Regression analysis-minimum principle-based energy management research method

Through regression analysis-minimum principle combined with Pontryagin minimum value principle, adaptively adjust comorphic variables to simplify energy management functions, the contradiction between the control effect and real-time nature of hybrid vehicle energy management strategies is solved, and more efficient is achieved fuel economy and real-time.

CN119089107APending Publication Date: 2024-12-06JILIN UNIVERSITY
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411225809.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

There is a contradiction between the control effect and real-time nature of hybrid vehicles. Rules-based strategies ensure real-timeness but cannot obtain global optimality. The large amount of calculation of the global optimal method makes it difficult to apply.

Method used

The energy management research method using regression analysis-minimum value principle, combined with the Pontryagin minimum value principle, adaptively adjust the comorphic variables, introduce fuel consumption model and current regression model, simplify the energy management function, and quickly obtain the analysis of Hamilton function through mathematical derivation Solution to achieve optimal power distribution.

Benefits of technology

It improves the fuel economy of energy management, while enhancing the real-time nature of the algorithm, reducing the solution difficulty and partial-direction calculation complexity, and achieving better control effect and real-time nature.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089107A_ABST
    Figure CN119089107A_ABST
Patent Text Reader

Abstract

The invention relates to an energy management research method based on a regression analysis-minimum value principle. In the predictive cruising process of the hybrid electric vehicle, information such as the vehicle speed, gears and external characteristics in the cruising process is obtained, and the vehicle required power is obtained in cooperation with speed planning; the method comprises the following steps: based on a core idea of introducing an equivalent factor and a co-state variable based on an ECMS algorithm, updating the co-state variable and the equivalent factor in real time according to a battery SOC and a battery capacity, and solving battery equivalent oil consumption through a Hamilton function of energy management; simplifying a vehicle key dynamic model by utilizing regression analysis, and further arranging a Hamilton function for energy management into a standard quadratic function form containing a control variable and a co-state variable in combination with a minimum value principle; the method comprises the following steps of: establishing a Hamilton function, quickly obtaining an analytical solution of the Hamilton function through mathematical derivation to obtain an optimal power distribution mode, establishing respective Hamilton functions of adjacent gears at the same time, comparing total fuel consumption Hamilton function values under the adjacent gears, and obtaining a corresponding optimal torque distribution mode when the minimum value is the optimal gear.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of hybrid vehicle energy management, and in particular relates to an energy management research method based on regression analysis and minimum principle. Background Art

[0002] In recent years, the growing conflict between energy shortages and environmental pollution has led countries to announce timetables for banning the sale of traditional fuel vehicles. This has prompted the automotive industry to accelerate the development of new energy vehicle technologies. Hybrid vehicles, currently the most mature new energy vehicle, offer reduced pollutant emissions while maintaining a longer driving range compared to traditional fuel vehicles.

[0003] Because hybrid electric vehicles (HEVs) have dual power sources, how to rationally allocate the output power of each source to improve energy efficiency while reducing fuel consumption has become a hot topic for many researchers. To achieve real-time optimization-based HEV energy management, researchers have developed instantaneous optimization energy management methods such as the Equivalent Consumption Minimum Strategy (ECMS), model predictive control, and neural networks. The ECMS method is based on the Pontryagin's Minimum Principle (PMP). By setting the fuel-electric equivalence factor, the vehicle's equivalent fuel consumption is minimized at each moment, thereby achieving instantaneous optimization. Current HEV energy management strategies primarily suffer from a conflict between control effectiveness and real-time performance. Rule-based energy management strategies can ensure real-time performance but cannot achieve global optimality. Global optimal energy management methods, such as dynamic programming, are computationally intensive, making them difficult to implement in real vehicles. Given the limitations of rule-based and global optimal methods, designing instantaneous optimization strategies that balance optimal control effectiveness and real-time performance is key to further exploring the energy-saving potential of HEVs.

[0004] Therefore, this paper proposes an energy management research method based on regression analysis-minimum principle. The ECMS method is derived through the PMP principle, the co-state variables are adaptively adjusted, and the fuel consumption model and current regression model are introduced to simplify the complex energy management function, thereby improving the fuel economy of energy management while enhancing the real-time performance of the algorithm solution. Summary of the Invention

[0005] 1. A regression analysis-minimum principle energy management research method, characterized by: During predictive cruising, based on an energy management strategy that minimizes vehicle equivalent fuel consumption, a hybrid vehicle utilizes regression analysis to simplify the vehicle's key dynamic model. This is combined with the Pontryagin minimum principle, while adaptively adjusting co-state variables and equivalent factors. Furthermore, through mathematical derivation, an analytical solution to the Hamiltonian function is rapidly obtained, resulting in an optimal power distribution scheme. The Hamiltonian function values ​​for total fuel consumption in adjacent gears are compared to determine the corresponding optimal gear and torque distribution scheme.

[0006] Step 1: Establish an equivalent fuel consumption minimization strategy model based on the minimum principle: The current vehicle speed, gear position, and external characteristic curve are used to determine the allowable torque range of the vehicle engine. The torque within the allowable engine range is discretized and combined with the vehicle power requirement obtained from speed planning to obtain a series of discrete torque distribution combinations. The real-time fuel consumption and battery power of each distribution method are then calculated. Based on the core concepts of the ECMS algorithm, which introduces equivalent factors and co-state variables, the battery equivalent fuel consumption is solved using the Hamiltonian function of energy management. The optimal control variable is determined by taking the minimum equivalent fuel consumption.

[0007] The entire control system defines the state variable as the battery SOC and the engine torque as the control variable. The Hamilton function and equivalent factors for energy management are defined as follows:

[0008] H(x(t),λ(t),u(t),t)=λ(t)f(x(t)),P bat (t),t)+B e (u(t))……(1.1)

[0009]

[0010] Where: State equation battery SOC changes:

[0011]

[0012] λ(t) is an n-dimensional covariate variable that satisfies the following conditions:

[0013] Where U oc is the open circuit voltage of the battery, R bat is the internal resistance of the battery, Q bat is the battery capacity, P bat Battery output power, B e is the engine fuel consumption, s(t) represents the equivalent factor at time t, Q lhv is the lower calorific value of the engine fuel, E bat is the total energy of the battery;

[0014] Step 2: Real-time update of co-state variables and equivalent factors: Based on the Pontryagin minimum principle, the Hamiltonian function is minimized at each instant. In other words, the global optimal problem of vehicle energy management is transformed into an instantaneous optimal problem. The key to solving the Hamiltonian function minimum in this paper is to obtain accurate co-state variables and update the equivalent factors in real time.

[0015]

[0016] Step 3: Establishing regression models for key components of hybrid vehicles: Based on the minimum principle in Step 1, the Hamiltonian function is rationally simplified through regression models. This paper replaces the state variable change equations in the complex Hamiltonian function with regression models of engine power and fuel consumption, and regression models of battery current with respect to SOC and motor mechanical power, to reduce the computational complexity of the function.

[0017] Among them, the regression model of engine power and fuel consumption is as follows:

[0018]

[0019] Regression analysis is performed on the current, motor mechanical power, and battery SOC, and a quadratic polynomial regression model of the battery current is established as follows:

[0020]

[0021] Step 4: Select the shift command and the motor output torque as the control quantity to indirectly obtain the engine output torque; when the car is cruising, the motor torque is a continuous variable and the gear is a discrete variable. The vehicle has three shift commands to choose from: maintaining the current gear, upshifting, and downshifting; the three shift commands correspond to three Hamilton functions. To achieve the optimal solution, the minimum values ​​of the three Hamilton functions are solved at the same time. The gear corresponding to the minimum Hamilton function value is the optimal gear, and the motor torque corresponding to the Hamilton function value is the optimal motor torque. Define three independent Hamilton functions as follows:

[0022]

[0023] Step 5: Optimal engine and motor torque distribution under different gears: First, the required driving force of the vehicle is obtained according to the speed planning, and then the required power of the vehicle is calculated, and P req =P e +P m ;

[0024] Incorporate the regression model analysis results into the Hamilton function of formula 1.1 as follows:

[0025]

[0026] In the formula, the Hamilton function value on the left side of equation 5.1 is the sum of the equivalent fuel consumption obtained by converting the engine fuel consumption and the electric energy consumption. The first term on the right side of the equation is the engine fuel consumption, and the second term on the right side of the equation is the equivalent fuel consumption obtained by converting the electric energy consumption. ij is the current regression model parameter b ij and battery capacity Q bat Calculated;

[0027] In order to facilitate calculation, the Hamilton function is further organized into a standard quadratic function form containing control variables and co-state variables:

[0028] Form 1: H(t) = A1(t)SOC 2 (t)+B1(t)SOC(t)+C1(t)………(5.2)Form 2: Where, the parameters of the quadratic function are as follows:

[0029] A1(t)=λ(t)p 20

[0030] B1(t)=λ(t)(p 10 +p 11 P m (t))

[0031]

[0032] A2(t)=a2+λ(t)p 02

[0033] B2(t)=(-a1-2a2P req (t))+λ(t)(p 01 +p 11 SOC(t))

[0034]

[0035] Step 6: After obtaining the quadratic function form of the Hamilton function in Equations 5.1 and 5.2, the optimal decision for the control quantity of the Hamilton function is the analytical solution of the quadratic function with respect to the motor output power. When the motor output power is the optimal motor output power at the current moment, the optimal motor operating point is obtained. At this time, the Hamilton function takes the minimum value, which means the optimal engine and motor output torque distribution is obtained. Considering the basic properties of the quadratic function, the analytical solution for energy management is:

[0036]

[0037] Where, P m,min (t) and P m,max (t) are the minimum and maximum mechanical power of the motor at the current moment, P m,opt (t) is the optimal motor output power at the current moment;

[0038] Step 7: Use the optimal motor output power solved by formula 6.1 to quickly and accurately solve the optimal power distribution method under a specific gear; when the mechanical efficiencies of adjacent gears are similar, use the Hamilton function of adjacent gears in formula 5.1 to calculate the corresponding optimization function value, and then compare the three function values ​​to obtain the corresponding optimal shifting behavior at the current moment, and then obtain the corresponding optimal torque distribution method and gear.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] 1. The energy management research method based on regression analysis and minimum principle described in this patent replaces the complex state quantity change equation with the vehicle regression model, effectively simplifying the complexity of partial derivative calculations, improving the control effect while reducing the difficulty of solution.

[0041] 2. The energy management research method based on regression analysis-minimum principle described in this patent adaptively adjusts the co-state variables based on the equivalent fuel consumption minimum strategy to obtain better fuel economy.

[0042] 3. This method quickly obtains the analytical solution of the Hamilton function through mathematical derivation, reduces discrete and iterative operations, improves the fuel economy of energy management, and enhances the real-time performance of the algorithm solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The following description of the embodiments will become easier to understand with reference to the accompanying drawings, in which:

[0044] Figure 1 A control flow chart of an energy management research method based on regression analysis-minimum principle according to an embodiment of the present invention;

[0045] Figure 2 A real-time solution flow chart of the energy management research method based on the regression analysis-minimum principle according to an embodiment of the present invention;

[0046] Figure 3 This is a simplified structural diagram of a P2 coaxial parallel hybrid vehicle applicable to an example of the present invention. DETAILED DESCRIPTION

[0047] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention.

[0048] Reference Attachment Figure 1 and Figure 2 To describe the energy management research method and solution process of regression analysis-minimum principle, but the present invention is not limited to these embodiments.

[0049] Step 1: Establish an equivalent fuel consumption minimization strategy model based on the minimum principle: The current vehicle speed, gear position, and external characteristic curve are used to determine the allowable torque range of the vehicle engine. The torque within the allowable engine range is discretized and combined with the vehicle power requirement obtained from speed planning to obtain a series of discrete torque distribution combinations. The real-time fuel consumption and battery power of each distribution method are then calculated. Based on the core concepts of the ECMS algorithm, which introduces equivalent factors and co-state variables, the battery equivalent fuel consumption is solved using the Hamiltonian function of energy management. The optimal control variable is determined by taking the minimum equivalent fuel consumption.

[0050] The entire control system defines the state variable as the battery SOC and the engine torque as the control variable. The Hamilton function and equivalent factors for energy management are defined as follows:

[0051] H(x(t),λ(t),u(t),t)=λ(t)f(x(t)),P bat (t),t)+B e (u(t))……(1.1)

[0052]

[0053] Where: State equation battery SOC changes:

[0054]

[0055] λ(t) is an n-dimensional covariate variable that satisfies the following conditions:

[0056] Where U oc is the open circuit voltage of the battery, R bat is the internal resistance of the battery, Q bat is the battery capacity, P bat Battery output power, B e is the engine fuel consumption, s(t) represents the equivalent factor at time t, Q lhv is the lower calorific value of the engine fuel, E batis the total energy of the battery;

[0057] Step 2: Real-time update of co-state variables and equivalent factors: Based on the Pontryagin minimum principle, the Hamiltonian function is minimized at each instant. In other words, the global optimal problem of vehicle energy management is transformed into an instantaneous optimal problem. The key to solving the Hamiltonian function minimum in this paper is to obtain accurate co-state variables and update the equivalent factors in real time.

[0058]

[0059] Step 3: Establishing regression models for key components of hybrid vehicles: Based on the minimum principle in Step 1, the Hamiltonian function is rationally simplified through regression models. This paper replaces the state variable change equations in the complex Hamiltonian function with regression models of engine power and fuel consumption, and regression models of battery current with respect to SOC and motor mechanical power, to reduce the computational complexity of the function.

[0060] Among them, the regression model of engine power and fuel consumption is as follows:

[0061]

[0062] Regression analysis is performed on the current, motor mechanical power, and battery SOC, and a quadratic polynomial regression model of the battery current is established as follows:

[0063]

[0064] Step 4: Select the shift command and the motor output torque as the control quantity to indirectly obtain the engine output torque; when the car is cruising, the motor torque is a continuous variable and the gear is a discrete variable. The vehicle has three shift commands to choose from: maintaining the current gear, upshifting, and downshifting; the three shift commands correspond to three Hamilton functions. To achieve the optimal solution, the minimum values ​​of the three Hamilton functions are solved at the same time. The gear corresponding to the minimum Hamilton function value is the optimal gear, and the motor torque corresponding to the Hamilton function value is the optimal motor torque. Define three independent Hamilton functions as follows:

[0065]

[0066] Step 5: Optimal engine and motor torque distribution under different gears: First, the required driving force of the vehicle is obtained according to the speed planning, and then the required power of the vehicle is calculated, and P req =P e +P m ;

[0067] Incorporate the regression model analysis results into the Hamilton function of formula 1.1 as follows:

[0068]

[0069] In the formula, the Hamilton function value on the left side of equation 5.1 is the sum of the equivalent fuel consumption obtained by converting the engine fuel consumption and the electric energy consumption. The first term on the right side of the equation is the engine fuel consumption, and the second term on the right side of the equation is the equivalent fuel consumption obtained by converting the electric energy consumption. ij is the current regression model parameter b ij and battery capacity Q bat Calculated;

[0070] In order to facilitate calculation, the Hamilton function is further organized into a standard quadratic function form containing control variables and co-state variables:

[0071] Form 1: H(t) = A1(t)SOC 2 (t)+B1(t)SOC(t)+C1(t)………(5.2)Form 2: Where, the parameters of the quadratic function are as follows:

[0072] A1(t)=λ(t)p 20

[0073] B1(t)=λ(t)(p 10 +p 11 P m (t))

[0074]

[0075] A2(t)=a2+λ(t)p 02

[0076] B2(t)=(-a1-2a2P req (t))+λ(t)(p 01 +p 11 SOC(t))

[0077]

[0078] Step 6: After obtaining the quadratic function form of the Hamilton function in Equations 5.1 and 5.2, the optimal decision for the control quantity of the Hamilton function is the analytical solution of the quadratic function with respect to the motor output power. When the motor output power is the optimal motor output power at the current moment, the optimal motor operating point is obtained. At this time, the Hamilton function takes the minimum value, which means the optimal engine and motor output torque distribution is obtained. Considering the basic properties of the quadratic function, the analytical solution for energy management is:

[0079]

[0080] Where, P m,min (t) and P m,max (t) are the minimum and maximum mechanical power of the motor at the current moment, P m,opt (t) is the optimal motor output power at the current moment;

[0081] Step 7: Use the optimal motor output power solved by formula 6.1 to quickly and accurately solve the optimal power distribution method under a specific gear; when the mechanical efficiencies of adjacent gears are similar, use the Hamilton function of adjacent gears in formula 5.1 to calculate the corresponding optimization function value, and then compare the three function values ​​to obtain the corresponding optimal shifting behavior at the current moment, and then obtain the corresponding optimal torque distribution method and gear.

[0082] This paper introduces the regression model into the Hamilton function, transforms the original function into a quadratic function, and then obtains the change law of the co-state variable. By using the basic properties of the quadratic function, the analytical solution of the function can be quickly obtained, and the optimal gear and torque distribution at the current moment can be derived.

[0083] The present invention frequently uses terms such as SOC, but does not exclude the possibility of using other terms. These terms are used only to more conveniently describe and explain the essence of the present invention; interpreting them as any additional limitations is contrary to the spirit of the present invention.

[0084] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

Claims

1. A regression analysis-minimum principle energy management research method, characterized by: During the predictive cruising process, hybrid vehicles introduce co-state variables about the state SOC based on the energy management strategy of minimizing the equivalent fuel consumption of the vehicle. Regression analysis is used to simplify the key power model of the vehicle, combined with the Pontryagin minimum principle, and the co-state variables are adaptively adjusted. Then, the analytical solution of the Hamilton function is quickly obtained through mathematical derivation, and the optimal power distribution method is obtained. The total fuel consumption Hamilton function values ​​under adjacent gears are compared to obtain the corresponding optimal gear and torque distribution method. Step 1: Establishment of the minimum equivalent fuel consumption strategy model based on the minimum principle: The current vehicle speed information, gear information, and external characteristic curve can be used to obtain the allowable torque range of the current vehicle engine; discretize the torque within the allowable range of the engine, and use the vehicle demand power obtained by speed planning to obtain a series of discrete torque distribution combinations, and calculate the real-time fuel consumption and battery power of each distribution method; Based on the core idea of ​​introducing equivalent factors and co-state variables based on the ECMS algorithm, the battery equivalent fuel consumption is solved through the Hamilton function of energy management, and the optimal control variable is obtained by taking the minimum equivalent fuel consumption; The entire control system defines the state quantity as the battery SOC and the engine torque as the control quantity, where the Hamilton function and equivalent factors of energy management are defined as follows: H(x(t),λ(t),u(t),t)=λ(t)f(x(t)),P bat (t),t)+B e (u(t))………(1.1) Where: State equation battery SOC changes: λ(t) is an n-dimensional covariate variable that satisfies the following conditions: Where U oc is the open circuit voltage of the battery, R bat is the internal resistance of the battery, Q bat is the battery capacity, P bat Battery output power, B e is the engine fuel consumption, s(t) represents the equivalent factor at time t, Q lhv is the lower calorific value of the engine fuel, E bat is the total energy of the battery; Step 2: Real-time update of co-state variables and equivalent factors: According to the Pontryagin minimum principle, minimize the Hamilton function at each instant. In other words, the global optimal problem of vehicle energy management is transformed into an instantaneous optimal problem. The key to solving the minimum value of the Hamilton function in this paper is to obtain accurate co-state variables and update the equivalent factors in real time. Step 3: Establishment of regression model of key components of hybrid vehicles: Based on the minimum principle in step 1, the Hamilton function is reasonably simplified through the regression model; this paper replaces the state quantity change equation in the complex Hamilton function with the regression model of engine power and fuel consumption and the regression model of battery current on SOC and motor mechanical power to reduce the calculation complexity of the function; Among them, the regression model of engine power and fuel consumption is as follows: The current, motor mechanical power, and battery SOC are subjected to regression analysis, and the battery current quadratic polynomial regression model is established as follows: Step 4: Select the shift command and the motor output torque as the control quantity to indirectly obtain the output torque of the engine; when the car is cruising, the motor torque is a continuous variable and the gear is a discrete variable. The vehicle has three options for shift commands, namely, maintaining the current gear, upshifting and downshifting; the three shift commands correspond to three Hamilton functions. To achieve the optimal solution, the minimum values ​​of the three Hamilton functions are solved at the same time. The gear corresponding to the minimum Hamilton function value is the optimal gear, and the motor torque corresponding to the Hamilton function value is the optimal motor torque; define three independent Hamilton functions as follows: Step 5: Optimal engine and motor torque distribution under different gears: First, the required driving force of the vehicle is obtained according to the speed planning, and then the required power of the vehicle is calculated, and P req =P e +P m ; Incorporate the regression model analysis results into the Hamilton function of formula 1.1, as follows: In the formula, the Hamilton function value on the left side of equation 5.1 is the sum of the equivalent fuel consumption obtained by converting the engine fuel consumption and the electric energy consumption. The first term on the right side of the equation is the engine fuel consumption, and the second term on the right side of the equation is the equivalent fuel consumption obtained by converting the electric energy consumption. ij is the current regression model parameter b ij With battery capacity Q bat Calculate and obtain; In order to facilitate calculation, the Hamilton function is further organized into a standard quadratic function form containing control variables and co-state variables: Form 1: H(t) = A1(t)SOC 2 (t)+B1(t)SOC(t)+C1(t)………(5.2)Form 2: In the formula, the parameters of the quadratic function are as follows: A1(t)=λ(t)p 20 B1(t)=λ(t)(p 10 +p 11 P m (t)) A2(t)=a2+λ(t)p 02 B2(t)=(-a1-2a2P req (t))+λ(t)(p 01 +p 11 SOC(t)) Step 6: After obtaining the quadratic function form of the Hamilton function in equations 5.1 and 5.2, the optimal decision of the control quantity of the Hamilton function number is the analytical solution of the quadratic function on the motor output power. When the motor output power is the optimal motor output power at the current moment, the optimal motor operating point is obtained. At this time, the Hamilton function obtains the minimum value, that is, the optimal engine and motor output torque distribution is obtained; considering the basic properties of the quadratic function, the analytical solution of energy management is: Where P m,min (t) and P m,max (t) are the minimum and maximum mechanical power of the motor at the current moment, P m,opt (t) is the optimal motor output power at the current moment; Step 7: Use the optimal motor output power solved by equation 6.1 to quickly and accurately solve the optimal power distribution method for a specific gear; when the mechanical efficiencies of adjacent gears are similar, use the Hamilton function of adjacent gears in equation 5.1 to calculate the corresponding optimization function value, and then compare the three function values ​​to obtain the corresponding optimal shifting behavior at the current moment, and then obtain the corresponding optimal torque distribution method and gear.

Citation Information

Cited By

  • Four-wheel-drive plug-in hybrid vehicle energy management method based on approximate minimum value principle

    CN121697607A

  • Energy management method for plug-in hybrid electric vehicle

    CN122519231A