| The traditional internal combustion engine vehicles have low efficiency and high pollution,and the pure electric vehicles have long charging time and poor environmental adaptability.Therefore,due to the advantages of long driving range,short refuelling time and no running pollution,fuel cell hybrid electric vehicles are expected to play a key role in the future clean energy transportation and alleviate the worldwide energy and environmental problems to a certain extent.Proton exchange membrane fuel cell(PEMFC)has become the first choice for automotive fuel cells owing to the high energy density and low temperature.At present,the powertrain system of fuel cell vehicles generally driven by two power sources,fuel cell and lithium battery.How to reasonably distribute energy between them so that the powertrain can meet the driving needs of vehicles and has superior economy and durability is very important.Affected by the system composition,working principle and technical limitations,the remaining useful life(RUL)of the fuel cell changes dramatically under on-board conditions.At the same time,due to the high cost of fuel cell system,only considering the economy of powertrain can no longer meet the needs of multiple performance optimizations.Therefore,a control strategy of powertrain considering the state of health(SOH)is proposed to solve the multi-objective optimization problem of complex systems under variable states.At present,there are few powertrain control strategies considering SOH in the literature,and there are still two problems.Firstly,the SOH of the PEMFC can not be correctly estimated on-board,which makes it impossible to transmit SOH to the controller as an input,which makes the optimal strategy solved by the controller unable to achieve optimal performance due to the deviation between the current state and the initial state of the system.Besides,the strategy can not optimize the SOH of the system,so as to prolong the RUL.Secondly,the global optimal results of the constructed strategy cannot match the complex and variable conditions,which will lead to the deviation of the strategy from the optimal power distribution point under the real vehicle conditions,and the optimization goal cannot be achieved.In order to solve the above problems,this paper establishes a set of adaptive energy management strategy(EMS)based on SOH online recognizer.The main research work and innovations of this paper are as follows:(1)Considering the difficulty of on-line SOH estimation,this paper analyzes the change mechanism of key components SOH of the powertrain,and constructs a performance degradation causal chain coupled with operating conditions and components,extracts the health indicators of each component,and forms a relatively complete SOH representation system of the powertrain with a double-layer structure of internal characteristics and output performance.(2)A SOH monitoring system is established for the two key power sources,fuel cell and battery,and an online SOH identifier is formed.To evaluate the SOH of PEMFCs,SOH is divided into two parts: the output characteristic characterized by voltage and the internal hydration characteristic characterized by impedance.For voltage estimation,the genetic algorithm is used to identify the parameters of the fitting model,and then the Kalman filter is used to construct the voltage attenuation estimation algorithm.Then,the environmental factor and its online update strategy are proposed for the voltage change under unsteady state conditions,and an online recognizer of fuel cell SOH can be quantitatively characterized in real time.For the impedance data,the Randles model is used to fit the impedance spectrum data after data preprocessing,and then a hydration state judgment model based on linear discriminant method is constructed as an offline supplement to the fuel cell SOH recognizer.An on-line SOH recognizer considering calendar aging and cycle aging of battery is constructed,and the SOH is represented by capacity change.(3)The fuel cell vehicle powertrain system used in the project is the research object.Based on MATLAB/Simulink,the mechanical model and electrical model of the vehicle are established.The model includes fuel cell,battery,motor,DC/DC,vehicle controller and driver,in which the focus is on modeling the fuel cell system.In this paper,the commonly used nine-order dynamic model is improved.Through the simulation results,it is verified that the model can characterize the dynamic response capability of fuel cells under different power,which provides a platform for the subsequent verification of EMSs.(4)To solve the problem that SOH is not considered in the construction of the EMS,two sets of algorithms are proposed in this paper,namely,the global optimal EMS considering SOH and the instantaneous optimal EMS considering SOH.Optimization methods are proposed for the current system state,that is,the strategy is modified according to the dynamic response capability of PEMFC,and the system degradation after SOH changes.Then,a comprehensive cost calculation function is constructed to replace the equivalent hydrogen consumption as the optimization objective.Results show that the deviation between the required and actual output power of PEMFC is smaller,so the system economy is improved.The control strategy considering performance degradation is also more economical.At the same time,considering SOH,the comprehensive use cost of the system is greatly reduced.(5)To solve the problem that the global optimal results can not match the complex working conditions,this paper adds an online driving pattern recognizer before the prediction,different from the method of converting the global dynamic programming into the stochastic dynamic programming by using the working condition prediction in the literature.Firstly,the characteristic parameters of working conditions are reduced by principal component analysis,and then the driving pattern are divided into urban,suburban and high-speed modes by hierarchical clustering.The learning vector quantization neural network is introduced to build an online driving pattern recognizer according to the typical driving cycle data.Then,according to the identified patterns,different adaptive control strategies are designed for each pattern,so that the controller can automatically switch to the characteristic optimization strategy under specific driving cycles,which includes the stochastic dynamic programming of predictive control and the instantaneous optimization strategy considering SOH.By comparing the results with three rule-based EMSs,it is verified that the optimization strategy established in this paper can effectively reduce the comprehensive operation cost and achieve the purpose of optimizing the SOH and economy.In general,based on the analysis of the causal chain of SOH coupled with the key components of the powertrain and the working conditions,the SOH representation system is proposed,and then the online SOH identifier is constructed to provide the quantified SOH for the controller.Then,according to the change of SOH and the dynamic response ability of fuel cell,an adaptive EMS suitable for complex and variable driving conditions is designed.This EMS can achieve the optimal comprehensive use cost,which comprehensive represents of economy and SOH.This set of optimal control flow considering SOH based on data acquisition and preprocessing,parameter identification and state estimation has certain application value. |
PDF Full Text Request