Energy management method of vehicle hybrid power system, electronic equipment and medium

Abstract

The invention provides an energy management method of a vehicle hybrid power system, electronic equipment and a medium. The energy management method of the vehicle hybrid power system comprises the steps that a vehicle hybrid power system model is established; establishing a target function with the minimum comprehensive value loss cost; power distribution of different power sources of the hybrid power system of the vehicle under different operation conditions is realized by utilizing a deep learning network model and combining a vehicle hybrid power system model, a target function, an actual operation condition and the like; according to the invention, the fuel economy and the power variable load influence of the hybrid power system are considered, and the output power among different power sources is better distributed in different driving processes, so that better fuel economy and durability are obtained. According to the hybrid power system, the adaptability of energy management under different operation conditions of the vehicle is high, so that different road conditions are dealt with.

Description

Technical Field

[0001] The present invention belongs to the technical field of new energy vehicles, and in particular relates to an energy management method, electronic equipment and medium for a vehicle hybrid power system. Background Art

[0002] Rail transit industry is a pillar industry in my country, but it also causes a lot of pollution. In order to deal with this problem, the application of fuel cells in the field of rail transit has become a current research hotspot. Fuel cells have the advantages of zero pollution and high efficiency, but their variable load energy is relatively weak, and they are often equipped with lithium batteries or capacitors to assist in energy supply to form a hybrid power system. It can not only make up for the lack of dynamic response of fuel cells, but also recover braking energy and improve the economic efficiency of train operation.

[0003] However, the price and service life of hybrid systems are important factors that restrict the commercialization of fuel cells, and fast-changing power loads often cause irreversible damage to hybrid systems. Therefore, reasonable power distribution can greatly increase the service life of hybrid systems.

[0004] At present, the energy management method of hybrid power system is limited to reducing the fuel consumption of the system, ignoring the harm of power load variation to the fuel cell system. In addition, when the vehicle runs on different routes, it often needs to be redesigned and calculated, which has poor adaptability. Summary of the invention

[0005] The purpose of the present invention is to address the deficiencies in the prior art and to provide an energy management method, electronic equipment and medium for a vehicle hybrid power system, wherein the energy management of the hybrid power system has high adaptability under different working conditions.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is:

[0007] A method for energy management of a vehicle hybrid power system comprises the following steps:

[0008] S1: Establishing a vehicle hybrid power system model, wherein the vehicle hybrid power system model includes a fuel cell degradation model and a lithium battery degradation model;

[0009] The fuel cell degradation model is as follows:

[0010]

[0011] Among them, m deg,fc is the performance degradation of the fuel cell, D high , D low With D change are the fuel cell losses under high power, low power and power variable load conditions, P cap,fcis the rated power of the fuel cell, U fcnor is the rated voltage of the fuel cell, U EOL is the voltage of the fuel cell when it is retired;

[0012] The lithium battery degradation model is as follows:

[0013]

[0014] Among them, m deg,bat is the performance degradation of lithium battery, i bat , U bat are the output current and voltage of the lithium battery respectively, and N is the number of charge and discharge cycles of the lithium battery.

[0015] S2: Establish an objective function based on hydrogen cost, fuel cell loss, lithium battery loss, average daily operating cost, and lithium battery charge SOC offset. The objective function is as follows:

[0016]

[0017] Where J is the total operating cost of the hybrid system, T is the operating time of the hybrid system, and J soc is the SOC offset of the lithium battery, J energy is the hydrogen consumption cost of the hybrid system, J fc is the fuel cell degradation cost, J bat is the degradation cost of lithium battery, J d is the average daily operating cost, k1, k2, k3, k4 are cost coefficients;

[0018] S3: inputting a standard operating condition into the hybrid power system model of the vehicle to obtain a hybrid power system output power curve under different conditions; the standard operating condition includes an operating speed curve, a required power curve, and a SOC curve when the vehicle is designed to operate;

[0019] S4: using the operating speed curve, the required power curve, the SOC curve, and the hybrid power system output power curve of the vehicle under the designed operating condition as inputs of the deep learning network model, and training the deep learning network model;

[0020] Iteratively optimizing and training the deep learning network model using the objective function to obtain a hybrid power system power distribution model;

[0021] S5: Inputting the real-time operation data of the vehicle into the power distribution model of the hybrid power system to obtain the output power of the fuel cell and the output power of the lithium battery in the hybrid power system, wherein the real-time operation data of the vehicle includes the operation speed curve, the demand power curve, and the SOC curve under the actual operation condition.

[0022] The present invention utilizes a deep learning network, combines the vehicle hybrid system model, the objective function of minimizing the comprehensive value loss cost, and the actual operating conditions, etc., to achieve the power allocation of the fuel cell system and the lithium battery system under different vehicle operating conditions; the present invention takes into account the fuel economy and the influence of the hybrid system power load variation, and better allocates the output power between different power sources during different driving processes to obtain better fuel economy and durability. The hybrid system energy management of the present invention has high adaptability under different vehicle operating conditions, so as to cope with different road conditions.

[0023] Furthermore, the fuel cell consumption is hydrogen consumption, and the instantaneous hydrogen consumption of the fuel cell is as follows:

[0024] C fc =0.01954P fc -0.31924

[0025] Among them, C fc is the instantaneous hydrogen consumption of the fuel cell, P fc is the fuel cell power.

[0026] Furthermore, the loss of lithium batteries is equivalent to hydrogen consumption, and the instantaneous hydrogen consumption of lithium batteries is as follows:

[0027]

[0028] In the formula, C bat is the instantaneous hydrogen consumption of lithium battery, P bat is the lithium battery power, η chg and η dis are the charging and discharging efficiency of lithium batteries, η chg,avg and η dis,avg are the average charging and discharging efficiency of lithium batteries, C fc,avg and P fc,avg are the average hydrogen consumption and average power of the fuel cell respectively.

[0029] Furthermore, the fuel cell degradation cost J fc and lithium battery degradation costJ bat Calculated by the following formula:

[0030]

[0031] Among them, Δt is the sampling time, ω fc is the unit price of fuel cell degradation cost, ω bat is the unit price of lithium battery degradation cost.

[0032] Furthermore, the different conditions include vehicle load, remaining amount of hydrogen in the fuel cell, and SOC of the lithium battery.

[0033] Furthermore, the deep learning network model includes an autoencoder and a GAN network connected in sequence, and the GAN network includes a generator and a discriminator.

[0034] Based on the same inventive concept, the present invention also provides an electronic device, including:

[0035] one or more processors;

[0036] A memory having one or more programs stored thereon, which, when executed by the one or more processors, enables the one or more processors to implement the steps of the energy management method of the vehicle hybrid power system.

[0037] Based on the same inventive concept, the present invention further provides a computer-readable storage medium storing a computer program, which implements the steps of the energy management method of a vehicle hybrid power system when the computer program is executed by a processor.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] The present invention utilizes a deep learning network based on autoencoder-generative adversarial, combined with a vehicle hybrid system model, an objective function of minimizing the comprehensive value loss cost, and actual operating conditions, to achieve power allocation of different power sources of the hybrid system under different vehicle operating conditions; the present invention takes into account fuel economy and the impact of power load variation of the hybrid system, and better allocates output power between different power sources during different driving processes to obtain better fuel economy and durability. The hybrid system energy management of the present invention has high adaptability under different operating conditions, so as to cope with different road conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A schematic flow chart of an energy management method for a vehicle hybrid power system of the present invention;

[0041] Figure 2 A schematic diagram of a deep learning network model based on AE-GAN of the present invention;

[0042] Figure 3 Schematic diagram of the autoencoder;

[0043] Figure 4 Schematic diagram of the GAN neural network. DETAILED DESCRIPTION

[0044] The present invention will be described in detail below in conjunction with the embodiments. It should be noted that the embodiments and features of the embodiments of the present invention can be combined with each other without conflict. For the convenience of description, if the words "upper", "lower", "left" and "right" appear in the following, they only indicate that the upper, lower, left and right directions are consistent with the drawings themselves, and do not limit the structure.

[0045] Example

[0046] At present, the energy management method of vehicle hybrid power systems (including fuel cell systems and lithium battery systems) performs well under the condition of a single line. However, for other lines, redesign is required. Therefore, the current vehicle hybrid power system energy management method has the disadvantage of poor adaptability. To solve this problem, this embodiment uses an energy management method based on deep learning of the autoencoder AE-generative adversarial network GNA. Unlike traditional energy management, this method can be self-updated according to the line conditions of the vehicle operation, thereby solving the defect that the existing hybrid power system energy management method adapts to a single working condition.

[0047] like Figure 1-Figure 4 , this embodiment provides a method for energy management of a vehicle hybrid power system, comprising the following steps:

[0048] 1. Establish a vehicle hybrid system model, including a fuel cell degradation model and a lithium battery degradation model; the fuel cell degradation model and the lithium battery degradation model can reasonably estimate the equivalent losses of the fuel cell and the lithium battery under operating conditions, and provide indicators and references for subsequent power distribution.

[0049] Fuel cell degradation model:

[0050]

[0051] Among them, m deg,fc is the performance degradation of the fuel cell, in kW; D high , D low With D change are the fuel cell losses under high power, low power and power variable load conditions, respectively. The power loss below 10% of the rated power is the low power condition, and the power loss above 90% of the rated power is the high power condition. cap,fc is the rated power of the fuel cell, in Kw, and in this embodiment, it is 4000kW; U fcnor is the rated voltage of the fuel cell, in μV, and in this embodiment, it is 468V; U EOL The voltage of the fuel cell when it is retired is expressed in μV. In this embodiment, it is 90% of the rated voltage.

[0052] Power load variation refers to the power variation caused by different vehicle operation modes. Rapid power load variation will accelerate the aging of fuel cells and even damage them. This technical solution can reasonably allocate the output power of fuel cells and lithium batteries according to the vehicle's required operating conditions, making the fuel cell power smoother, reducing fuel cell losses, and thus extending the service life of the hybrid system.

[0053] The fuel cell loss is the hydrogen consumption. The fuel cell hydrogen consumption is obtained based on experimental fitting. Its instantaneous hydrogen consumption is as follows:

[0054] C fc =0.01954P fc -0.31924

[0055] Among them, C fc is the instantaneous hydrogen consumption of the fuel cell, P fc is the fuel cell power.

[0056] Lithium battery degradation model:

[0057]

[0058] Among them, m deg,bat is the performance degradation of lithium batteries, in kWh; i bat , U bat are the output current and voltage of the lithium battery respectively; N is the number of charge and discharge cycles of the lithium battery, which is 70,000 in this embodiment.

[0059] For the convenience of calculation, the loss of lithium battery is equivalent to hydrogen consumption, and its instantaneous hydrogen consumption is:

[0060]

[0061] Among them, C bat is the instantaneous hydrogen consumption of lithium battery; P bat is the lithium battery power, when it is less than 0, it is charging, and when it is greater than 0, it is discharging; η chg and η dis are the charging and discharging efficiency of lithium batteries respectively; η chg,avg and η dis,avg are the average charging and discharging efficiencies of lithium batteries, both of which are 0.95 in this example; C fc,avg and P fc,avg are the average hydrogen consumption and average power of the fuel cell respectively.

[0062] The calculation formula for lithium battery charging and discharging efficiency is:

[0063]

[0064] Where η dis is the charging efficiency of lithium battery, η chgis the discharge efficiency of lithium battery; R int is the internal resistance of the lithium battery, which is 40 mΩ in this embodiment.

[0065] 2. Based on hydrogen cost, fuel cell loss, lithium battery loss, daily average operating cost, and SOC offset, establish an objective function to minimize the comprehensive value loss cost;

[0066]

[0067] Where J is the total operating cost of the vehicle hybrid system, T is the operating time of the hybrid system; J soc is the SOC offset of the lithium battery; J energy is the hydrogen consumption cost of the hybrid system; J fc is the fuel cell degradation cost; J bat is the degradation cost of lithium battery; J d is the average daily operating cost; k1, k2, k3, k4 are cost coefficients, is the price of hydrogen.

[0068] The lithium battery SOC offset can be obtained by comparing the current lithium battery SOC with the SOC baseline value; the average daily operating cost is obtained by dividing the operating cost of the hybrid system, such as maintenance cost, battery replacement cost, etc., according to the service life of the hybrid system, to each day. The cost coefficient is selected based on experience according to the different emphasis on different costs of the hybrid system.

[0069] Fuel cell degradation costJ fc and lithium battery degradation costJ bat Calculated by the following formula:

[0070]

[0071] Among them, Δt is the sampling time, ω fc is the unit price of fuel cell degradation cost, ω bat is the unit price of lithium battery degradation cost. The unit price of fuel cell degradation cost and lithium battery degradation cost are calculated based on the purchase cost of fuel cells and lithium batteries, as well as the degree of degradation of fuel cells and lithium batteries when they are retired. The formula is:

[0072]

[0073] In the formula, ω fc / bat is the unit price of fuel cell degradation cost or lithium battery degradation cost; cost fc / bat is the purchase price of the fuel cell or lithium battery; ii is the discount rate, which is 10% in this embodiment; T fc / bat It is the service life of the fuel cell or lithium battery in years.

[0074] The objective function aims to find a balance between fuel consumption and power source loss during vehicle operation. If the minimum hydrogen consumption is pursued blindly, it will inevitably lead to an increase in the power source output burden, which will in turn affect the life of the hybrid system; on the contrary, emphasizing the smooth output of the hybrid system will often increase the vehicle's fuel usage. Therefore, converting it into a cost for measurement can more reasonably guide the hybrid system output, ensuring the life of the power source while reducing fuel usage.

[0075] 3. Input the standard operating conditions into the vehicle hybrid power system model to obtain the hybrid power system output power curve under different conditions; the standard operating conditions include the operating speed curve, demand power curve, and SOC curve of the vehicle's designed operating conditions; different conditions include vehicle load, remaining hydrogen, and lithium battery SOC; after inputting the vehicle's standard operating conditions into the vehicle hybrid power system model, the specific parameters of the vehicle can be changed to simulate different conditions when the vehicle is driving, and the solution can be performed under different conditions to obtain the output power of the hybrid power system.

[0076] Select the vehicle's required power P load and lithium battery SOC as state space variables; select hybrid system output power as action space variable, and limit it to 0-400kW, and discretize it in intervals of 10kW. In this embodiment, the hybrid system output power is 400Kw, which can be changed according to the design power upper limit of the vehicle hybrid system. As shown in the following formula:

[0077]

[0078] State space variables and action space variables are used to solve standard working conditions.

[0079] 4. The operating speed curve, required power curve, SOC curve, and hybrid power system output power curve of the vehicle under the designed operating condition are used as the input of the deep learning network model based on AE-GAN, and the deep learning network model based on AE-GAN is trained. The deep learning network model based on AE-GAN includes an autoencoder and a GAN network connected in sequence, and the GAN network includes a generator and a discriminator;

[0080] The objective function is iteratively optimized to train the deep learning network model based on AE-GAN, and the power distribution model of the hybrid power system is obtained;

[0081] The real-time operation data of the vehicle is input into the power distribution model of the hybrid power system to obtain the output power of the fuel cell and the output power of the lithium battery in the hybrid power system. The real-time operation data of the vehicle includes the operating speed curve, demand power curve, SOC curve, etc. under the actual operating conditions of the vehicle. The real-time operation data of the vehicle, related power data, voltage data, current data, etc. can be obtained through sensors on the vehicle.

[0082] The training process of the deep learning network model based on AE-GAN includes:

[0083] The autoencoder can perform data fusion on the selected decay features, and its steps can be divided into:

[0084] 1) The operating speed curve, required power curve, SOC curve, and hybrid power system output power curve of the vehicle under the designed operating conditions are used as the original data set and input into the autoencoder;

[0085] 2) Extract characteristic parameters from the original data set;

[0086] 3) Normalize the data and perform feature fusion;

[0087] The input sample is X, the weight matrix between the input layer and the coding layer is W, the coding layer node bias is bm, and the node activation function is g(·). Then the calculation formula of the potential representation H is:

[0088] H=g(WX+b m )

[0089] The feature fusion data is used as the input of the GAN network. During the model training process, the GAN network contains the generator G and the discriminator D modules. The optimization objective function in the game process is:

[0090]

[0091] Among them, V(D,G) is the optimization objective function, x and z are the real data and the data generated by the generator, respectively, and P data With P z is the distribution of real data and generated data, D(x) and G(z) are the probability distribution functions of real data and generated data respectively, is the expected value of the real data, is the expected value of the noise data, and D(G(z)) is the discriminator’s judgment result on the generated data.

[0092] During the training process, the loss functions of the generator and the discriminator are:

[0093]

[0094] Among them, LG With L D are the loss functions of the generator and the discriminator respectively.

[0095] The parameters of the deep learning network model based on AE-GAN are set as follows: the number of iterations is 500, the generator learning rate is set to 0.0001; the discriminator learning rate is set to 0.5; and the latent space dimension is set to 100.

[0096] The trained hybrid power system power distribution model is output to the vehicle top-level control module. The model is continuously updated according to the vehicle's real-time operating data to better adapt to different operating conditions.

[0097] When applied in real time, the vehicle's real-time operating data can be used to update and supplement the model, thereby improving the adaptability of the energy management method to working conditions.

[0098] The use of a deep learning method based on an autoencoder-generative adversarial network not only extends the service life of the hybrid system, but also has better adaptability to working conditions, thereby coping with different road conditions.

[0099] Another embodiment of the present invention provides an electronic device, including:

[0100] one or more processors;

[0101] The memory stores one or more programs, and when the one or more programs are executed by one or more processors, the one or more processors implement the steps of the energy management method of the vehicle hybrid power system.

[0102] In some implementations, the memory may be a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory.

[0103] In some other implementations, the processor may be a central processing unit (CPU), a digital signal processor (DSP), or other general-purpose processors of various types, which are not limited herein.

[0104] Another embodiment of the present invention provides a computer-readable storage medium storing a computer program, which implements the steps of the energy management method of a vehicle hybrid power system when the computer program is executed by a processor.

[0105] The contents explained in the above embodiments should be understood as these embodiments are only used to more clearly illustrate the present invention, and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

Claims

1. A method for energy management of a vehicle hybrid power system, characterized in that: The following steps are involved: S1: Establishing a vehicle hybrid power system model, wherein the vehicle hybrid power system model includes a fuel cell degradation model and a lithium battery degradation model; The fuel cell degradation model is as follows: The lithium battery degradation model is as follows: S2: Establish an objective function based on hydrogen cost, fuel cell loss, lithium battery loss, average daily operating cost, and SOC offset. The objective function is as follows: S3: inputting a standard operating condition into the hybrid power system model of the vehicle to obtain a hybrid power system output power curve under different conditions; the standard operating condition includes an operating speed curve, a required power curve, and a SOC curve when the vehicle is designed to operate; S4: using the operating speed curve, the required power curve, the SOC curve, and the hybrid power system output power curve of the vehicle under the designed operating condition as inputs of the deep learning network model, and training the deep learning network model; Iteratively optimizing and training the deep learning network model using the objective function to obtain a hybrid power system power distribution model; S5: inputting the real-time operation data of the vehicle into the power distribution model of the hybrid power system to obtain the output power of the fuel cell and the output power of the lithium battery in the hybrid power system, wherein the real-time operation data of the vehicle includes the operation speed curve, the demand power curve, and the SOC curve of the vehicle under the actual operation condition; Among them, m deg,fc is the performance degradation of the fuel cell, D high , D low With D change are the fuel cell losses under high power, low power and power variable load conditions, P cap,fc is the rated power of the fuel cell, U fcnor is the rated voltage of the fuel cell, U EOL is the voltage of the fuel cell when it is retired; m deg,bat is the performance degradation of lithium battery, i bat , U bat are the output current and voltage of the lithium battery, respectively; N is the number of charge and discharge cycles of the lithium battery; J is the total operating cost of the hybrid system, T is the operating time of the hybrid system; J soc is the SOC offset of the lithium battery; J energy is the hydrogen consumption cost of the hybrid system, J fc is the fuel cell degradation cost, J bat is the degradation cost of lithium battery, J d is the average daily operating cost, k1, k2, k3, k4 are cost coefficients.

2. The energy management method of a vehicle hybrid power system according to claim 1, characterized in that: The fuel cell consumption is hydrogen consumption. The instantaneous hydrogen consumption of the fuel cell is as follows: C fc =0.01954P fc -0.31924 Among them, C fc is the instantaneous hydrogen consumption of the fuel cell, P fc is the fuel cell power.

3. The energy management method of a vehicle hybrid power system according to claim 2, characterized in that: Converting the loss of lithium batteries into hydrogen consumption, the instantaneous hydrogen consumption of lithium batteries is as follows: Among them, C bat is the instantaneous hydrogen consumption of lithium battery, P bat is the lithium battery power, η chg and η dis are the charging and discharging efficiency of the lithium battery system, η chg,avg and η dis,avg are the average charging and discharging efficiency of lithium batteries, C fc,avg and P fc,avg are the average hydrogen consumption and average power of the fuel cell respectively.

4. The energy management method of a vehicle hybrid power system according to claim 1, characterized in that: Fuel cell degradation costJ fc and lithium battery degradation costJ bat Calculated by the following formula: Among them, Δt is the sampling time, ω fc is the unit price of fuel cell degradation cost, ω bat is the unit price of lithium battery degradation cost.

5. The energy management method of a vehicle hybrid power system according to claim 1, characterized in that: The different conditions include vehicle load, remaining amount of hydrogen in the fuel cell, and SOC of the lithium battery.

6. The energy management method of a vehicle hybrid power system according to claim 1, characterized in that: The deep learning network model includes an autoencoder and a GAN network connected in sequence, and the GAN network includes a generator and a discriminator.

7. An electronic device, characterized in that: include: one or more processors; A memory having one or more programs stored thereon, which, when the one or more programs are executed by the one or more processors, enables the one or more processors to implement the steps of the method according to any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that: The computer program is stored therein, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images