CN-121989909-A - Self-adaptive switching method of multi-mode power system of hybrid commercial vehicle

CN121989909ACN 121989909 ACN121989909 ACN 121989909ACN-121989909-A

Abstract

The invention discloses a self-adaptive switching method of a multi-mode power system of a hybrid commercial vehicle, which relates to the technical field of automobile power control and comprises the specific steps of multi-dimensional information acquisition, acquisition of multi-type information by means of sensors and the like and processing and constructing of a data set; the method comprises the steps of constructing a five-dimensional model, dividing a five-dimensional parameter quantization interval, calculating an adaptation degree to generate a candidate set, screening a dynamic weight game and an optimal mode, constructing a three-dimensional objective function screening optimal mode, predictively pre-switching and adjusting, constructing a model pre-judging and pre-adjusting power source, dynamically compensating and adjusting component loss, and dynamically monitoring and compensating out-of-standard loss in real time.

Inventors

WANG RUISUO
LIAN GUOFU
JIANG JIBIN
YU XIWEI

Assignees

福建理工大学

Dates

Publication Date: 20260508
Application Date: 20260120

Claims (10)

1. The self-adaptive switching method of the multi-mode power system of the hybrid commercial vehicle is characterized by comprising the following specific steps of: S100, multidimensional information acquisition, namely acquiring load, road surface, power source, component loss and task mode information through a sensor and an interface, preprocessing and extracting features, and constructing a data set containing running states and task attributes; S200, constructing a five-dimensional model, namely dividing a five-dimensional parameter quantization interval containing a task mode, total load, road surface characteristics, power source efficiency and component loss based on the data set, calculating the adaptation degree of the five-dimensional parameter and the power mode through a five-dimensional parameter coupling adaptation algorithm, and generating a power mode candidate set; S300, constructing a three-dimensional objective function comprising energy efficiency, loss and response speed, calculating the power mode candidate set based on the objective function, and screening out an optimal power mode with the highest objective function value; S400, predictive pre-switching adjustment, namely constructing a predictive model by utilizing historical operation data and task attributes, pre-judging load change, road surface characteristics and task demand fluctuation of a future road section, calculating a response speed correction value by combining a response correction algorithm, and pre-starting, pre-closing or parameter pre-adjusting a power source according to a pre-judging result and the optimal power mode; S500, dynamic compensation adjustment of the component loss, namely monitoring the component loss parameter in real time in the operation process of the optimal power mode, triggering a compensation mechanism in combination with the current task mode, and dynamically adjusting the power distribution proportion or the operation mode until the loss parameter returns to a safety interval when the component loss exceeds the standard.
2. The self-adaptive switching method of the multi-mode power system of the hybrid commercial vehicle is characterized in that in S100, in multi-dimensional information acquisition, 4 suspension pressure sensors arranged on front and rear axles of a cargo compartment respectively acquire all suspension pressure data in real time to calculate a total load value and a gravity center offset coefficient, road gradient, friction coefficient and pit level are generated in real time through fusion of a 16-wire vehicle-mounted laser radar and a high-precision navigation map, engine rotating speed and load rate are acquired in real time through an engine ECU, torque and rotating speed of a front motor and a rear motor are acquired in real time through a double-motor controller, engine cylinder abrasion amount, battery capacity attenuation rate and motor winding temperature rise accumulation value are calculated in real time through a loss monitoring module, a manually selected task mode is acquired through a vehicle-mounted touch screen and a user interaction interface of a mobile phone APP, and distribution timeliness requirements and cargo types in orders are automatically identified through an API interface of a logistics order system.
3. The self-adaptive switching method of the multi-mode power system of the hybrid commercial vehicle according to claim 1 is characterized in that in the step S200, in the five-dimensional model construction, the specific contents of a five-dimensional parameter quantization interval are that a task mode quantization interval is divided into an extreme fuel-saving mode, a standard balance mode, a time-limited delivery mode and a component maintenance mode, a total load quantization interval is divided into no-load, medium-load and full-load, a road surface characteristic quantization interval is divided into a gradient, a friction coefficient and a pothole grade, a power source efficiency quantization interval is divided into an engine high-efficiency area and a motor high-efficiency area, and a component loss quantization interval is divided into normal, critical and excessive.
4. The adaptive switching method of the multi-mode power system of the hybrid commercial vehicle according to claim 1, wherein in the step S200 of constructing the five-dimensional model, the expression of the five-dimensional parameter coupling matching algorithm is as follows: , wherein, The adaptation degree of the kth power mode is 0-1, For the task mode quantization factor, As a result of the overall load quantization factor, As a coupling factor for the characteristics of the road surface, As a factor of the efficiency of the power source, For the component loss quantization factor, The dynamic weights representing the task pattern dimensions, As a dynamic weight of the overall load dimension, Dynamic weights for component loss dimensions, and Screening Forms a candidate set.
5. The adaptive switching method of a multi-mode power system of a hybrid commercial vehicle according to claim 1, wherein in the step S300, the expression of the three-dimensional objective function in the dynamic weight game and the optimal mode screening is: , wherein, The value of the objective function is indicated, The total energy efficiency is indicated by the term, Indicating the total loss of the material and, The overall response speed is indicated and, The energy efficiency weight is represented as a weight of the energy efficiency, The loss weight is represented by a weight of the loss, Response speed weight, and 。
6. The adaptive switching method of a multi-mode power system of a hybrid commercial vehicle according to claim 5, wherein in the step S300, the energy efficiency weight is determined in the dynamic weight game and the optimal mode screening The calculation formula of (2) is as follows: , wherein, Representing the task energy efficiency sensitivity coefficient, Representing total loss, energy efficiency weight The calculation formula of (2) is as follows: , wherein, Is a task loss sensitivity coefficient, response speed weight The calculation formula of (2) is as follows: , wherein, Is a task response coefficient.
7. The adaptive switching method of the multi-mode power system of the hybrid commercial vehicle according to claim 1, wherein in the step S400, the predictive pre-switching adjustment, the predictive model is an LSTM neural network model based on time sequence feature fusion, and specifically includes: The input layer parameters comprise historical operation data, current task attributes, task modes and real-time traffic information, wherein the historical operation data comprise load change curves, road surface feature libraries and power mode switching histories of the same route for the past 3 months; the feature processing layer is used for carrying out normalization processing on the input parameters, distributing weights for different features through an attention mechanism and extracting time sequence features; The network structure comprises 3 layers of LSTM units, wherein each layer is connected with a Dropout layer to prevent overfitting, and the output layer is a full-connection layer; output parameters, namely load change trend, road surface characteristic change and task demand fluctuation of a road section of 500 meters in the future; The training process comprises adopting a sliding window method, adopting a parameter adjustment mode of self-adaptive learning rate, enabling the initial learning rate to be 0.001, reducing the mean value of the square difference between the predicted value and the actual value to the minimum value by continuously adjusting model parameters, stopping training when the mean value of the square difference of a verification set is continuously less than or equal to 0.02, and carrying out incremental update on the model every 7 days based on newly added operation data.
8. The adaptive switching method of a multi-mode power system of a hybrid commercial vehicle according to claim 7, wherein in the step S400, the training of the predictive model is: The data preparation comprises the steps of collecting historical operation data and newly-added operation data, integrating the historical operation data into a training data set, wherein the historical operation data comprises the same-route load change, road surface characteristic measured values and power mode switching records; The sliding window division comprises the steps of dividing a training data set by adopting a sliding window method, setting the size of the window and the sliding step length, selecting data from each window according to a set proportion as the training set, and taking the rest data as the verification set; The model structure initialization comprises the steps of setting a network to comprise a plurality of layers of time sequence processing units, setting a random data discarding mechanism after each layer, setting discarding proportion, and setting output nodes corresponding to the prediction parameters by an output layer; Iterative training, namely setting an initial learning rate by adopting an optimization mode of the self-adaptive learning rate, calculating a prediction result by using training set data input models in each iteration, and adjusting network parameters by comparing actual values; stopping training, namely stopping current training when the deviation of the verification set continuously meets the set condition, and storing the network parameter at the moment as a current version of the model; And incremental updating, namely adding the newly added operation data into a training data set at regular intervals, repeating incremental training based on the current model version, and updating network parameters.
9. The adaptive switching method of a multi-mode power system of a hybrid commercial vehicle according to claim 1, wherein in the step S400, the expression of the response correction algorithm is: , wherein, Indicating the total response speed after the correction, Indicating the power mode switching delay time, The pre-judgment deviation coefficient is represented, Is the response correction coefficient.
10. The adaptive switching method of the multi-mode power system of the hybrid commercial vehicle according to claim 1, wherein in the step S500, the component loss dynamic compensation adjustment is specifically implemented by a compensation mechanism: In the component maintenance mode, when the abrasion loss of an engine cylinder reaches 30%, starting alternating load, namely continuously starting a motor before 3 times, starting a motor after 4 times, switching the motor for more than or equal to 5 seconds, and starting balanced charging when the battery capacity decays to 10% and the vehicle is parked at night; In the time-limited delivering mode, when the temperature rise of the motor winding reaches 60 ℃, the torque equipartition of the double motors is triggered, and the total torque drop amplitude is less than or equal to 10%; In the standard balance mode, when the component loss reaches a critical value, the engine load rate is limited to be less than or equal to 80 percent, and the motor torque is limited to be less than or equal to 200N and m.

Description

Self-adaptive switching method of multi-mode power system of hybrid commercial vehicle Technical Field The invention relates to the technical field of automobile power control, in particular to a self-adaptive switching method of a multi-mode power system of a hybrid commercial vehicle. Background Along with the continuous promotion of the transportation industry to energy consumption control and operation efficiency requirement, hybrid commercial car is because having traditional fuel power and electrically driven advantage concurrently, become the important direction of coping industry demand, the operating condition of this type of commercial car has apparent complexity, not only face the frequent fluctuation of load, the big problem of road surface condition difference, still need the task mode of adaptation diversification, the different tasks have apparent difference to the requirement of power take off, present, the trade no longer limits to basic drive function to the performance requirement of hybrid commercial car driving system, more emphasize the cooperation that realizes energy efficiency maximize under different operating conditions, the component loss is minimized and response speed is optimized, however, in commercial car actual operation, running state and task attribute's dynamic change make traditional power mode switching mode be difficult to accurate adaptation, a set of self-adaptation switching technique that can integrate the multidimension degree information, realize dynamic optimization, in order to promote power system's wholeness and running performance, satisfy the operation demand under the complicated operating condition. The traditional power switching method of the hybrid commercial vehicle has various limitations, firstly, the information acquisition dimension is single, only the parameters of a power source or partial working condition indexes are often concerned, the key information such as load, road surface, component loss and task mode cannot be fully integrated, the perception of the running state and task mode of the system is not comprehensive, further, the power mode adaptation decision lacks sufficient data support, secondly, a fixed weight or single target optimization strategy is mostly adopted in the decision process, the relation of dynamic balance energy efficiency, loss and response speed is difficult to change according to the actual working condition, the situation that a certain performance index is emphasized and other indexes are ignored easily occurs, multi-target cooperative optimization cannot be realized, furthermore, the prejudging capability of future working conditions is lacking, only the current working condition change can be passively responded, the power switching is lagged behind the actual requirement, the problem of unsmooth power connection easily occurs, meanwhile, the compensation mechanism of the component loss is not perfect, the running mode cannot be timely adjusted in combination with the task mode and the loss state, the excessive loss of the components is easy to cause, and the long-term stability and service life of the system are easily affected. Based on the above, in order to solve the above technical problems, the present disclosure provides an adaptive switching method of a multi-mode power system of a hybrid commercial vehicle. Disclosure of Invention The invention aims to make up the defects of the prior art and provides a self-adaptive switching method of a multi-mode power system of a hybrid commercial vehicle, wherein load, road surface, power source, component loss and task mode multidimensional information are collected through a sensor and an interface, and a data set is constructed after preprocessing; dividing a five-dimensional parameter quantization interval based on a data set, generating a power mode candidate set by coupling a proper matching algorithm, then constructing a three-dimensional objective function, screening an optimal power mode by dynamic weight game, predicting future working conditions by using a prediction model, pre-starting, pre-closing or pre-adjusting parameters of a power source by combining a response correction algorithm, and finally monitoring component loss in real time to trigger a compensation mechanism to adjust an operation mode. The invention provides a self-adaptive switching method of a multi-mode power system of a hybrid commercial vehicle, which aims to solve the technical problems and comprises the following specific steps: S100, multidimensional information acquisition, namely acquiring load, road surface, power source, component loss and task mode information through a sensor and an interface, preprocessing and extracting features, and constructing a data set containing running states and task attributes; S200, constructing a five-dimensional model, namely dividing a five-dimensional parameter quantization interval containing a task mode, total load, road surface charact