Speed-energy integrated optimization control method for intelligent fuel cell vehicles

Abstract

The present invention relates to an integrated optimization control method for vehicle speed and energy of an intelligent fuel cell vehicle, the method comprising: obtaining traffic information on the road during the driving process of the main vehicle, predicting the speed trajectory of the front vehicle at the future moment based on the ELM algorithm; taking the economy and comfort of the fuel cell vehicle as the optimization target, integrating the main vehicle speed planning and energy management strategy into an integrated architecture, and solving the optimal control problem based on the MPC framework and applying the PMP principle; in addition, the present invention proposes a control architecture based on multi-core parallel computing, that is, allocating the front vehicle speed prediction part to one core of the multi-core chip, and allocating the integrated energy management algorithm to another core of the multi-core chip, so as to significantly reduce the calculation burden of the control system and enhance the ability of information processing at different time scales. The present invention fully considers the uncertainty of the traffic environment, implements the vehicle speed planning and energy management strategy in a dynamic environment, and realizes the integrated optimization control of vehicle speed planning and energy management.

Description

Technical Field

[0001] The invention relates to the technical field of fuel cell vehicle energy management, and in particular to a speed-energy integrated optimization control method for an intelligent fuel cell vehicle. Background Art

[0002] Hydrogen fuel cell vehicles mainly use hydrogen as fuel, and generate electricity through the chemical reaction between hydrogen and water to drive the vehicle. Compared with traditional vehicles, fuel cell vehicles have no pollution. Compared with pure battery vehicles, fuel cells have high energy density, strong environmental adaptability, and long driving range. They have become an important direction for the development of new energy vehicles. However, the low power density and low dynamic characteristics of fuel cells require power battery packs as auxiliary energy storage devices to provide peak power and recover braking energy. Faced with complex and changing road environments and traffic needs, it is of great significance to better realize the power distribution control of multiple energy sources and maximize their economic potential.

[0003] The rapid development of interconnected and autonomous driving technologies has further enhanced the huge potential of fuel cell vehicle energy management, especially in complex urban traffic environments, which plays an important role in reducing energy consumption and improving transportation efficiency. In this context, speed planning provides unprecedented opportunities to improve energy efficiency. More and more studies have incorporated vehicle speed planning into the research system of automotive energy management. When speed planning and EMS are jointly optimized, it is usually called co-optimization. Co-optimization methods can achieve energy savings under various driving conditions while considering vehicle motion and powertrain. Compared with not considering speed planning, the co-optimization of speed planning and energy management can significantly reduce the on-off cycles, frequent load change conditions, and excessive or low propulsion power requirements of fuel cell vehicles, which are the main causes of fuel cell degradation and high operating costs. Therefore, it is necessary to develop advanced co-optimization methods that are conducive to improving vehicle efficiency, reducing costs, and further developing and large-scale commercialization of FCEVs.

[0004] At present, the coordinated optimization of vehicle speed planning and energy management mostly adopts multi-level hierarchical control. Hierarchical control can reduce the computational cost, but the coupling relationship between vehicle speed and power distribution control is not properly handled during the optimization process. If the vehicle-level and powertrain-level control (vehicle speed and energy management) can be optimized at the same time, the energy utilization efficiency and the dynamic performance of the system can be improved. Therefore, the optimal control strategy in energy management plays an important role in the performance and cost reduction of the whole vehicle, and the integrated real-time optimization control of vehicle speed and energy management has positive practical significance.

[0005] At present, most of the research on the coordinated optimization of vehicle speed planning and energy management is concentrated on fuel vehicles and pure electric vehicles or hybrid vehicles driven by fuel and batteries. There are relatively few studies on more environmentally friendly and efficient fuel cell hybrid vehicles. The four have differences in topological structure and energy system operation, and the ways of dealing with optimization problems are also different. Therefore, it is necessary to study the coordinated optimization problem of vehicle speed planning and energy management of fuel cell hybrid vehicles.

[0006] Traditionally, speed planning methods and EMS methods (through convex programming, Pontryagin minimum principle (PMP) and dynamic programming (DP)) express the control problem of vehicle speed or powertrain EMS as an optimization problem for a given or known driving environment. However, the vehicle cannot obtain all driving environment information in advance during actual movement. The speed planning and powertrain EMS results generated by global optimization may not be feasible in the current actual driving environment and are only suitable for reference comparison and cannot be applied in practice. Therefore, the flexibility and adaptability of these methods are limited, and the accuracy and generalization ability of these methods need to be further improved.

[0007] At present, the coordinated optimization of vehicle speed planning and energy management mostly adopts multi-level hierarchical control. Hierarchical control can reduce the computational cost, but the coupling relationship between vehicle speed and power distribution control is not properly handled during the optimization process. The control of the powertrain should not be limited to the energy management level. All control degrees of freedom in the powertrain will affect the speed optimization in ecological driving. The optimal speed planning result given by the upper layer may not be feasible or the effect is poor (such as insufficient power supply in the lower energy management), which cannot guarantee the best economy in the entire journey.

[0008] At present, the coordinated optimization of vehicle speed and energy management has the problems of high computational complexity and low computational efficiency, which makes the optimization algorithm unable to be actually applied in real scenarios. Therefore, it is necessary to solve the real-time calculation problem of the optimization algorithm. Summary of the invention

[0009] In view of the above shortcomings, the present invention proposes a speed-energy integrated optimization control method for intelligent fuel cell vehicles, provides an energy management control strategy under dynamic traffic scenarios, improves the energy management optimization potential and control strategy to comprehensively consider the multi-objective optimization problem of the economy and comfort of fuel cell vehicles.

[0010] To achieve the above object, the present invention provides the following technical solutions:

[0011] Terminology explanation: Some of the English abbreviations involved in this invention are explained as follows:

[0012] V2V, vehicle-to-vehicle communication.

[0013] V2I, vehicle-to-infrastructure communication.

[0014] ELM, Extreme Learning Machine.

[0015] MPC, Model Predictive Control.

[0016] PMP, Pontryagin's Minimum Principle.

[0017] FCEV, Fuel Cell Electric Vehicle.

[0018] EMS, Energy Management System.

[0019] DCDC stands for Direct Current to Direct Current Convertor.

[0020] A speed-energy integrated optimization control method for an intelligent fuel cell vehicle, the method comprising the following steps:

[0021] By interacting with the surrounding environment (V2I) and surrounding vehicles (V2V), the traffic information on the road is obtained during the driving process of the main vehicle, and the speed trajectory of the leading vehicle in the future is predicted based on the ELM algorithm;

[0022] Taking the economy and comfort of fuel cell vehicles as optimization goals, the main vehicle speed planning and energy management strategy are integrated into an integrated architecture, and the PMP principle is used to solve the optimal control problem based on the MPC framework.

[0023] As a further technical solution of the present invention, the step of predicting the speed trajectory of the preceding vehicle at a future time based on the ELM algorithm includes:

[0024] During driving, vehicles often encounter situations where there are vehicles ahead. The driving behavior of the vehicle ahead has a very important impact on the driving of the vehicle. The vehicle plans its own speed under the premise of considering the behavior of the vehicle ahead, and uses the speed of the vehicle ahead in the future to guide the driving behavior of the vehicle, so as to better achieve the economical and energy-saving driving of the vehicle. The present invention adopts the ELM algorithm to predict the speed of the vehicle ahead, and provides a basis for the formulation of the vehicle speed-energy integrated control strategy.

[0025] The present invention obtains traffic information on the road during the driving process of the main vehicle through V2V and V2I, and obtains the speed of the preceding vehicle according to the output of the preceding vehicle speed prediction model, ignoring the problems such as communication delay and instrument error caused by the two technologies. The training of the ELM algorithm is a process of inputting enough samples into the network and adjusting the weights of each node of the network using historical data to make the output value of the model consistent with the expected value. Based on the above analysis, the input data of the preceding vehicle speed prediction model in the present invention are the preceding vehicle historical speed, the preceding vehicle historical acceleration, and the speed limit and traffic information of the traffic section (X=[ v p ( k-r ), a p ( k-r ),…, v lim ( k-r ), q ( k-r )], r= 1 …N r , N r is the historical time step length), and the output data is the speed of the preceding vehicle in the predicted time domain ( , , N p is the prediction time domain step length).

[0026] As a further technical solution of the present invention, the training set of the preceding vehicle speed prediction model comes from actual speed data collected on regular routes, and the gap between the predicted value and the actual value is compared through the verification set. If the gap is within a preset range, a trained preceding vehicle speed prediction model is obtained; the trained preceding vehicle speed prediction model predicts the preceding vehicle speed based on network information and uses it as a dynamic influencing factor on the driving of the vehicle, and is used in the speed-energy integrated control strategy of the vehicle.

[0027] As a further technical solution of the present invention, considering the changes in surrounding environment information and driving data and the problem of computing load in the future, the speed prediction part of the leading vehicle is allocated to a core of the multi-core chip (denoted as CPU0), and the surrounding environment information and vehicle driving data during the vehicle driving process are stored, and added to the ELM algorithm model as a new data set to update the parameters of the ELM algorithm model; in particular, this process is not a real-time update, but an operation is performed in a long time domain after storing a certain data set. After the model is changed, the new speed prediction model is used to predict the leading vehicle until the next training and updating of the leading vehicle speed prediction model.

[0028] As a further technical solution of the present invention, the steps of taking the economy and comfort of fuel cell vehicles as optimization goals, integrating the main vehicle speed planning and energy management strategy into an integrated architecture, and solving the optimal control problem based on the MPC framework and the PMP principle include:

[0029] By using the vehicle network to obtain the surrounding traffic information and the predicted speed data of the preceding vehicle obtained by the ELM algorithm, the fuel cell vehicle speed-energy integrated optimization problem is established based on the MPC framework and the PMP principle, and the optimal control sequence that meets the system constraints is obtained. The goal of the fuel cell vehicle speed-energy integrated optimization problem is to find the optimal control sequence that satisfies the objective function. J Minimum optimal control sequence u* ( k ), the present invention is based on the fuel cell vehicle economy R EMS and comfort R c To optimize the target, the vehicle speed v h and power batteries SOC is the state variable, with the acceleration of the main vehicle a h and fuel cell power P fc As the control variable, solving the optimal control problem under the constraints is described as follows:

[0030] ;

[0031] is the hydrogen consumption of the fuel cell, which is determined by the fuel cell power P fc Functional representation of ; k is the total travel discrete node; N P To predict the time domain step size, ; w is the weight coefficient, which is used to adjust the relative importance of economy and comfort and needs to be determined through simulation experiments; a hmin is the minimum acceleration allowed, a hmax is the maximum acceleration allowed;

[0032] The state equation is described as follows:

[0033] Formula 2: ;

[0034] U oc is the open circuit voltage of the power battery, which is SOC Functional representation of ; R b is the internal resistance of the power battery, which is also composed of SOC Functional representation of ; Q bat is the battery capacity of the power battery; Δt is a single time step ；P dem is the required power of the main vehicle, which can be calculated by formula 4;

[0035] The constraints are described as follows:

[0036]

[0037] d is the distance between the main vehicle and the vehicle in front, and the specific expression is shown in Formula 4; d min is the minimum clearance to avoid collision, and its specific value is calculated by formula 4; d max It is the biggest gap in ensuring traffic efficiency; SOC low , SOC high They are SOC The upper and lower limits allowed; P fcmin , P fcmax are the upper and lower limits of the fuel cell power; P bmin , P bmax is the upper and lower limits of the power of the power battery, Δ P fc Constraints on fuel cell power fluctuations: ,This value is based on the fuel cell life consideration; d 0 is the initial distance to the vehicle ahead; v 0 , SOC 0 , SOC tf are the initial conditions and terminal values ​​of the main vehicle;

[0038] Formula 4: ;

[0039] v p is the speed of the preceding vehicle; t h0 is the time interval; s 0 is the static clearance; m veh The quality of the main vehicle; f r is the friction coefficient; g is the acceleration due to gravity;α is the slope angle; C D is the drag coefficient; A f is the windward area; ρ a is the air density; δ is the rotation mass conversion factor;

[0040] According to the PMP principle, the necessary condition for solving the optimal control problem is , introduce Hamiltonian H and define it as:

[0041] Formula 5: ;

[0042] Then, the necessary conditions for the vehicle system to achieve optimal power distribution and optimal vehicle speed at the same time are as follows:

[0043]

[0044] p 1. p 2 is the covariate variable of the Hamiltonian function, which needs to provide an initial value and is determined according to the third and fourth necessary conditions of formula 6;

[0045] Hamiltonian H with respect to two control variables a h and P fc It has a convex structure, which indicates that the optimal combination of two control variables can be selected in the calculation time step to minimize the Hamiltonian; according to the state equation and variable equation in formula 6, the selected control variable will affect the state variable and co-state variable of the next time step; a similar process will be repeated in the next prediction time step and control time step to obtain the trajectory of the parameters in the prediction period. It is not easy to obtain an analytical solution by using the PMP algorithm to solve the optimal control problem. It can be numerically solved by an iterative method. The present invention chooses to use a shooting method for iterative solution;

[0046] Based on the MPC framework, the present invention adopts the PMP principle to solve the optimal control problem as a set of instantaneous optimization. T p Within each forecast range Np ,in T p = N p Δt , the control is solved to minimize the cumulative energy consumption; however, due to the evolution of the disturbance, only the control strategy T c will be executed, where Tc = N c Δt c , N c To control the step size, Δt c is the single-step control time. After effective control is performed, the prediction range moves forward and another instantaneous optimization begins, thereby achieving the best power distribution and the best vehicle speed;

[0047] In this stage, the integrated energy management algorithm is allocated to another core of the multi-core chip (denoted as CPU1) to significantly reduce the computational burden of the control system and enhance the ability to process information at different time scales.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] 1. Energy management control strategy in dynamic traffic scenarios.

[0050] Compared with energy management under deterministic driving conditions, the uncertainty of the traffic environment is taken into account, and the vehicle speed planning and energy management strategy are performed under dynamic traffic conditions based on the changes in the speed and traffic information of the preceding vehicle. The accuracy of the preceding vehicle speed prediction model is improved and the computational load of the optimization algorithm is reduced by distributing the dynamic traffic environment (preceding vehicle speed prediction) processing and the integrated energy management algorithm to different cores of the multi-core chip.

[0051] 2. Enhance the potential for energy management optimization.

[0052] Compared with hierarchical optimization energy management, it fully considers the optimality of the optimization algorithm, can better reveal the coupling relationship between the vehicle level and the system power level, construct external adaptive mechanism (between the vehicle and the driving environment) and internal adaptive mechanism (between the vehicle and the powertrain), and further enhance the energy management optimization potential.

[0053] 3. The control strategy comprehensively considers the multi-objective optimization problem of fuel cell vehicle economy and comfort.

[0054] Taking into full consideration the mutual constraints between speed planning and powertrain control, taking the economy and comfort of fuel cell vehicles as the optimization goals, constructing a multi-objective optimization problem and determining the constraints, integrating the optimization control of vehicle speed planning and energy management, and reasonably allocating the power between various energy sources, the optimality of the strategy is achieved while ensuring the real-time performance of the strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A diagram of the powertrain of a fuel cell vehicle provided in an embodiment of the present invention.

[0056] Figure 2The overall flow chart of vehicle speed-energy integrated optimization control provided by an embodiment of the present invention.

[0057] Figure 3 A diagram of the network structure for predicting the speed of a preceding vehicle provided in an embodiment of the present invention.

[0058] Figure 4 A control strategy flow chart provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0059] The technical solution of this patent is further described in detail below in conjunction with specific implementation methods.

[0060] See also Figures 1 to 4 , a speed-energy integrated optimization control method for an intelligent fuel cell vehicle, the method comprising the following steps:

[0061] By interacting with the surrounding environment (V2I) and surrounding vehicles (V2V), the traffic information on the road is obtained during the driving process of the main vehicle, and the speed trajectory of the leading vehicle in the future is predicted based on the ELM algorithm;

[0062] Taking the economy and comfort of fuel cell vehicles as optimization goals, the main vehicle speed planning and energy management strategy are integrated into an integrated architecture, and the PMP principle is used to solve the optimal control problem based on the MPC framework.

[0063] This embodiment realizes the coordinated optimization control of fuel cell vehicles, improves the vehicle economy on the basis of ensuring driving safety and comfort, and meets the requirements of real-time applications. In addition, in order to improve the online computing efficiency of the control system, the present invention uses a multi-core chip controller for automotive standards and proposes a control architecture based on multi-core parallel computing, that is, the front vehicle speed prediction part is allocated to one core of the multi-core chip (represented as CPU0), and the integrated energy management algorithm is allocated to another core of the multi-core chip (represented as CPU1), so as to significantly reduce the computing burden of the control system and enhance the ability to process information at different time scales.

[0064] Figure 1 It refers to the powertrain of a hydrogen fuel cell vehicle, which is mainly composed of a hydrogen storage tank, a fuel cell, a power battery, a drive motor, a differential, a DCDC, etc.

[0065] Figure 2 It shows the overall flow chart of vehicle speed-energy integrated optimization control, which specifically includes the following contents:

[0066] The steps of predicting the speed trajectory of the preceding vehicle at a future time based on the ELM algorithm include:

[0067] During driving, vehicles often encounter situations where there are vehicles ahead. The driving behavior of the vehicle ahead has a very important impact on the driving of the vehicle. The vehicle plans its own speed under the premise of considering the behavior of the vehicle ahead, and uses the speed of the vehicle ahead in the future to guide the driving behavior of the vehicle, so as to better achieve the economical and energy-saving driving of the vehicle. The present invention adopts the ELM algorithm to predict the speed of the vehicle ahead, and provides a basis for the formulation of the vehicle speed-energy integrated control strategy.

[0068] The present invention obtains traffic information on the road during the driving process of the main vehicle through V2V and V2I, and obtains the speed of the preceding vehicle according to the output of the preceding vehicle speed prediction model, ignoring the problems such as communication delay and instrument error caused by the two technologies. The training of the ELM algorithm is a process of inputting enough samples into the network and adjusting the weights of each node of the network using historical data to make the output value of the model consistent with the expected value. Based on the above analysis, the input data of the preceding vehicle speed prediction model in the present invention are the preceding vehicle historical speed, the preceding vehicle historical acceleration, and the speed limit and traffic information of the traffic section ( is the historical time step length), and the output data is the speed of the preceding vehicle in the predicted time domain ( , , N p is the prediction time domain step length).

[0069] In this embodiment, Figure 3 This is the network structure diagram of the preceding vehicle speed prediction. The training set of the preceding vehicle speed prediction model comes from the actual speed data collected on regular routes, and the gap between the predicted value and the actual value is compared through the validation set. If the gap is within the preset range, the trained preceding vehicle speed prediction model is obtained; the trained preceding vehicle speed prediction model predicts the preceding vehicle speed according to the network information and uses it as a dynamic influencing factor on the driving of the vehicle, and is used in the speed-energy integrated control strategy of the vehicle.

[0070] In this embodiment, considering the changes in surrounding environment information and driving data and the problem of computing load in the future, the front vehicle speed prediction part is allocated to a core of the multi-core chip (denoted as CPU0), and the surrounding environment information and vehicle driving data during the vehicle driving process are stored, and added to the ELM algorithm model as a new data set to update the parameters of the ELM algorithm model; in particular, this process is not a real-time update, but an operation is performed in a long time domain after storing a certain data set. After the model is changed, the new speed prediction model is used to predict the front vehicle until the next training and updating of the front vehicle speed prediction model.

[0071] In this embodiment, the present invention utilizes V2V and V2I technologies to achieve information interaction with the surrounding environment and vehicles, obtains the driving status information of the leading vehicle (the current speed, acceleration, and historical driving data of the leading vehicle), traffic information (road traffic speed limit and traffic flow density information), and utilizes the ELM algorithm to predict the speed trajectory of the leading vehicle at future moments; in particular, the leading vehicle speed prediction part is allocated to one core (CPU0) of the multi-core processor to optimize computing efficiency, and the leading vehicle speed prediction model is regularly trained and updated to further improve its accuracy and adaptability.

[0072] The steps of taking the fuel cell vehicle economy and comfort as optimization goals, integrating the main vehicle speed planning and energy management strategy into an integrated architecture, and solving the optimal control problem based on the MPC framework and the PMP principle include:

[0073] By using the vehicle network to obtain the surrounding traffic information and the predicted speed data of the preceding vehicle obtained by the ELM algorithm, the fuel cell vehicle speed-energy integrated optimization problem is established based on the MPC framework and the PMP principle, and the optimal control sequence that meets the system constraints is obtained. The goal of the fuel cell vehicle speed-energy integrated optimization problem is to find the optimal control sequence that satisfies the objective function. J Minimum optimal control sequence u* ( k ), the present invention is based on the fuel cell vehicle economy R EMS and comfort R c To optimize the target, the vehicle speed v h and power batteries SOC is the state variable, with the acceleration of the main vehicle a h and fuel cell power P fc As the control variable, solving the optimal control problem under the constraints is described as follows:

[0074]

[0075] is the hydrogen consumption of the fuel cell, which is determined by the fuel cell power P fc Functional representation of ; k is the total travel discrete node; N P To predict the time domain step size, ; w is the weight coefficient, which is used to adjust the relative importance of economy and comfort and needs to be determined through simulation experiments; a hmin is the minimum acceleration allowed, a hmaxis the maximum acceleration allowed;

[0076] The state equation is described as follows:

[0077] Formula 2: ;

[0078] U oc is the open circuit voltage of the power battery, which is SOC Functional representation of ; R b is the internal resistance of the power battery, which is also composed of SOC Functional representation of ; Q bat is the battery capacity of the power battery; Δt is the single time step; P dem is the required power of the main vehicle, which can be calculated by formula 4;

[0079] The constraints are described as follows:

[0080]

[0081] d is the distance between the main vehicle and the vehicle in front, and the specific expression is shown in Formula 4; d min is the minimum clearance to avoid collision, and its specific value is calculated by formula 4; d max It is the biggest gap in ensuring traffic efficiency; SOC low , SOC high They are SOC The upper and lower limits allowed; P fcmin , P fcmax are the upper and lower limits of the fuel cell power; P bmin , P bmax is the upper and lower limits of the power of the power battery, Δ P fc Constraints on fuel cell power fluctuations: ,This value is based on the fuel cell life consideration; d 0 is the initial distance to the vehicle ahead; v 0 , SOC 0 , SOC tf are the initial conditions and terminal values ​​of the main vehicle;

[0082] Formula 4:

[0083] v p is the speed of the preceding vehicle; t h0 is the time interval; s 0 is the static clearance; m veh The quality of the main vehicle; f r is the friction coefficient; g is the acceleration due to gravity; α is the slope angle; C D is the drag coefficient; A f is the windward area; ρ a is the air density; δ is the rotation mass conversion factor;

[0084] According to the PMP principle, the necessary condition for solving the optimal control problem is , introduce Hamiltonian H and define it as:

[0085] Formula 5:

[0086] ;

[0087] Then, the necessary conditions for the vehicle system to achieve optimal power distribution and optimal vehicle speed at the same time are as follows:

[0088]

[0089] p 1. p 2 is the covariate variable of the Hamiltonian function, which needs to provide an initial value and is determined according to the third and fourth necessary conditions of formula 6;

[0090] Hamiltonian H with respect to two control variables a h and P fc It has a convex structure, which indicates that the optimal combination of two control variables can be selected in the calculation time step to minimize the Hamiltonian; according to the state equation and variable equation in Formula 6, the selected control variable will affect the state variable and co-state variable of the next time step; a similar process will be repeated in the next prediction time step and control time step to obtain the trajectory of the parameters in the prediction period. It is not easy to obtain an analytical solution by using the PMP algorithm to solve the optimal control problem. It can be numerically solved by an iterative method. The present invention chooses to use the shooting method for iterative solution. The specific flow chart of the vehicle speed-energy integrated optimization control strategy is as follows Figure 4 As shown;

[0091] Based on the MPC framework, the present invention adopts the PMP principle to solve the optimal control problem as a set of instantaneous optimization. T p Within each prediction horizon Np, T p = N p Δt , the control is solved to minimize the cumulative energy consumption; however, due to the evolution of the disturbance, only the control strategy T c will be executed, where T c = N c Δt c , N c To control the step size, Δt c is the single-step control time. After effective control is performed, the prediction range moves forward and another instantaneous optimization begins, thereby achieving the best power distribution and the best vehicle speed;

[0092] In this stage, the integrated energy management algorithm is allocated to another core of the multi-core chip (denoted as CPU1) to significantly reduce the computational burden of the control system and enhance the ability to process information at different time scales.

[0093] In this embodiment, the present invention establishes a vehicle dynamics model based on the collected traffic information and the driving status information of the preceding vehicle and the vehicle, integrates the main vehicle speed planning and energy management strategy into an integrated architecture, considers the mutual constraints between speed planning and powertrain control, and can construct an external adaptive mechanism (between the vehicle and the driving environment) and an internal adaptive mechanism (between the vehicle and the powertrain), takes the economy and comfort of fuel cell vehicles as the optimization goals, constructs a multi-objective optimization problem and determines the constraints, solves the optimization problem based on the MPC algorithm framework and the PMP principle, and obtains the optimal vehicle speed planning and the optimal power allocation of each energy. In particular, the integrated energy management algorithm is allocated to another core (CPU1) of the multi-core processor, thereby greatly reducing the computational burden of the control system and improving the ability to process information at different time scales.

[0094] It should be noted that, in the present invention, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0095] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A speed-energy integrated optimization control method for an intelligent fuel cell vehicle, characterized in that: The method comprises the following steps: By interacting with the surrounding environment and surrounding vehicles, the traffic information on the road is obtained during the driving process of the main vehicle, and the speed trajectory of the leading vehicle in the future is predicted based on the ELM algorithm; Taking the economy and comfort of fuel cell vehicles as the optimization goals, the main vehicle speed planning and energy management strategy are integrated into an integrated architecture, and the PMP principle is used based on the MPC framework to solve the optimal control problem; Using an automotive-grade multi-core chip controller, a control architecture based on multi-core parallel computing is proposed, which allocates the front vehicle speed prediction part to one core of the multi-core chip and allocates the integrated energy management algorithm to another core of the multi-core chip. The steps of taking the fuel cell vehicle economy and comfort as optimization goals, integrating the main vehicle speed planning and energy management strategy into an integrated architecture, and solving the optimal control problem based on the MPC framework and the PMP principle include: Based on the MPC framework, the PMP principle is used to establish the fuel cell vehicle speed-energy integrated optimization problem, and the optimal control sequence that meets the system constraints is obtained. The goal of the fuel cell vehicle speed-energy integrated optimization problem is to find the optimal control sequence u*(k) that minimizes the objective function J, and to optimize the fuel cell vehicle economy R EMS and comfort R c As the optimization target, the vehicle speed v h and the power battery SOC as state variables, and the main vehicle acceleration a h and fuel cell power P fc As the control variable, solving the optimal control problem under the constraints is described as follows: J=min(R c +w R EMS ) Formula 1: Where: m H2 is the hydrogen consumption of the fuel cell, which is determined by the fuel cell power P fc Function representation; k is the total travel discrete node; N P is the prediction time domain step size, f = 1…N P ; w is the weight coefficient, which is used to adjust the relative importance of economy and comfort, a hmin is the minimum acceleration allowed, a hmax is the maximum acceleration allowed; The state equation is described as follows: Formula 2: U oc is the open circuit voltage of the power battery, which is characterized by the function of SOC; R b is the internal resistance of the power battery, which is also characterized by the function of SOC; Q bat is the battery capacity of the power battery; Δt is the single-step time step; P dem is the required power of the main vehicle, which can be calculated by formula 4; The constraints are described as follows: Formula 3: d is the distance between the main vehicle and the vehicle in front, and its specific expression is shown in Equation 4; d min is the minimum clearance to avoid collision, and its specific value is calculated by formula 4; d max Is the biggest gap in ensuring traffic efficiency; SOC low , SOC high are the upper and lower limits of SOC respectively; P fcmin , P fcmax is the upper and lower limits of the fuel cell power; P bmin , P bmax is the upper and lower limits of the power of the power battery; ΔP fc The fuel cell power fluctuation constraint: ΔP fc (k+f)=P fc (k+f)-P fc (k+f-1), this value is based on the fuel cell life; d0 is the initial distance from the vehicle in front; v0, SOC0, SOC tf are the initial conditions and terminal values ​​of the main vehicle; Formula 4: v p is the speed of the preceding vehicle; t h0 is the time interval; s0 is the static gap; m veh is the mass of the main vehicle; f r is the friction coefficient; α is the slope angle; g is the acceleration due to gravity; C D A is the drag coefficient; f is the frontal area; ρ a is the air density; δ is the rotational mass conversion factor; According to the PMP principle, the necessary condition for solving the optimal control problem is δJ = 0. If the Hamiltonian H is introduced, it is defined as: Formula 5: Then, the necessary conditions for the vehicle system to achieve optimal power distribution and optimal vehicle speed at the same time are as follows: Formula 6: p1 and p2 are covariates of the Hamiltonian function; Hamiltonian H with respect to two control variables a h and P fc It has a convex structure and is solved iteratively using the shooting method; Based on the MPC framework, the PMP principle is used to solve the optimal control problem as a set of instantaneous optimizations. p Within each prediction range Np, where T p =N p Δt, after solving, minimizes the cumulative energy consumption; however, due to the evolution of disturbances, only the control strategy in the control period T c will be executed, where T c =N c Δt c , N c To control the step size, Δt c For single-step control time, after effective control is performed, the prediction horizon moves forward and another instantaneous optimization begins, thereby achieving the best power distribution and the best vehicle speed.

2. The intelligent fuel cell vehicle speed-energy integrated optimization control method according to claim 1 is characterized in that: The step of predicting the speed trajectory of the preceding vehicle at a future time based on the ELM algorithm comprises: The traffic information on the road is obtained during the driving process of the main vehicle, and the speed of the preceding vehicle is obtained according to the output of the preceding vehicle speed prediction model. The input data of the preceding vehicle speed prediction model are the preceding vehicle's historical speed, the preceding vehicle's historical acceleration, the speed limit of the traffic section, and the traffic information. The output data is the preceding vehicle's speed in the predicted time domain.

3. The intelligent fuel cell vehicle speed-energy integrated optimization control method according to claim 2 is characterized in that: The training set of the preceding vehicle speed prediction model comes from the actual speed data collected on regular routes, and the gap between the predicted value and the actual value is compared through the validation set. If the gap is within the preset range, a trained preceding vehicle speed prediction model is obtained; the trained preceding vehicle speed prediction model predicts the preceding vehicle speed based on the network information and uses it as a dynamic influencing factor on the driving of the vehicle, and is used in the vehicle speed-energy integrated control strategy of the vehicle.

4. The intelligent fuel cell vehicle speed-energy integrated optimization control method according to claim 1 is characterized in that: The present invention proposes a control architecture based on multi-core parallel computing by using a multi-core chip controller for automotive grade, that is, the steps of allocating the front vehicle speed prediction part to one core of the multi-core chip and allocating the integrated energy management algorithm to another core of the multi-core chip include: The leading vehicle speed prediction part is allocated to one core of the multi-core chip, and the surrounding environment information and vehicle driving data during the vehicle driving process are stored, and added to the ELM algorithm model as a new data set to update the parameters of the ELM algorithm model; after the model is changed, the new speed prediction model is used to predict the leading vehicle until the next training and update of the leading vehicle speed prediction model, and the integrated energy management algorithm is allocated to another core of the multi-core chip. The multi-core CPUs share hardware resources to ensure data communication.

Images