CN-121993589-A - Gear shifting control method, device, equipment and storage medium for electric mining truck

CN121993589ACN 121993589 ACN121993589 ACN 121993589ACN-121993589-A

Abstract

The application discloses a gear shifting control method, device and equipment for a mining electric truck and a storage medium. The method comprises the steps of constructing a system dynamics model based on a vehicle electric drive assembly structure, training a physical information neural network by taking the system dynamics model as physical constraint in an active synchronization stage of a gear shifting process, embedding the trained physical information neural network into a model prediction control framework as a prediction model, carrying out rolling optimization solution based on a preset multi-objective optimization function to obtain an optimal control sequence, and controlling based on the optimal control sequence. Aiming at the nonlinear characteristics of the gear shifting process of the mining truck under extreme working conditions, particularly the accurate control of an active synchronization stage, the application realizes the accurate coordination control of the gear shifting process, and effectively solves the problems of torque interruption, impact vibration, low energy efficiency and the like of the pure electric mining truck under the extreme working conditions such as heavy load, low speed, large gradient, uneven road surface and the like in the gear shifting process.

Inventors

XIN YING
LI YONGCHANG
WANG JIN
CHEN WEIYAO
WANG HANRUI
DAI YUXIN

Assignees

潍柴动力股份有限公司

Dates

Publication Date: 20260508
Application Date: 20260123

Claims (10)

1. The gear shifting control method for the mining electric truck is characterized by comprising the following steps of: constructing a system dynamics model based on a vehicle electric drive assembly structure; in an active synchronization stage of a gear shifting process, training a physical information neural network by taking the system dynamics model as physical constraint; Embedding the physical information neural network after training into a model prediction control framework as a prediction model, and carrying out rolling optimization solution based on a preset multi-objective optimization function to obtain an optimal control sequence; and controlling based on the optimal control sequence.
2. The method of claim 1, wherein constructing a system dynamics model based on the vehicle electric drive assembly structure comprises: Analyzing structural parameters of the vehicle electric drive assembly; Based on the structural parameters of the electric drive assembly, a longitudinal dynamics model of the whole vehicle is established, and the longitudinal dynamics model of the whole vehicle integrates the output torque, the transmission ratio, the rolling resistance, the air resistance, the total mass of the vehicle, the wheel radius, the gradient resistance and the braking moment parameters of a plurality of driving motors; and determining a system dynamics model of an active synchronization stage based on the whole vehicle longitudinal dynamics model.
3. The method of claim 1, further comprising, prior to training a physical information neural network using the system dynamics model as a physical constraint during an active synchronization phase of a shift process: dividing a gear shifting process into an unloading torque stage, an off-shift stage, an active synchronization stage, an on-shift stage and a torque recovery stage; And monitoring the state of the gear shifting process in real time, and determining that the vehicle enters the active synchronization stage.
4. The method of claim 1, wherein training a physical information neural network with the system dynamics model as a physical constraint comprises: Based on the system dynamics model, performing state decoupling, and establishing a system state equation comprising state quantity, control quantity and output; discretizing the system state equation to obtain a discrete state transition equation; based on the discrete state transition equation as a physical constraint, training a physical information neural network.
5. The method of claim 4, wherein training a physical information neural network based on the discrete state transfer equation as a physical constraint comprises: Substituting the predicted output value of the physical information neural network into the discrete state transition equation to calculate physical loss; calculating a data loss based on a data set comprising a system state, a control input and a measured output; Constructing a composite loss function based on the physical loss and the data loss; And optimizing the composite loss function through a back propagation algorithm, and updating network weight parameters to obtain the trained physical information neural network.
6. The method of claim 1, further comprising, prior to performing the rolling optimization solution based on the preset multi-objective optimization function: determining a control target of a model predictive control framework, wherein the control target comprises rotation speed synchronization precision, rotation angle synchronization precision, gear shifting stability and minimum transient torsional vibration; And carrying out weighted summation on each control target based on the quadratic cost function and a preset weight coefficient to obtain the multi-target optimization function.
7. The method according to claim 1, wherein embedding the trained physical information neural network as a prediction model into a model prediction control framework, performing a rolling optimization solution based on a preset multi-objective optimization function to obtain an optimal control sequence, comprises: taking the physical information neural network as a prediction model, and embedding the physical information neural network into a model prediction control framework; at each rolling optimization moment of model predictive control, predicting the system state through the physical information neural network according to the current system state and a control input sequence in a future control time domain to obtain a state prediction sequence; Based on the state prediction sequence, combining a multi-objective optimization function and system constraint, solving the limited time domain optimal control problem on line in each control period to obtain an optimal control sequence.
8. A mining electric truck shift control device, comprising: the dynamics model construction module is used for constructing a system dynamics model based on the vehicle electric drive assembly structure; The neural network training module is used for training a physical information neural network by taking the system dynamics model as physical constraint in an active synchronization stage of a gear shifting process; the optimization solving module is used for embedding the physical information neural network after training into a model prediction control framework as a prediction model, and carrying out rolling optimization solving based on a preset multi-objective optimization function to obtain an optimal control sequence; and the gear shifting control module is used for controlling based on the optimal control sequence.
9. An electronic device comprising a processor and a memory storing program instructions, the processor being configured, when executing the program instructions, to perform the mining electric truck shift control method of any one of claims 1-7.
10. A computer readable medium having stored thereon computer readable instructions executable by a processor to implement a mining electric truck shift control method according to any one of claims 1 to 7.

Description

Gear shifting control method, device, equipment and storage medium for electric mining truck Technical Field The application relates to the technical field of vehicle control, in particular to a gear shifting control method, device and equipment for an electric mining truck and a storage medium. Background Along with the rapid development of mine electrification and intelligent technology, the pure electric mining truck is used as high-efficiency and environment-friendly heavy transportation equipment and is widely applied to scenes such as surface mines. However, its complex driveline presents multiple challenges during frequent shifts. The mine truck operates under extreme working conditions such as heavy load, low speed, large gradient and uneven road surface, and the traditional gear shifting control strategy is difficult to simultaneously consider smoothness, response speed and energy efficiency. Particularly under the high-load condition, torque interruption and impact vibration in the gear shifting process can cause accelerated abrasion of transmission parts, reduce the reliability of the system and influence the running efficiency of the whole vehicle. In the prior art, the gear shifting control of the mining truck mainly adopts logic control or a traditional PID control strategy based on rules, and the methods have insufficient modeling on nonlinear dynamic characteristics of the system and are difficult to adapt to working condition changes. Disclosure of Invention The embodiment of the application provides a gear shifting control method, device and equipment for an electric mining truck and a storage medium, which at least solve the technical problem that the gear shifting process of the mining truck is difficult to realize in the related technology. According to an aspect of an embodiment of the present application, there is provided a shift control method for an electric mining truck, including: constructing a system dynamics model based on a vehicle electric drive assembly structure; in an active synchronization stage of a gear shifting process, training a physical information neural network by taking the system dynamics model as physical constraint; Embedding the physical information neural network after training into a model prediction control framework as a prediction model, and carrying out rolling optimization solution based on a preset multi-objective optimization function to obtain an optimal control sequence; and controlling based on the optimal control sequence. In one embodiment, constructing a system dynamics model based on a vehicle electric drive assembly structure includes: Analyzing structural parameters of the vehicle electric drive assembly; Based on the structural parameters of the electric drive assembly, a longitudinal dynamics model of the whole vehicle is established, and the longitudinal dynamics model of the whole vehicle integrates the output torque, the transmission ratio, the rolling resistance, the air resistance, the total mass of the vehicle, the wheel radius, the gradient resistance and the braking moment parameters of a plurality of driving motors; and determining a system dynamics model of an active synchronization stage based on the whole vehicle longitudinal dynamics model. In one embodiment, before the training of the physical information neural network by taking the system dynamics model as a physical constraint in the active synchronization stage of the gear shifting process, the method further comprises: dividing a gear shifting process into an unloading torque stage, an off-shift stage, an active synchronization stage, an on-shift stage and a torque recovery stage; And monitoring the state of the gear shifting process in real time, and determining that the vehicle enters the active synchronization stage. In one embodiment, training a physical information neural network with the system dynamics model as a physical constraint includes: Based on the system dynamics model, performing state decoupling, and establishing a system state equation comprising state quantity, control quantity and output; discretizing the system state equation to obtain a discrete state transition equation; based on the discrete state transition equation as a physical constraint, training a physical information neural network. In one embodiment, training a physical information neural network based on the discrete state transition equation as a physical constraint, comprising: Substituting the predicted output value of the physical information neural network into the discrete state transition equation to calculate physical loss; calculating a data loss based on a data set comprising a system state, a control input and a measured output; Constructing a composite loss function based on the physical loss and the data loss; And optimizing the composite loss function through a back propagation algorithm, and updating network weight parameters to obtain the trained physical information neur