METHOD AND SYSTEM FOR CONTROL OF BATTERY THERMAL MANAGEMENT OF ELECTRIC DRIVE BULLDOZERS BASED ON DEEP LEARNING

Abstract

A method and system for control of BTM of electric drive bulldozers based on deep learning includes: obtaining a micro-load segment at the current moment in a load spectrum and predicting a micro-load segment at the next moment using a Markov chain model; carrying out a weighting calculation of the segment using the segment at the current moment and a corresponding weight thereof with the predicted segment and a corresponding weight thereof, and then calculating motor speed, motor torque, and a state of charge (SOC) of battery under the segment after being weighted; and taking a real-time battery outlet water temperature, occupant cabin temperature, occupant cabin target temperature and ambient temperature as well as the motor speed, motor torque and SOC, so as to obtain a control strategy of the electric compressor speed, and further obtain a method for control of BTM of the electric drive bulldozer.

Claims

1. A method for control of battery thermal management (BTM) of electric drive bulldozers based on deep learning, comprising: obtaining a micro-load segment at the current moment in a load spectrum and predicting a micro-load segment at the next moment using a Markov chain model; carrying out a weighting calculation of the micro-load segment using the micro-load segment at the current moment and a corresponding weight thereof with the predicted micro-load segment and a corresponding weight thereof, and then calculating a motor speed, a motor torque, and a state of charge (SOC) of battery under the micro-load segment after being weighted; and taking a real-time battery outlet water temperature, an occupant cabin temperature, an occupant cabin target temperature and an ambient temperature as well as the motor speed, the motor torque and the SOC which are obtained by calculation after the micro-load segment being weighted as input quantities of an electric compressor rotating speed prediction model to predict an electric compressor speed, so as to obtain a control strategy of the electric compressor speed, and further obtain a method for control of BTM of the electric drive bulldozer.

2. The method according to claim 1, wherein before predicting the micro-load segment at the next moment using the Markov chain model, further comprising: building a probability transfer matrix of each type of working condition using the micro-load segments obtained from historical data of existing load spectrums.

3. The method according to claim 2, wherein during the prediction of the next micro-load segment using the Markov chain model, selecting the working condition with the maximum transfer probability as the prediction result of the micro-load segment at the next moment based on the micro-load segment at the current moment of the load spectrum.

4. The method according to claim 1, wherein a calculation formula of the weight of the predicted micro-load segment is:
W.sub.t+1=−α+βe.sup.∇x.sup.t.sup./γ, wherein, ∇x.sub.t=|x.sub.t+1−x.sub.t| and 0≤∇x.sub.t≤M, x.sub.t is the micro-load segment of the electric drive bulldozer at the t moment, x.sub.t+1 is the micro-load segment of the electric drive bulldozer at the t+1 moment, W.sub.t+1 is the weight of the t+1 moment, a correlation coefficient of the weights of the loads at the t moment and the t+1 moment is −1, the weighted micro load segment is: x.sub.t×(1−W.sub.t+1)+x.sub.t+1×W.sub.t+1, wherein, α, β, γ and M are constants.

5. The method according to claim 1, wherein the electric compressor speed prediction model is obtained by training by a support vector machine (SVM) algorithm improved by a dual-population adaptive genetic algorithm (DPAGA).

6. The method according to claim 1, wherein during the training of the electric compressor speed prediction model, setting the initial parameters of the SVM algorithm improved by DPAGA, such as population number, maximum iteration times, crossover probability, generation gap, etc., then using a parameters combination of a penalty factor generated randomly and a variance of Radial Basis Function (RBF) as an initial population, and then performing the operations of selection, crossover, and dual-population adaptive mutation for the population of each generation to find the parameters combination of the penalty factor and the variance of RBF that minimizes the error of the SVM algorithm.

7. The method according to claim 1, wherein the training samples for training the electric compressor speed prediction model are obtained through a joint simulation and operation of a one-dimensional thermal management software and a three-dimensional thermal management software.

8. A system for control of BTM of electric drive bulldozers based on deep learning, comprising: a micro-load segment prediction module, being configured to obtain a micro-load segment at the current moment in a load spectrum and predicting a micro-load segment at the next moment using a Markov chain model; a weighted working condition parameters calculation module, being configured to carry out a weighting calculation of the micro-load segment using the micro-load segment at the current moment and a corresponding weight thereof with the predicted micro-load segment and a corresponding weight thereof, and then calculating a motor speed, a motor torque, and a state of charge (SOC) of battery under the micro-load segment after being weighted; and an electric compressor speed prediction module, being configured to take a real-time battery outlet water temperature, an occupant cabin temperature, an occupant cabin target temperature and an ambient temperature as well as the motor speed, the motor torque and the SOC which are obtained by calculation after the micro-load segment being weighted as input quantities of an electric compressor rotating speed prediction model to predict an electric compressor speed, to obtain a method for control of BTM of the electric drive bulldozer.

9. A computer-readable storage medium, having a computer program stored thereon; when the program is executed by a processor, implements the steps of the method for control of BTM of electric drive bulldozers based on deep learning of claim 1.

10. A computer device, comprising a memory, a processor and a computer program stored in the memory and runnable on the processor; when the processor executes the program, implements the steps of the method for control of BTM of electric drive bulldozers based on deep learning of claim 1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary examples of the present invention and descriptions thereof are used to explain the present invention, and do not constitute an improper limitation of the present invention.

[0027] FIG. 1 is a flow diagram of a method for control of BTM of electric drive bulldozers based on deep learning of examples of the present invention; and

[0028] FIG. 2 is a training result of a SVM model for the compressor speed control strategy of examples of the present invention.

DETAILED DESCRIPTION

[0029] The present invention will now be further described with reference to the accompanying drawings and examples.

[0030] It should be pointed out that the following detailed descriptions are all illustrative and are intended to provide further descriptions of the present invention. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those usually understood by a person of ordinary skill in the art to which the present invention belongs.

[0031] It should be noted that the terms used herein are merely used for describing specific implementations, and are not intended to limit exemplary implementations of the present invention. As used herein, the singular form is also intended to include the plural form unless the context clearly dictates otherwise. In addition, it should further be understood that, terms “comprise” and/or “comprising” used in this specification indicate that there are features, steps, operations, devices, components, and/or combinations thereof.

Example 1

[0032] As shown in FIG. 1, the present example provides a method for control of battery thermal management (BTM) of electric drive bulldozers based on deep learning, comprising the steps of:

[0033] Step S101: obtaining a micro-load segment at the current moment in a load spectrum and predicting a micro-load segment at the next moment using a Markov chain model.

[0034] Specifically, before predicting the micro-load segment at the next moment using the Markov chain model, further comprising: [0035] building a probability transfer matrix of each type of working condition using the micro-load segments obtained from historical data of existing load spectrums.

[0036] The predicting the micro-load segments of the electric drive bulldozer using the Markov chain model, and building the probability transfer matrix of each type of the working conditions based on the available data, comprising: setting a micro-load in current state to be s.sub.i (i=1, 2, . . . , p), a micro-load in next state to be s.sub.j (j=1, 2, . . . , q), and letting a micro-load in the state at t moment be S.sub.t, a micro-load in the state at t+1 moment be S.sub.t+1. Then, the transfer probability of transferring from the S.sub.t=s.sub.i in current state to the S.sub.t+1=s.sub.j in next state can be expressed as

[00001]$S_t=s_i{S}_{t+1}=s_{j}P\left(S_{t+1}=s_j|S_t=s_i\right)=P_{\mathrm{ij}}=\frac{N_{\mathrm{ij}}}{\underset{j=1}{\overset{q}{.Math.}}{N}_{\mathrm{ij}}},$

wherein P.sub.ij is the transfer probability for the micro-load in current state s.sub.i reaches to the micro-load in next state s.sub.j, N.sub.ij is the number of events for the micro-load in current state s.sub.i reaches to the micro-load in next state s.sub.j, and

[00002]$\underset{j=1}{\overset{q}{.Math.}}{N}_{\mathrm{ij}}$

is the total number of events for the micro-load in current state s.sub.i reaches to any micro-load in the next state.

[0037] In the specific implementation, during the prediction of the next micro-load segment using the Markov chain model, selecting the working condition with the maximum transfer probability as the prediction result of the micro-load segment at the next moment based on the micro-load segment at the current moment of the load spectrum.

[0038] Specifically, the selecting the working condition with the maximum transfer probability as the prediction result of the micro-load segment at the next moment s.sub.j based on the micro-load segment at the current moment s.sub.i, is: S.sub.t+1.sup.P=arg max.sub.s.sub.j P(S.sub.t+1=s.sub.j|S.sub.t=s.sub.i), so obtaining the first-order Markov chain prediction model.

[0039] Step S102: carrying out a weighting calculation of the micro-load segment using the micro-load segment at the current moment and a corresponding weight thereof with the predicted micro-load segment and a corresponding weight thereof, and then calculating a motor speed, a motor torque, and a state of charge (SOC) of battery under the micro-load segment after being weighted.

[0040] A correlation coefficient between the load weight at the t moment and the load weight at the t+1 moment is −1, and an objective weight is determined according to the magnitude of the variability of the load spectrums of the two load segments; wherein, the larger the difference, the more information the t+1 moment provides and the greater the role it can play in the comprehensive evaluation, and the greater its weight. The weight at the t+1 moment is:

W.sub.t+1=−α+βe.sup.∇x.sup.t.sup./γ, [0041] wherein, ∇x.sub.t=|x.sub.t+1−x.sub.t| and 0≤∇x.sub.t≤M, x.sub.t is the micro-load segment of the electric drive bulldozer at the t moment, x.sub.t+1 is the micro-load segment of the electric drive bulldozer at the t+1 moment, W.sub.t+1 is the weight of the t+1 moment, a correlation coefficient of the weights of the loads at the t moment and the t+1 moment is −1, the weighted micro-load segment is: x.sub.t×(1−W.sub.t+1)+x.sub.t+1×W.sub.t+1, wherein, α, β, γ and M are constants.

[0042] For example, W.sub.t+1=−0.0316+0.02989e.sup.∇x.sup.t.sup./27.728, W.sub.t+1=−0.0316+0.02989e.sup.∇x.sup.t.sup./27.728 M=80.

[0043] It should be noted that the specific values of α, β, γ and M, the skilled person in the field can be specifically set according to the actual situation, and will not be repeated here.

[0044] Step S103: taking a real-time battery outlet water temperature, an occupant cabin temperature, an occupant cabin target temperature and an ambient temperature as well as the motor speed, the motor torque and the SOC which are obtained by calculation after the micro-load segment being weighted as input quantities of an electric compressor rotating speed prediction model to predict an electric compressor speed, to obtain a method for control of the BTM of the electric drive bulldozer.

[0045] In the present example, the electric compressor speed prediction model is obtained by training by a SVM algorithm improved by DPAGA.

[0046] During the performance of the mutation operation, the DPAGA divides the population into two sub-populations, wherein the sub-population with lower fitness performs a self-adaptive Cauchy mutation and the population with higher fitness performs a self-adaptive Gaussian mutation to complete an optimization search process. The individual i is updated to:

[00003]$x_i^{\prime }=\{\begin{array}{cc}{x}_i+\mathrm{range}.Math.F\left(x_i\right).Math.N_i\left(0,1\right)& F\left(x_i\right)\le 0.5\\ x_i+\mathrm{range}.Math.F\left(x_i\right).Math.C_i\left(1,0\right)& F\left(x_i\right)>0.5\end{array};$

[0047] wherein,

[00004]$F\left(x_i\right)=\frac{f\left(x_i\right)-f_{\min }}{f_{\max }-f_{\min }}$

is the proportional transformation function, ƒ(x.sub.i) is the fitness function value of the individual x.sub.i, ƒ.sub.min and ƒ.sub.max are the minimum value and the maximum value of the fitness function of each individual in the current iteration of the population, respectively (taking the minimum value as an example, the smaller the fitness function value, the better the individual). x.sub.i and x.sub.i′ are the i-th chromosomes before and after the mutation, respectively; range is the moving range of the individual, N.sub.i(0,1) is the random number of Gaussian distribution, and C.sub.i(0,1) is the random number of the Cauchy distribution.

[0048] During the training of the electric compressor speed prediction model, setting the initial parameters of the SVM algorithm improved by DPAGA, such as population number, maximum iteration times, crossover probability, generation gap, etc., then using a parameters combination of a penalty factor generated randomly and a variance of Radial Basis Function (RBF) as an initial population, and then performing the operations of selection, crossover, and dual-population adaptive mutation for the population of each generation to find the parameters combination of the penalty factor and the variance of RBF that minimizes the error of the SVM algorithm.

[0049] Wherein, the training samples used for training the electric compressor speed prediction model by the improved SVM are obtained through a joint simulation and operation of one-dimensional thermal management software and three-dimensional thermal management software. The different ambient temperatures, vehicle speeds, battery heat dissipations (equivalent to motor speed, motor torque and battery output power), occupant cabin temperatures and occupant cabin target temperatures, which are set for the BTM subsystem, are used as the training sample inputs, the compressor speeds obtained by the simulation and which meet the battery safety temperature requirements and their corresponding duty cycles are used as the training sample outputs, and then the above-mentioned inputs and outputs are used as the training samples of the SVM training the prediction model.

[0050] The prediction models of compressor speed under different working conditions can be obtained through the training by the SVM algorithm improved by DPAGA, and then the method for control of the BTM of the electric drive bulldozers may be formed.

[0051] Based on the control strategy of BTM obtained by the above-mentioned method, the BTMS for electric drive bulldozers is finally formed by combining the evaporator, the condenser, the microchannel heat exchanger of battery, the electronic expansion valve and other components.

Example 2

[0052] The present example provides a system for control of BTM of electric drive bulldozers based on deep learning, specifically comprising the modules of: [0053] a micro-load segment prediction module, being configured to obtain a micro-load segment at the current moment in a load spectrum and predicting a micro-load segment at the next moment using a Markov chain model; [0054] a weighted working condition parameters calculation module, being configured to carry out a weighting calculation of the micro-load segment using the micro-load segment at the current moment and a corresponding weight thereof with the predicted micro-load segment and a corresponding weight thereof, and then calculating a motor speed, a motor torque, and a state of charge (SOC) of battery under the micro-load segment after being weighted; and [0055] an electric compressor speed prediction module, being configured to take a real-time battery outlet water temperature, an occupant cabin temperature, an occupant cabin target temperature and an ambient temperature as well as the motor speed, the motor torque and the SOC which are obtained by calculation after the micro-load segment being weighted as input quantities of an electric compressor rotating speed prediction model to predict an electric compressor speed, to obtain a method for control of BTM of the electric drive bulldozer.

[0056] It should be noted here that the modules in the system for control of BTM of electric drive bulldozers based on deep learning of the present example correspond to the steps in the method for control of the BTM of electric drive bulldozers based on deep learning of Example 1, and the specific implementation process is the same, which will not be repeated here.

Example 3

[0057] The present example provides a computer-readable storage medium, having a computer program stored thereon; when the program is executed by a processor, implements the steps of a method for control of the BTM of electric drive bulldozers based on deep learning as described above.

Example 4

[0058] The present example provides a computer device, comprising a memory, a processor and a computer program stored in the memory and runnable on the processor; when the processor executes the program, implements the steps of a method for control of the BTM of electric drive bulldozers based on deep learning as described above.

[0059] Those skilled in the art should understand that the examples of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of hardware examples, software examples, or examples combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk memory, optical memory, etc.) containing computer usable program codes.

[0060] The present invention is described with reference to methods, devices (systems) and flowcharts and/or block diagrams of computer program products according to the examples of the present invention. It should be understood that each of the processes and/or boxes in the flowchart and/or block diagram, and the combination of the processes and/or boxes in the flowchart and/or block diagram, may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, a specialized computer, an embedded processor, or other programmable data processing device to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing the functions specified in one process or multiple processes of the flowchart and/or one box or multiple boxes of the block diagram.

[0061] These computer program instructions may also be stored in a computer-readable memory capable of directing the computer or other programmable data processing apparatus to operate in a particular manner such that the instructions stored in such the computer-readable memory produce an article of manufacture comprising an instruction device that implements the function specified in one process or a plurality of processes of the flowchart and/or in one box or a plurality of boxes of the block diagram.

[0062] These computer program instructions may also be loaded onto a computer or other programmable data processing device to enable a series of operational steps to be performed on the computer or other programmable device to generate a computer implemented process, so that instructions executed on a computer or other programmable device provide steps for implementing functions specified in one process or a plurality of processes of the flowchart and/or in one box or a plurality of boxes of the block diagram.

[0063] Those skilled in the art can understand that the realization of all or part of the processes in the methods of the above examples can be accomplished by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it may comprise the processes of the examples of the above methods. The storage medium may be a disk, optical disc, Read-only memory (ROM) or random access memory (RAM), etc.

[0064] The foregoing descriptions are merely preferred examples of the present invention, but not intended to limit the present invention. A person skilled in the art may make various alterations and variations to the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.