Implementation method and system for energy management strategy of hybrid power mining truck

Abstract

The invention relates to a hybrid power mining truck energy management strategy implementation method and system. The hybrid power mining truck energy management strategy implementation method comprises the steps that a power system model of a hybrid power mining truck is constructed according to kinetic parameters of the hybrid power mining truck; based on the power system model, carrying out modeling on the driving working condition of each stage of the hybrid power mining truck to form a target function; solving the target function by using a dynamic programming algorithm to obtain a global optimal energy management strategy of the hybrid power mining truck under each working condition; inputting the global optimal energy management strategy under each working condition as training data into a BP neural network model for training to obtain an energy distribution model; and optimizing an energy management strategy of the target hybrid power mining truck by using the energy distribution model. By using the dynamic programming algorithm and the BP neural network, the energy management strategy of the hybrid power mining truck can be dynamically adjusted according to different working conditions, and the adaptive capacity of the truck in various environments is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of hybrid vehicle energy management technology, and in particular to a method and system for implementing an energy management strategy for a hybrid mining truck. Background Art

[0002] As a key equipment in mining, the performance of hybrid mining trucks directly affects the production efficiency and operating costs of mines. Due to the complexity and variability of the mining operating environment, hybrid mining trucks need to face a variety of working conditions during operation, such as heavy load uphill, light load downhill, low speed and high torque driving, etc. These working conditions put forward extremely high requirements on energy management strategies, which must ensure both efficient use of energy and real-time response to meet the actual needs of mining operations.

[0003] Traditional energy management strategies are mostly based on rule control or simple optimization algorithms, which are difficult to adapt to complex and changeable working conditions, resulting in unreasonable energy distribution and poor fuel economy. In recent years, with the rapid development of machine learning technology, more and more researchers have begun to explore the application of machine learning algorithms to the energy management of hybrid vehicles. However, directly applying existing machine learning algorithms to the energy management of hybrid mining trucks still faces many challenges, such as high algorithm model complexity, poor real-time performance, and insufficient generalization ability. Summary of the invention

[0004] In order to solve the above problems, an object of the embodiments of the present invention is to provide a method and system for implementing an energy management strategy for a hybrid mining truck.

[0005] A method for implementing an energy management strategy for a hybrid mining truck, comprising:

[0006] A power system model of a hybrid mining truck is constructed according to the dynamic parameters of the hybrid mining truck;

[0007] Modeling the driving conditions of the hybrid mining truck at each stage based on the power system model to form an objective function;

[0008] The objective function is solved using a dynamic programming algorithm to obtain a global optimal energy management strategy for a hybrid mining truck under various operating conditions;

[0009] The global optimal energy management strategy under each working condition is input into the BP neural network model as training data for training to obtain the energy allocation model;

[0010] The energy allocation model is used to optimize the energy management strategy of a target hybrid mining truck.

[0011] Preferably, the objective function is constructed with the engine fuel consumption as the performance index function, the running time of the driving condition as the system stage variable, the hydraulic accumulator pressure change rate as the system state variable, and the driving torque of the hydraulic pump / motor as the system control variable.

[0012] Preferably, the global optimal energy management strategy under each working condition is input as training data into the BP neural network model for training to obtain the energy allocation model, including:

[0013] Classify the global optimal energy management strategy under each working condition to obtain a set of working condition categories;

[0014] The set of each working condition category is input into the BP neural network model for training to obtain the energy distribution model under each working condition category.

[0015] Preferably, the global optimal energy management strategy under each working condition is classified using a fuzzy C-means clustering algorithm based on a simulated annealing genetic algorithm to obtain a set of working condition categories.

[0016] Preferably, the set of each operating condition category is input into the BP neural network model respectively, and the BP neural network is optimized using a nested genetic algorithm to obtain the energy distribution model under each operating condition category.

[0017] Preferably, when the nested genetic algorithm is used to optimize the parameter configuration of the BP neural network, the outer loop is responsible for optimizing the number of neurons in the hidden layer, and the inner genetic algorithm is responsible for optimizing the initial weights and biases of the BP neural network.

[0018] The present invention also provides a hybrid mining truck energy management strategy implementation system, comprising:

[0019] A power model building module, used for building a power system model of a hybrid mining truck according to the dynamic parameters of the hybrid mining truck;

[0020] An objective function building module is used to use the power system model to model the driving conditions of each stage of the hybrid mining truck to form an objective function;

[0021] An optimal energy management strategy solving module is used to solve the objective function using a dynamic programming algorithm to obtain a global optimal energy management strategy for the hybrid mining truck under various working conditions;

[0022] A training module is used to input the global optimal energy management strategy under various working conditions as training data into the BP neural network model for training to obtain an energy allocation model;

[0023] The energy management strategy optimization module is used to optimize the energy management strategy of the target hybrid mining truck using the energy allocation model.

[0024] The present invention also provides an electronic device, comprising a bus, a transceiver, a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory and the processor are connected via the bus, and wherein the computer program, when executed by the processor, implements the steps in the above-mentioned method for implementing an energy management strategy for a hybrid power mining truck.

[0025] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps in the above-mentioned method for implementing an energy management strategy for a hybrid power mining truck.

[0026] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0027] The present invention relates to a method for implementing an energy management strategy for a hybrid mining truck. Compared with the prior art, the present invention can dynamically adjust the energy management strategy of a hybrid mining truck according to different working conditions by using a dynamic programming algorithm and a BP neural network, thereby improving the adaptability of the truck in various environments.

[0028] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0030] Figure 1 The system structure diagram of the hybrid power mining truck provided by the present invention shows the core components of the hybrid power mining truck and their connection methods.

[0031] Figure 2 The hybrid mining truck power system model diagram provided by the present invention shows the construction process of the hybrid mining truck power system model.

[0032] Figure 3The overall solution process diagram of the DP-based energy management strategy provided by the present invention shows the calculation process of solving the global optimal energy management strategy of the hybrid mining truck through the reverse calculation and forward calculation of the dynamic programming algorithm.

[0033] Figure 4 The BP neural network topology diagram provided by the present invention shows the basic structure and parameter configuration of the BP neural network.

[0034] Figure 5 The block diagram of the multi-BP neural network model based on working condition identification provided by the present invention shows the structure of the multi-BP neural network model based on working condition identification. DETAILED DESCRIPTION

[0035] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0036] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0037] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0038] See also Figure 1-5 , a method for implementing an energy management strategy for a hybrid mining truck, comprising:

[0039] Step 1: Construct a power system model of a hybrid mining truck according to the dynamic parameters of the hybrid mining truck;

[0040] like Figure 1 As shown, the present invention first constructs a power system model of a hybrid mining truck based on its structure and working principle. The model covers mathematical models of key components such as the engine, transmission, clutch, hydraulic pump / motor, hydraulic accumulator, and feedback element. Through these mathematical models, the operating state of a hybrid mining truck under different working conditions can be accurately simulated, including but not limited to heavy load uphill, light load downhill, and low speed and high torque driving. This step provides a solid foundation for the subsequent energy management strategy design.

[0041] Specifically, the construction of the power system model involves a detailed description of the dynamic and static characteristics of each key component. For example, the relationship between the engine's fuel consumption rate, power output, speed and torque; the speed ratio change and transmission efficiency of the transmission; the impact of the clutch engagement and disengagement process on power transmission; the relationship between the flow, pressure and torque of the hydraulic pump / motor; and the charging and discharging process of the hydraulic accumulator, etc. These models not only consider the interaction between components, but also the impact of changes in operating conditions on component performance, thereby ensuring the high fidelity and practicality of the model. The specific power system model is shown in the attached figure. Figure 2 shown.

[0042] Step 2: Modeling the driving conditions of the hybrid mining truck at each stage based on the power system model to form an objective function;

[0043] In the present invention, the engine fuel consumption is taken as the performance index function, the operating time of the driving condition is selected as the system stage variable, the hydraulic accumulator pressure change rate is selected as the system state variable, and the driving torque of the hydraulic pump / motor is selected as the system control variable. The driving condition of each stage of the hybrid mining truck is modeled to form an objective function.

[0044] Step 3: Using a dynamic programming algorithm to solve the objective function to obtain a global optimal energy management strategy for the hybrid mining truck under various working conditions;

[0045] The dynamic programming algorithm can find the global optimal energy management strategy, that is, the torque distribution ratio between the engine and the hydraulic pump / motor under different working conditions. The reverse calculation process starts from the terminal state and gradually calculates the optimal control decision for each stage; the forward calculation process gradually calculates the state trajectory and performance indicators of the system based on the optimal control decision sequence obtained by reverse calculation. To achieve this goal, the control variable T of each stage needs to be p / m_k The process can be regarded as a multi-stage decision-making problem, in which the entire driving condition is divided into N stages, and the state variable SOC of each stage is further subdivided into multiple grid points [SOC 1 ,SOC 2 …SOCi …]. Therefore, the problem of solving the global optimal energy management strategy is transformed into finding the optimal control variable decision [T p / m_1 ,T p / m_2 …T p / m_j …] corresponding fuel consumption [J 1 ,J 2 …J j …] problem. Under the DP framework, this problem is solved by the reverse solution process. Specifically, first, for the last stage N-1, the fuel cost corresponding to each control variable point under each state point is calculated, and the minimum fuel cost J corresponding to each state point is determined. i *. Then, the process is recursively extended to the N-2 stage, and the state point of the next stage corresponding to each control variable point is found, and the minimum fuel cost J of them and is calculated. i *, as the minimum fuel cost cumulative value of the state point at time N-2. This process continues until stage 0, and the minimum fuel cost cumulative value J of each state is obtained i *. This reverse calculation process is shown in the attached Figure 3 As shown in the figure, the solid line represents the reverse solution path from stage N-1 to stage 0. After the reverse calculation is completed, the forward calculation is performed. The calculation process starts from stage 0 and finds the cumulative value J of the minimum fuel consumption according to the initial state given by the system (such as SOC3). 3 *, and record its corresponding state point SOC 3 , optimal control variable T p / m_1 and the optimal fuel cost J 1 . Then, solve the state point value of the first stage. If the exact corresponding point cannot be found, use the interpolation method or the nearest minimum point to calculate, and record the corresponding state value, optimal control variable value and fuel cost value. This process continues until the N-1th stage, and finally obtains the optimal state value, optimal control variable value and optimal fuel cost value in each stage. This forward calculation process is shown in the attached figure. Figure 3 As shown, the dotted line represents the forward solution path from stage 0 to stage N-1.

[0046] Step 4: Input the global optimal energy management strategy under each working condition as training data into the BP neural network model for training to obtain the energy allocation model;

[0047] Due to the high computational complexity of the dynamic programming algorithm, it is difficult to directly apply it to real-time control. Therefore, the present invention uses a BP neural network to approximate the global optimal energy management strategy obtained by the dynamic programming algorithm. The BP neural network is a multi-layer feedforward neural network that adjusts the network weights and thresholds through a back propagation algorithm to minimize the output error.

[0048] In the present invention, firstly, a training sample set of a BP neural network is constructed according to the global optimal solution obtained by a dynamic programming algorithm. Then, the structure and parameters of the BP neural network are designed, including the number of neurons in the input layer, hidden layer and output layer, the activation function, and the initialization method of weights and thresholds. The input layer contains state information such as the required torque, vehicle speed and SOC value at the previous moment; the output layer contains the engine torque distribution ratio at the current moment. The number of neurons in the hidden layer and the selection of the activation function have an important influence on the performance of the network, so they need to be optimized through subsequent steps.

[0049] like Figure 4-5 As shown, in order to improve the performance of the BP neural network, the present invention uses a nested genetic algorithm to optimize the parameter configuration of the BP neural network. The nested genetic algorithm is an optimization method that combines a genetic algorithm and a local search algorithm, and has the characteristics of strong global search capability and fast convergence speed. In the present invention, a nested genetic algorithm framework is designed to combine the advantages of the genetic algorithm with the training process of the BP neural network. The framework consists of two nested parts: an outer loop and an inner genetic algorithm. The outer loop is responsible for optimizing the number of neurons in the hidden layer, and finds the optimal number of neurons by continuously changing the number of neurons and evaluating the network performance; the inner genetic algorithm is responsible for optimizing the initial weights and thresholds of the network, and finds the optimal weight and threshold configurations through genetic operations (such as selection, crossover, mutation) and local search algorithms (such as gradient descent).

[0050] Through the optimization of nested genetic algorithms, the optimal number of hidden layer neurons, initial weights and thresholds can be automatically and efficiently determined, thereby improving the learning efficiency and performance of the BP neural network.

[0051] The specific process of optimizing the BP neural network is as follows:

[0052] (1) Initialization: Set the parameters of the genetic algorithm (population size is 40, crossover rate is 0.7, mutation rate is 0.01, number of iterations is 100) and the search range of the number of neurons in the outer loop.

[0053] (2) Outer loop operation: For each candidate value of the number of neurons, start the inner genetic algorithm.

[0054] (3) Inner genetic algorithm operation:

[0055] ①Encoding: Encode the initial weights and thresholds of the network into chromosomes.

[0056] ②Fitness evaluation: Use the training data to train the BP neural network and calculate the error on the validation set as the individual fitness.

[0057] ③Genetic operations: perform selection, crossover and mutation operations according to fitness to generate a new population.

[0058] ④ Iteration: Repeat the genetic operation until the stopping condition is met.

[0059] (4) Result evaluation and selection: After the outer loop is completed, the optimal fitness under different numbers of neurons is compared, and the optimal configuration is selected as the final parameters of the BP neural network.

[0060] (5) Network training and testing: Retrain the BP neural network using the optimal parameters and evaluate its performance on the test set.

[0061] Specifically, before training the BP neural network model, the present invention proposes a multi-BP neural network model based on working condition recognition for the energy management problem of hybrid mining trucks under complex and changeable open-air working conditions. The model uses an intelligent working condition recognition module to determine the working condition category of the current input sample in real time, and automatically calls the corresponding BP neural network sub-model for calculation.

[0062] First, the fuzzy C-means clustering algorithm based on simulated annealing genetic algorithm is used to classify the operating condition data, and the complex and changeable operating conditions are divided into multiple representative categories. Then, for each operating condition category, an optimized BP neural network sub-model is independently constructed and trained. These sub-models have the same network structure and parameter configuration optimization method, but the training sample set and training process are customized according to the characteristics of each operating condition category.

[0063] The multi-BP neural network model based on working condition identification has two processes: offline training and online simulation:

[0064] (1) Offline training process

[0065] 1) Dataset acquisition: The optimal energy management rule dataset under various driving conditions is obtained through the DP algorithm.

[0066] 2) Cluster analysis: The FCM-GASA algorithm was used to divide the data set into 10 categories.

[0067] 3) Sub-model construction: For each category, a BP neural network sub-model is constructed, and the appropriate network structure and parameters are set according to the NGA-BP algorithm.

[0068] 4) Sub-model training: Use the training data of the corresponding category to fully train each sub-model and save the trained model parameters.

[0069] (2) Online simulation process

[0070] 1) Real-time data input: obtain the current required torque (Tr(k)), vehicle speed (v(k)) and the SOC value of the previous moment (SOC(k-1)).

[0071] 2) Working condition identification: The current data is classified through the fuzzy C-means clustering algorithm of the simulated annealing genetic algorithm to determine the working condition category to which it belongs.

[0072] 3) Sub-model calling: According to the working condition category, call the corresponding BP neural network sub-model for calculation.

[0073] 4) Optimal distribution: Each sub-model outputs the optimized torque distribution result under its own working condition, and the total output torque distribution of the hybrid mining truck is obtained through integration.

[0074] In actual applications, the working condition recognition module determines the working condition category of the current input sample based on the current required torque, vehicle speed, and SOC value at the previous moment, and calls the corresponding BP neural network sub-model for calculation. This mechanism not only greatly improves the processing efficiency and flexibility of the model, improves the generalization ability, but also ensures that stable and accurate control results can be obtained under different working conditions.

[0075] Step 5: Optimize the energy management strategy of the target hybrid mining truck using the energy allocation model.

[0076] Based on the BP neural network model and working condition identification module constructed in the above steps, the present invention realizes an adaptive machine learning energy management strategy for hybrid mining trucks. In the real-time control process, the state information such as the required torque, vehicle speed and SOC value at the previous moment are first obtained. Then, the current working condition category is determined by the working condition identification module, and the corresponding BP neural network sub-model is called for calculation. Finally, according to the output result of the BP neural network sub-model, the engine torque distribution ratio at the current moment is determined to realize real-time and efficient energy management.

[0077] The energy management strategy is highly adaptive and real-time, and can automatically adjust the energy distribution ratio according to the needs under different working conditions, thereby maximizing the fuel economy of hybrid mining trucks. At the same time, due to the use of advanced machine learning technologies such as BP neural network and nested genetic algorithm, the strategy also has good generalization ability and robustness, and can adapt to the complex and changing mining operating environment.

[0078] The present invention also provides a hybrid mining truck energy management strategy implementation system, comprising:

[0079] A power model building module, used for building a power system model of a hybrid mining truck according to the dynamic parameters of the hybrid mining truck;

[0080] An objective function building module is used to use the power system model to model the driving conditions of each stage of the hybrid mining truck to form an objective function;

[0081] An optimal energy management strategy solving module is used to solve the objective function using a dynamic programming algorithm to obtain a global optimal energy management strategy for the hybrid mining truck under various working conditions;

[0082] A training module is used to input the global optimal energy management strategy under various working conditions as training data into the BP neural network model for training to obtain an energy allocation model;

[0083] The energy management strategy optimization module is used to optimize the energy management strategy of the target hybrid mining truck using the energy allocation model.

[0084] The present invention also provides an electronic device, comprising a bus, a transceiver, a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory and the processor are connected via the bus, and is characterized in that when the computer program is executed by the processor, the steps in the above-mentioned method for implementing an energy management strategy for a hybrid mining truck are implemented. Compared with the prior art, the beneficial effects of the electronic device provided by the present invention are the same as the beneficial effects of the method for implementing an energy management strategy for a hybrid mining truck described in the above-mentioned technical solution, and are not elaborated herein.

[0085] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps in the above-mentioned method for implementing an energy management strategy for a hybrid mining truck are implemented. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present invention are the same as the beneficial effects of the method for implementing an energy management strategy for a hybrid mining truck described in the above-mentioned technical solution, which will not be repeated here.

[0086] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technical solution that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for implementing a hybrid mining truck energy management strategy, characterized in that: include: A power system model of a hybrid mining truck is constructed according to the dynamic parameters of the hybrid mining truck; Modeling the driving conditions of the hybrid mining truck at each stage based on the power system model to form an objective function; The objective function is solved using a dynamic programming algorithm to obtain a global optimal energy management strategy for a hybrid mining truck under various operating conditions; The global optimal energy management strategy under each working condition is input into the BP neural network model as training data for training to obtain the energy allocation model; The energy allocation model is used to optimize the energy management strategy of a target hybrid mining truck.

2. A method for implementing a hybrid mining truck energy management strategy according to claim 1, characterized in that: The objective function is constructed with engine fuel consumption as the performance index function, the operating time of the driving condition as the system stage variable, the hydraulic accumulator pressure change rate as the system state variable, and the driving torque of the hydraulic pump / motor as the system control variable.

3. A method for implementing a hybrid mining truck energy management strategy according to claim 2, characterized in that: The global optimal energy management strategy under each working condition is input as training data into the BP neural network model for training to obtain the energy allocation model, including: Classify the global optimal energy management strategy under each working condition to obtain a set of working condition categories; The set of each working condition category is input into the BP neural network model for training to obtain the energy distribution model under each working condition category.

4. A method for implementing a hybrid mining truck energy management strategy according to claim 3, characterized in that: The fuzzy C-means clustering algorithm based on simulated annealing genetic algorithm is used to classify the global optimal energy management strategy under each working condition to obtain the set of working condition categories.

5. A method for implementing a hybrid mining truck energy management strategy according to claim 4, characterized in that: The set of each working condition category is input into the BP neural network model respectively, and the BP neural network is optimized using a nested genetic algorithm to obtain the energy distribution model under each working condition category.

6. A method for implementing a hybrid mining truck energy management strategy according to claim 5, characterized in that: When the nested genetic algorithm is used to optimize the parameter configuration of the BP neural network, the outer loop is responsible for optimizing the number of neurons in the hidden layer, and the inner genetic algorithm is responsible for optimizing the initial weights and biases of the BP neural network.

7. A hybrid mining truck energy management strategy implementation system, characterized in that: include: A power model building module, used for building a power system model of a hybrid mining truck according to the dynamic parameters of the hybrid mining truck; An objective function building module is used to use the power system model to model the driving conditions of each stage of the hybrid mining truck to form an objective function; An optimal energy management strategy solving module is used to solve the objective function using a dynamic programming algorithm to obtain a global optimal energy management strategy for the hybrid mining truck under various working conditions; A training module is used to input the global optimal energy management strategy under various working conditions as training data into the BP neural network model for training to obtain an energy allocation model; The energy management strategy optimization module is used to optimize the energy management strategy of the target hybrid mining truck using the energy allocation model.

8. An electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, wherein: When the computer program is executed by the processor, the steps in the method for implementing an energy management strategy for a hybrid mining truck as described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the method for implementing an energy management strategy for a hybrid mining truck as described in any one of claims 1 to 6 are implemented.

Images