, and the driver characteristics are determined from the energy consumption perspective. The average value of the driver energy consumption represents the characteristics of the electric quantity consumed by different drivers, and the standard deviation of the driver energy consumption represents the stable condition of the electric quantity consumed by different drivers.
The specific calculation method is as follows: firstly, grouping data of the same driver name in training data, solving the average value and standard deviation of 'energy consumption' in each group of data, and taking the value as a characteristic value representing the energy consumption of the driver; finally, a driver attribute feature data set is formed, namely, each driver name corresponds to one driver energy consumption average value and one driver energy consumption standard deviation.
And simultaneously, inserting the driver attribute characteristic parameters into the corresponding training data.
Step five: the raw data of the new journey is processed.
Firstly, according to the nature of the new journey of the electric bus, selecting the training data of the same route, driving direction and working day, and calculating the training data according to the second step, the third step and the fourth step to obtain a departure time characteristic data set, a vehicle performance characteristic data set and a driver attribute characteristic data set; and then, respectively searching in corresponding data sets according to departure time, vehicle number and driver name in the data to be predicted to obtain a departure time characteristic parameter, a vehicle performance characteristic parameter and a driver attribute characteristic parameter of a new journey, and inserting the characteristic value into the original data to finish initial processing of the predicted data.
Step six: and calculating the Euclidean distance between the prediction data and the training data.
Firstly, the characteristic value data of training data and prediction data are standardized, and the data standardization principle is as follows:
in the formula, x i ' is a normalized data value, x i Representing the raw data value, σ represents the raw data standard deviation, n represents the number of data samples in the data set, x mean Representing the average of data objects x in the raw data.
Then, the principle of the euclidean distance is used:
wherein x is prediction data, C i For the ith training data, x j Representing the j-th characteristic value, C, in the prediction data ij Represents the jth eigenvalue in the ith training data, and m represents the number of eigenvalues. The above equation can calculate the distance between the predicted data eigenvalue and all the training data eigenvalues one by one.
Step seven: the predicted energy consumption for the new trip is given.
Referring to fig. 4, the optimal data number of the shortest distance between the prediction data and the training data is obtained.
The specific calculation method is as follows: firstly, arranging the Euclidean distances calculated in the step six in an ascending order, and respectively taking the first k data with the minimum distance; secondly, calculating the average value of the energy consumption values in the k data as the energy consumption of the prediction data, and further calculating the average error of the prediction data; finally, 12 groups of data in different directions of different lines are trained respectively to obtain average errors corresponding to different k values. And selecting the k value with the minimum average error as the optimal k value, and finally selecting k =7.
And selecting the training data with the shortest distance from the first 7 training data, and taking the average value of the energy consumption as the predicted energy consumption of the new journey, namely completing the prediction of the journey energy consumption of the electric bus.
Claims (1)
1. A method for quickly predicting travel energy consumption of an electric bus based on characteristic data is characterized by comprising the following steps:
the method comprises the following steps: fusing real-time electric quantity data of the electric bus and road order data of the electric bus according to the timestamp and the vehicle number;
the method for acquiring the real-time electric quantity data of the electric bus comprises the following steps: "vehicle number", "electric quantity", and "time stamp" corresponding to real-time electric quantity; the electric bus road list data comprises: "date", "route name", "vehicle number", "driver name", "direction of travel", "route type", "actual departure time", "actual arrival time";
the method for fusing the real-time electric quantity data of the electric bus and the road list data of the electric bus comprises the following steps: firstly, converting a time stamp in the real-time electric quantity data of the electric bus into a form of year/month/day, minute/second; then, according to data of 'date', 'actual departure time', 'actual arrival time' and 'vehicle number' in the electric bus route order data, 'electric quantity' in the real-time electric quantity data of the electric bus is matched into the electric bus route order data, and the data of 'departure time electric quantity' and 'arrival time electric quantity' of the vehicle are added; finally, performing data primary processing, including deleting the data with null values and calculating basic data; wherein calculating the base data comprises: calculating the running time of the vehicle running the journey, and calculating the 'vehicle running time' of the vehicle according to the existing data, namely subtracting the actual departure time from the actual arrival time of the vehicle; calculating the 'consumed electric quantity' of the vehicle in the travel, namely subtracting the electric quantity of the arrival time from the electric quantity of the departure time of the vehicle to finally form training data;
the training data parameters include: "date", "route name", "vehicle number", "driver name", "traveling direction", "route type", "actual departure time", "departure time electric quantity", "actual arrival time", "arrival time electric quantity", "vehicle operation time", and "consumed electric quantity";
step two: extracting characteristic parameters of the departure time;
the departure schedule features are primarily represented by vehicle operating time; the vehicle running time is closely related to the vehicle departure time and the road traffic state, so that the vehicle running time can be expressed as a departure schedule characteristic; predicting the vehicle running time according to the actual departure time of the electric bus in the training data;
the prediction method comprises the following steps: firstly, a departure time characteristic data set is constructed, wherein the data set comprises: the system comprises the following steps of ' departure time ' and ' running time ', wherein the departure time ' is a time period formed by the time interval from the earliest departure time to the latest departure time of a vehicle in training data, and the ' running time ' is data to be solved; secondly, arranging actual departure time in the training data according to an ascending order, grouping the actual departure time according to the same departure time by taking 'minutes' as a minimum time unit, averaging the running time corresponding to the same departure time, taking the running time as the running time of the departure vehicle at the moment, and splicing the obtained running time into a departure time characteristic data set according to the 'departure time'; finally, some blank data exist in the starting time characteristic data set, so that the running time of two adjacent departure times is filled by adopting an equal difference method, and the running time can be smoothed to form a complete starting time characteristic data set; finally, each departure time is formed to have a departure time characteristic corresponding to the departure time characteristic;
meanwhile, inserting the characteristic parameters of the departure time into corresponding training data;
step three: extracting vehicle performance characteristic parameters;
the data characterizing the vehicle performance characteristics in the experimental data is the "vehicle number"; however, the vehicle number cannot visually distinguish the difference of energy consumption of different vehicles, so that the average value and the standard deviation of the energy consumption of the electric bus are used for replacing the vehicle characteristics of the electric bus, namely the vehicle characteristics are digitalized, and the vehicle characteristics are determined from the energy consumption perspective; the average value of the vehicle energy consumption represents the characteristics of the electric quantity consumed by different vehicles, and the standard deviation of the vehicle energy consumption represents the stable condition of the electric quantity consumed by different vehicles;
the specific calculation method is as follows: firstly, grouping data with the same vehicle number in training data, solving the average value and standard deviation of 'energy consumption' in each group of data, and taking the value as a characteristic value representing the energy consumption of the vehicle; finally, forming a vehicle performance characteristic data set, namely, each vehicle number corresponds to a vehicle energy consumption average value and a vehicle energy consumption standard deviation;
meanwhile, inserting the vehicle performance characteristic parameters into corresponding training data;
step four: extracting driver attribute parameters;
the data representing the attribute characteristics of the driver in the experimental data is 'driver name'; the name of the driver cannot visually represent the characteristics of energy consumption of the driver for driving the vehicle, so that the average value and the standard deviation of the energy consumption of the driver are used for replacing the operation characteristics of the driver, namely the characteristics of the driver are digitalized, and the characteristics of the driver are determined from the energy consumption perspective; the average value of driver energy consumption represents the characteristics of electric quantity consumed by different drivers, and the standard deviation of driver energy consumption represents the stable condition of the electric quantity consumed by different drivers;
the specific calculation method is as follows: firstly, grouping data of the same driver name in training data, solving an average value and a standard deviation of 'energy consumption' in each group of data, and taking the value as a characteristic value representing the energy consumption of the driver; finally, forming a driver attribute feature data set, namely, each driver name corresponds to a driver energy consumption average value and a driver energy consumption standard deviation;
simultaneously, inserting the attribute characteristic parameters of the driver into corresponding training data;
step five: processing the original data of the new journey;
firstly, according to the nature of the new journey of the electric bus, selecting the training data of the same route, the same driving direction and the same working day, and calculating the training data according to the second step, the third step and the fourth step to obtain a departure time characteristic data set, a vehicle performance characteristic data set and a driver attribute characteristic data set; then, respectively searching in corresponding data sets according to departure time, vehicle number and driver name in data to be predicted to obtain departure time characteristic parameters, vehicle performance characteristic parameters and driver attribute characteristic parameters of a new journey, and inserting characteristic values into original data to complete initial processing of predicted data;
step six: calculating Euclidean distance between the prediction data and the training data;
firstly, the characteristic value data of training data and prediction data are standardized, and the data standardization principle is as follows:
in formula (II), x' i For normalized data values, x i Representing the raw data value, sigma representing the raw data standard deviation, n representing the number of data samples in the data set, x mean Represents the average value of data object x in the raw data;
then, the principle of the euclidean distance is used:
wherein x is prediction data, C i For the ith training data, x j Representing the j-th characteristic value, C, in the prediction data ij Representing the jth characteristic value in the ith training data, and m represents the quantity of the characteristic values; the above formula calculates the distances between the predicted data characteristic values and all the training data characteristic values one by one;
step seven: giving the predicted energy consumption of the new journey;
before predicting the energy consumption of the new journey of the electric bus, determining the optimal data number of the shortest distance between the predicted data and the training data; the specific calculation method is as follows: firstly, arranging the Euclidean distances calculated in the step six in an ascending order, and respectively taking the first k data with the minimum distance; secondly, calculating the average value of the energy consumption values in the k data as the energy consumption of the prediction data, and further calculating the average error of the prediction data; finally, training 12 groups of data in different directions of different lines respectively to obtain average errors corresponding to different k values; selecting a k value with the minimum average error as an optimal k value, and finally selecting k =7;
and selecting the first 7 training data with the shortest distance, and taking the average value of the energy consumption as the predicted energy consumption of the new journey, namely completing the prediction of the journey energy consumption of the electric bus.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211299932.1A CN115759347A (en) | 2022-10-24 | 2022-10-24 | Method for quickly predicting travel energy consumption of electric bus based on characteristic data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211299932.1A CN115759347A (en) | 2022-10-24 | 2022-10-24 | Method for quickly predicting travel energy consumption of electric bus based on characteristic data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115759347A true CN115759347A (en) | 2023-03-07 |
Family
ID=85352808
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211299932.1A Pending CN115759347A (en) | 2022-10-24 | 2022-10-24 | Method for quickly predicting travel energy consumption of electric bus based on characteristic data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115759347A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116872776A (en) * | 2023-06-21 | 2023-10-13 | 隆瑞三优新能源汽车科技有限公司 | Bus charging power distribution method and device, electronic equipment and medium |
| CN116872776B (en) * | 2023-06-21 | 2024-05-14 | 隆瑞三优新能源汽车科技有限公司 | Bus charging power distribution method and device, electronic equipment and medium |
-
2022
- 2022-10-24 CN CN202211299932.1A patent/CN115759347A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116872776A (en) * | 2023-06-21 | 2023-10-13 | 隆瑞三优新能源汽车科技有限公司 | Bus charging power distribution method and device, electronic equipment and medium |
| CN116872776B (en) * | 2023-06-21 | 2024-05-14 | 隆瑞三优新能源汽车科技有限公司 | Bus charging power distribution method and device, electronic equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110126841B (en) | Pure electric vehicle energy consumption model prediction method based on road information and driving style | |
| CN111666715B (en) | Electric automobile energy consumption prediction method and system | |
| CN109927709B (en) | Vehicle driving route working condition determining method, energy management method and system | |
| CN108761509B (en) | Automobile driving track and mileage prediction method based on historical data | |
| Zhang et al. | Driving cycles construction for electric vehicles considering road environment: A case study in Beijing | |
| Li et al. | Prediction of electric bus energy consumption with stochastic speed profile generation modelling and data driven method based on real-world big data | |
| CN103745110B (en) | Method of estimating operational driving range of all-electric buses | |
| CN110348644B (en) | Method for predicting construction requirements of electric vehicle charging facilities | |
| CN102044149A (en) | City bus operation coordinating method and device based on time variant passenger flows | |
| CN110491158A (en) | A kind of bus arrival time prediction technique and system based on multivariate data fusion | |
| Guo et al. | A novel energy consumption prediction model with combination of road information and driving style of BEVs | |
| Zhou et al. | A multiscale and high-precision LSTM-GASVR short-term traffic flow prediction model | |
| CN114048920A (en) | Site selection layout method, device, equipment and storage medium for charging facility construction | |
| CN115935672A (en) | Fuel cell automobile energy consumption calculation method fusing working condition prediction information | |
| CN116187161A (en) | Intelligent energy management method and system for hybrid electric bus in intelligent networking environment | |
| CN113642768A (en) | Vehicle running energy consumption prediction method based on working condition reconstruction | |
| CN113435777A (en) | Planning method and system for electric operating vehicle charging station | |
| CN111723871B (en) | Estimation method for real-time carriage full load rate of bus | |
| CN113085832B (en) | Energy management method for extended range hybrid vehicle | |
| Zhang et al. | Neural Network based Vehicle Speed Prediction for Specific Urban Driving | |
| CN111784027A (en) | Urban range electric vehicle charging demand prediction method considering geographic information | |
| CN115759347A (en) | Method for quickly predicting travel energy consumption of electric bus based on characteristic data | |
| CN116629425A (en) | Method and device for calculating vehicle energy consumption, computer readable medium and electronic equipment | |
| Liu et al. | Path planning method for electric vehicles based on freeway network | |
| CN113127591B (en) | Position prediction method based on Transformer and LSTM |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |