A hybrid vehicle energy management control system and method thereof
Through the Gaussian process regression model, and combined with the radial-based neural network output the optimal equivalent factor, the problem of poor adaptability of hybrid vehicle energy management strategies in complex environments is solved, and the optimal energy management and fuel consumption strategies are achieved, which improves the The driving efficiency and economicality of hybrid vehicles.
Patent Information
- Application Number
- CN202210896254.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-07-28
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2042-07-28
AI Technical Summary
The existing hybrid vehicle energy management strategies are poorly adaptable in complex practical working environments, making it difficult to achieve optimal energy management.
The Gaussian process regression model is used to predict vehicle speed, and combined with the radial basis neural network outputs the optimal equivalent factor, and the sensor unit collects vehicle speed and SOC data to achieve the control of the strategy of minimum fuel consumption.
It improves the adaptive ability of energy management, ensures optimal energy management in complex environments, improves the accuracy of predicting vehicle speeds and the accuracy of optimal equivalent factors, and ensures that hybrid vehicles drive according to the optimal energy management strategy.
Smart Images

Figure CN115257695B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of hybrid electric vehicles, and in particular to a hybrid electric vehicle energy management control system and method thereof. Background Art
[0002] Hybrid electric vehicles have great potential in alleviating the increasingly serious problems of global warming, oil consumption and environmental pollution. Hybrid electric vehicles are equipped with a variety of energy storage systems and power sources, which provide design freedom for efficient power output of electric motors and internal combustion engines. The energy management system of hybrid electric vehicles controls the coordinated operation of the engine and electric motor based on comprehensive consideration of factors such as the state of charge (SOC) of the power source, vehicle torque demand, engine performance and vehicle speed, so as to achieve efficient distribution of the required power between the two power sources. The energy management strategy plays an important role in coordinating the energy distribution of different power sources, which is of great significance to give full play to the structural advantages of the hybrid system.
[0003] At present, the energy management strategies of hybrid vehicles mainly include rule-based energy management strategies and optimization-based energy management strategies. Among them, rule-based energy management strategies can be subdivided into deterministic rules and fuzzy rules, and optimization-based energy management strategies can be subdivided into global optimization and instantaneous optimization. Rule-based energy management strategies are widely used in real vehicle energy management due to their simplicity, reliability, and strong interpretability. However, the formulation of this energy management strategy relies on experience and has poor adaptability to working conditions. Although the energy management strategy based on optimization can theoretically obtain the optimal energy management under known working conditions, the actual working environment of the vehicle is complex and changeable, which limits the wide application of this method. Summary of the invention
[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a hybrid vehicle energy management control system and method thereof, which predicts the vehicle speed based on Gaussian process regression and combines the equivalent fuel consumption minimum strategy to achieve optimal energy management control of the hybrid vehicle.
[0005] The objective of the present invention can be achieved through the following technical scheme: a hybrid electric vehicle energy management control system, comprising a sensor unit and a vehicle controller, wherein the sensor unit is used to collect driving data of the hybrid electric vehicle, including vehicle speed and SOC data; the vehicle controller outputs a control quantity corresponding to a fuel consumption minimum strategy based on the vehicle speed and SOC data collected by the sensor unit, through vehicle speed prediction and optimal equivalent factor selection, so as to control the driving state of the hybrid electric vehicle.
[0006] Furthermore, the vehicle controller includes a vehicle speed prediction module and an optimal equivalent factor decision module. The vehicle speed prediction module is based on a Gaussian process regression model and is used to predict the vehicle speed. The optimal equivalent factor decision module is based on a radial basis neural network and is used to output an optimal equivalent factor corresponding to a minimum fuel consumption strategy.
[0007] A hybrid electric vehicle energy management control method comprises the following steps:
[0008] S1, obtaining historical vehicle speed and historical SOC data of the hybrid vehicle;
[0009] S2. Based on the historical vehicle speed data, the parameters of the Gaussian process regression model are estimated to establish a vehicle speed prediction model based on Gaussian process regression;
[0010] S3. Use the vehicle speed prediction model to obtain the predicted vehicle speed, take the minimum cumulative equivalent fuel consumption and the balance of low power SOC as the optimization goal, use the optimization algorithm to solve and obtain the optimal equivalent factor, and establish a database including SOC, prediction data, and optimal equivalent factor;
[0011] S4, using the database to train the radial basis neural network;
[0012] S5. During the actual operation, the actual vehicle speed and actual SOC data of the hybrid vehicle are collected through the sensor unit, and the predicted driving speed is obtained using the output of the vehicle speed prediction model;
[0013] The predicted driving speed and actual SOC data are input into the trained radial basis neural network, and the corresponding driving optimal equivalent factor is output. Combined with the actual power demand corresponding to the actual vehicle speed, the optimal control amount of the hybrid vehicle can be determined;
[0014] S6. Based on the optimal control amount, the driving state of the hybrid vehicle is controlled accordingly.
[0015] Furthermore, the step S2 specifically includes the following steps:
[0016] S21, dividing the sequence of historical vehicle speed data into several segments, and taking the original data of the first part of the vehicle speed sequence of each segment as input data and the original data of the second part as output data;
[0017] S22. Using the input data and output data of each segment, the parameters of the Gaussian process regression model are estimated to establish a vehicle speed prediction model based on Gaussian process regression.
[0018] Furthermore, the mathematical expression of the Gaussian process regression model is as follows:
[0019] y=f(x)+ε
[0020] f(x)~GP(μ,∑)
[0021] ε~N(0,σ 2 )
[0022] Among them, x is the input data, y is the output data, f(x) represents a Gaussian process, and ε is the residual that conforms to the Gaussian distribution.
[0023] Furthermore, the step S3 specifically includes the following steps:
[0024] S31, using the vehicle speed prediction model to predict the vehicle speed of each segment in a set future period of time, inputting the predicted vehicle speed into the power demand model to obtain the power demand corresponding to the hybrid vehicle in the set future period of time;
[0025] S32, taking the minimum cumulative equivalent fuel consumption and the balance of low power SOC as the optimization goal, using the optimization algorithm to solve the optimal electric energy consumption equivalent factor of different segments under each power SOC;
[0026] S33. A database is established by combining the predicted vehicle speed, SOC and the optimal electric energy consumption equivalent factor.
[0027] Furthermore, the power demand model is specifically:
[0028]
[0029] Among them, F req is the driving force required by the car, Cd is the drag coefficient, A is the frontal area, v is the vehicle speed, m is the vehicle mass, g is the acceleration of gravity, θ is the road slope, f is the rolling resistance coefficient, and t is the time.
[0030] Furthermore, the objective function of the equivalent fuel consumption strategy with the minimum consumption in step S32 is:
[0031] u * =argmin(ΔE fuel (u)+sΔE elec (u))
[0032] Among them, u * is the optimal control quantity, ΔE fuel is the instantaneous fuel consumption, ΔE elec is the instantaneous power consumption, and s is the equivalent factor.
[0033] Furthermore, the input data of the radial basis function neural network includes SOC and predicted vehicle speed, and the output data of the radial basis function neural network is an optimal equivalent factor.
[0034] Furthermore, the mathematical model of the radial basis neural network is specifically:
[0035]
[0036] Among them, x' is the input variable, y' is the output variable, and w i , β i 、c i is the parameter of the radial basis neural network, and h is the number of hidden layer neurons in the radial basis neural network.
[0037] Compared with the prior art, the present invention has the following advantages:
[0038] 1. In the present invention, the vehicle controller outputs the control quantity corresponding to the minimum fuel consumption strategy through vehicle speed prediction and optimal equivalent factor selection according to the vehicle speed and SOC data collected by the sensor unit to control the driving state of the hybrid vehicle. By setting a vehicle speed prediction module and an optimal equivalent factor decision module in the vehicle controller, the vehicle speed prediction module is used based on the Gaussian process regression model to predict the vehicle speed; the optimal equivalent factor decision module is used based on the radial basis neural network to output the optimal equivalent factor corresponding to the minimum fuel consumption strategy, thereby realizing the prediction of the driving speed of the hybrid vehicle, thereby increasing the adaptive ability of the equivalent fuel consumption minimum strategy, and even in the face of the complex and changeable actual working environment of the vehicle, it is still possible to ensure the optimal energy management.
[0039] 2. The present invention can effectively improve the accuracy of predicting vehicle speed by acquiring historical vehicle speed data that can represent the vehicle driving conditions to train a Gaussian process regression model, thereby ensuring the accuracy of the optimal equivalent factor output by the subsequent radial basis neural network, thereby controlling the hybrid vehicle to complete driving according to the optimal energy management strategy.
[0040] 3. The present invention uses the predicted vehicle speed to calculate the power demand, and takes the minimum cumulative equivalent fuel consumption and the maintenance of the balance of low power SOC as the optimization goal, adopts the optimization algorithm to solve the optimal electric energy consumption equivalent factor of different segments under each power SOC, and then combines the predicted vehicle speed and SOC data to establish a database, and uses the database to train the radial basis function neural network to further ensure the reliability of the optimal equivalent factor output by the radial basis function neural network. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Figure 1 It is a schematic diagram of the system structure of the present invention;
[0042] Figure 2 It is a schematic diagram of the method flow of the present invention;
[0043] Figure 3 A schematic diagram of the process of determining the optimal energy management strategy in the present invention;
[0044] Figure 4 It is a schematic diagram of the operation control principle of the embodiment;
[0045] Figure 5 A schematic diagram of an application framework in an embodiment;
[0046] Description of the markings in the figure:
[0047] 1. Hybrid electric vehicle, 2. Sensor unit, 3. Vehicle controller, 301. Vehicle speed prediction module, 302. Optimal equivalent factor decision module. DETAILED DESCRIPTION
[0048] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.
[0049] Example
[0050] like Figure 1 As shown, a hybrid vehicle energy management control system includes a sensor unit 2 and a vehicle controller 3. The sensor unit 2 is used to collect driving data of the hybrid vehicle 1, including vehicle speed and SOC data; the vehicle controller 3 outputs a control quantity corresponding to a minimum fuel consumption strategy based on the vehicle speed and SOC data collected by the sensor unit 2, through vehicle speed prediction and optimal equivalent factor selection, so as to control the driving state of the hybrid vehicle 1.
[0051] Among them, the vehicle controller 3 includes a vehicle speed prediction module 301 and an optimal equivalent factor decision module 302. The vehicle speed prediction module 301 is based on a Gaussian process regression model and is used to predict the vehicle speed. The optimal equivalent factor decision module 302 is based on a radial basis neural network and is used to output the optimal equivalent factor corresponding to the minimum fuel consumption strategy.
[0052] The above system is applied in practice to realize a hybrid electric vehicle energy management control method, such as Figure 2 As shown, the following steps are included:
[0053] S1, obtaining historical vehicle speed and historical SOC data of the hybrid vehicle;
[0054] S2. Based on the historical vehicle speed data, the parameters of the Gaussian process regression model are estimated to establish a vehicle speed prediction model based on Gaussian process regression;
[0055] S3. Use the vehicle speed prediction model to obtain the predicted vehicle speed, take the minimum cumulative equivalent fuel consumption and the balance of low power SOC as the optimization goal, use the optimization algorithm to solve and obtain the optimal equivalent factor, and establish a database including SOC, prediction data, and optimal equivalent factor;
[0056] S4, using the database to train the radial basis neural network;
[0057] S5. During the actual operation, the actual vehicle speed and actual SOC data of the hybrid vehicle are collected through the sensor unit, and the predicted driving speed is obtained using the output of the vehicle speed prediction model;
[0058] The predicted driving speed and actual SOC data are input into the trained radial basis neural network, and the corresponding driving optimal equivalent factor is output. Combined with the actual power demand corresponding to the actual vehicle speed, the optimal control amount of the hybrid vehicle can be determined;
[0059] S6. Based on the optimal control amount, the driving state of the hybrid vehicle is controlled accordingly.
[0060] This embodiment applies the above technical solution, such as Figure 3 to Figure 5 As shown, the main contents are:
[0061] (1) Use sensors to collect data from real hybrid vehicles during real-world driving, including information about vehicle speed and SOC changes over time;
[0062] The driving speed sequence is divided into several segments for parameter estimation of the Gaussian process regression model to establish a speed prediction model. Each segment is divided into two parts. The first part of the speed sequence is used as the prediction model input x, and the second part of the sequence is used as the accurate output y.
[0063] The mathematical expression of the Gaussian process regression model is as follows:
[0064] y=f(x)+ε
[0065] f(x)~GP(μ,∑)
[0066] ε~N(0,σ 2 )
[0067] Where f(x) is a Gaussian process and ε is the residual that conforms to the Gaussian distribution;
[0068] (2) The Gaussian process regression model is used to predict the driving data of each segment for a period of time in the future, and the data is used as the input of the vehicle power demand model. The power demand model takes into account the longitudinal dynamics of the vehicle, including rolling resistance, wind resistance, acceleration resistance, and inertial resistance. The mathematical expression of the power demand model is as follows:
[0069]
[0070] In the formula, F reqis the driving force of automobile demand, Cd is the drag coefficient, A is the frontal area, v is the vehicle speed, m is the vehicle mass, g is the acceleration of gravity, θ is the road slope, f is the rolling resistance coefficient, and t is the time;
[0071] The vehicle power requirement calculated based on the predicted vehicle speed is regarded as the vehicle power requirement for a period of time in the future;
[0072] Then, taking the minimum cumulative equivalent fuel consumption and maintaining the balance of low power SOC as the optimization goal, the optimization algorithm is used to solve the optimal electric energy consumption equivalent factor of different segments under each power SOC. The mathematical expression of the equivalent fuel consumption strategy is as follows:
[0073] u * =argmin(ΔE fuel (u)+sΔE elec (u))
[0074] In the formula, u * is the optimal control quantity, ΔE fuel is the instantaneous fuel consumption, ΔE elec is the instantaneous power consumption, s is the equivalent factor;
[0075] Then, a database including SOC, prediction data, and optimal equivalent factors was established;
[0076] (3) The established database is used to train the radial basis function neural network. The input is the SOC value and the predicted vehicle speed. The output is the optimal equivalent factor. The mathematical model of the radial basis function neural network is as follows:
[0077]
[0078] In the formula, x' is the input variable, y' is the output variable, and w i , β i 、c i is the parameter of the neural network, h is the number of neurons in the hidden layer;
[0079] (4) During actual operation, the vehicle records the driving data (including vehicle speed, SOC, etc.) for a period of time in the past, uses the Gaussian process regression model to predict the vehicle speed for a period of time in the future, inputs the predicted data and the current SOC value into the trained radial basis neural network, and outputs the optimal equivalent factor. Then, according to the equivalent fuel consumption minimization strategy in step (2), the optimal control amount of the hybrid vehicle is determined to control the vehicle driving.
[0080] In this embodiment, taking a series hybrid vehicle as an example, compared with a pure electric vehicle, a series hybrid vehicle has an additional range extender that can charge the battery or power the drive motor, thereby increasing the vehicle's driving range; the energy management strategy of a series hybrid vehicle is mainly to control the output power of the range extender.
[0081] The application process includes:
[0082] First, use sensors to collect data from actual hybrid vehicles in the real world, including vehicle speed, SOC changes over time, etc.
[0083] Then the driving speed sequence is divided into several segments (for example, a 15-second sequence is one segment) for parameter estimation of the Gaussian process regression model to establish a speed prediction model. Each segment is divided into two parts. The first part of the speed sequence (for example, 10 seconds) is used as the prediction model input x, and the second part of the sequence (for example, 5 seconds) is used as the accurate output y.
[0084] Among them, the mathematical expression of the Gaussian process regression model is as follows:
[0085] P(Y|f(X),X)~N(Y|h(X) T β+f(X),σ 2 )
[0086] f(x)~GP(0,k(X,X))
[0087] Where X is the input of the training set, Y is the output of the training set, h(X) is a linear function that maps the training set to a high-dimensional space, f(x) is a Gaussian process with a mean of 0, k(·) is the kernel function of the Gaussian process, and β, σ 2 is the model parameter. According to the relevant theory of Gaussian process regression, it is predicted that under the new input x new Output y new The expectation is as follows:
[0088]
[0089] α=(k(X,X)+σ 2 ) -1 (Y-Hβ)
[0090] Second, the Gaussian process regression model is used to predict the driving data of each segment in the future and use it as the input of the following vehicle power demand model; considering the longitudinal dynamics of the vehicle, including rolling resistance, wind resistance, acceleration resistance, and inertial resistance, the mathematical expression of the power demand model is as follows:
[0091]
[0092] In the formula, F req is the driving force of automobile demand, Cd is the drag coefficient, A is the frontal area, v is the vehicle speed, m is the vehicle mass, g is the acceleration of gravity, θ is the road slope, f is the rolling resistance coefficient, and t is the time;
[0093] The vehicle power demand calculated based on the predicted speed is regarded as the power demand of the vehicle in the future. The minimum cumulative equivalent fuel consumption and the low power SOC are used to maintain balance, and linear superposition with different weights is used as the optimization goal. The optimization algorithm (such as the quadratic programming algorithm, etc.) is used to solve the optimal electric energy consumption equivalent factor of different predicted speed segments under each power SOC. The equivalent fuel consumption strategy is mathematically expressed as follows:
[0094] u * =argmin(ΔE fuel (u)+sΔE elec (u))
[0095] In the formula, u * is the optimal control quantity, ΔE fuel is the instantaneous fuel consumption, ΔE elec is the instantaneous power consumption, s is the equivalent factor;
[0096] Finally, a database including SOC, prediction data, and optimal equivalent factors is established;
[0097] 3. Use the established database to train the radial basis neural network. The input is the SOC value and the predicted driving data. The output is the optimal equivalent factor. The mathematical model of the radial basis neural network is as follows:
[0098]
[0099] In the formula, x' is the input variable, y' is the output variable, and w i , β i 、c i is the parameter of the neural network, h is the number of neurons in the hidden layer;
[0100] 4. During actual operation, the vehicle records the driving data of the past period (including vehicle speed, SOC, etc.), uses the Gaussian process regression model to predict the vehicle speed in the future, and inputs the predicted data and the current SOC value into the trained radial basis neural network, which outputs the optimal equivalent factor. Then, based on the strategy of minimizing equivalent fuel consumption, the optimal range extender output power of the series hybrid vehicle is determined to control vehicle driving.
[0101] In summary, the technical solution can predict the actual working conditions, so that the energy management strategy has predictive ability and adaptability. Since the prediction means for the vehicle driving conditions are added, the adaptability of the equivalent fuel consumption minimum strategy can be increased;
[0102] This technical solution selects driving data that can represent the vehicle's driving conditions to train the Gaussian process regression model, which can improve the accuracy of prediction and thus improve the economy of hybrid vehicles;
[0103] This technical solution can be combined with vehicle networking technology. The actual running vehicles can upload data to the cloud, and this technical solution can be used in the cloud to optimize the energy management strategy.
Claims
1. A hybrid electric vehicle energy management control method, characterized in that: The following steps are involved: S1, obtaining historical vehicle speed and historical SOC data of the hybrid vehicle; S2. Based on the historical vehicle speed data, the parameters of the Gaussian process regression model are estimated to establish a vehicle speed prediction model based on Gaussian process regression; S3. Use the vehicle speed prediction model to obtain the predicted vehicle speed, take the minimum cumulative equivalent fuel consumption and the balance of low power SOC as the optimization goal, use the optimization algorithm to solve and obtain the optimal equivalent factor, and establish a database including SOC, prediction data and optimal equivalent factor; S4, using the database to train the radial basis neural network; S5. During the actual operation, the actual vehicle speed and actual SOC data of the hybrid vehicle are collected through the sensor unit, and the predicted driving speed is obtained using the output of the vehicle speed prediction model; The predicted driving speed and actual SOC data are input into the trained radial basis neural network, and the corresponding driving optimal equivalent factor is output. Combined with the actual power demand corresponding to the actual vehicle speed, the optimal control amount of the hybrid vehicle can be determined; S6, based on the optimal control amount, correspondingly controlling the driving state of the hybrid vehicle; The step S2 specifically includes the following steps: S21, dividing the sequence of historical vehicle speed data into several segments, and taking the original data of the first part of the vehicle speed sequence of each segment as input data and the original data of the second part as output data; S22, using the input data and output data of each segment, estimating the parameters of the Gaussian process regression model, and establishing a vehicle speed prediction model based on Gaussian process regression; The step S3 specifically comprises the following steps: S31, using the vehicle speed prediction model to predict the vehicle speed of each segment in a set future period of time, inputting the predicted vehicle speed into the power demand model to obtain the power demand corresponding to the hybrid vehicle in the set future period of time; S32, taking the minimum cumulative equivalent fuel consumption and the balance of low power SOC as the optimization goal, using the optimization algorithm to solve the optimal electric energy consumption equivalent factor of different segments under each power SOC; S33, combining the predicted vehicle speed, SOC and the optimal electric energy consumption equivalent factor to establish a database; The objective function of the equivalent fuel consumption strategy with minimum consumption in step S32 is: he * =argmin(ΔE fuel (u)+sΔE elec (he)) Among them, u * is the optimal control quantity, ΔE fuel is the instantaneous fuel consumption, ΔE elec is the instantaneous power consumption, and s is the equivalent factor.
2. A hybrid vehicle energy management control method according to claim 1, characterized in that: The mathematical expression of the Gaussian process regression model is as follows: y=f(x)+ε f(x)~GP(μ,∑) ε~N(0,σ 2 ) Among them, x is the input data, y is the output data, f(x) represents a Gaussian process, and ε is the residual that conforms to the Gaussian distribution.
3. A hybrid vehicle energy management control method according to claim 1, characterized in that: The power demand model is specifically: Among them, F req is the driving force required by the car, Cd is the drag coefficient, A is the frontal area, v is the vehicle speed, m is the vehicle mass, g is the acceleration of gravity, θ is the road slope, f is the rolling resistance coefficient, and t is the time.
4. A hybrid vehicle energy management control method according to claim 1, characterized in that: The input data of the radial basis function neural network includes SOC and predicted vehicle speed, and the output data of the radial basis function neural network is an optimal equivalent factor.
5. A hybrid vehicle energy management control method according to claim 4, characterized in that: The mathematical model of the radial basis neural network is specifically: Among them, x' is the input variable, y' is the output variable, and w i , β i 、c i is the parameter of the radial basis neural network, and h is the number of hidden layer neurons in the radial basis neural network.
6. A hybrid electric vehicle energy management control system, used to implement the hybrid electric vehicle energy management control method as claimed in claim 1, characterized in that: It includes a sensor unit and a vehicle controller. The sensor unit is used to collect driving data of the hybrid vehicle, including vehicle speed and SOC data. The vehicle controller outputs a control quantity corresponding to a minimum fuel consumption strategy based on the vehicle speed and SOC data collected by the sensor unit, through vehicle speed prediction and optimal equivalent factor selection, so as to control the driving state of the hybrid vehicle.
7. A hybrid vehicle energy management control system according to claim 6, characterized in that: The vehicle controller includes a vehicle speed prediction module and an optimal equivalent factor decision module. The vehicle speed prediction module is based on a Gaussian process regression model and is used to predict the vehicle speed. The optimal equivalent factor decision module is based on a radial basis neural network and is used to output an optimal equivalent factor corresponding to a minimum fuel consumption strategy.
