CN-121978955-A - Gardening multifunctional working machine working condition self-adaptive energy management system based on multi-source information

CN121978955ACN 121978955 ACN121978955 ACN 121978955ACN-121978955-A

Abstract

The invention discloses a gardening multifunctional working machine working condition self-adaptive energy management system based on multi-source information, which relates to the technical field of intelligent agricultural machinery equipment and comprises the following modules: the multisource perception fusion module is used for integrating an electronic identification interface and a multisource field sensor array, constructing a working condition identification system for multisource information fusion, and accurately identifying working modes of the gardening multifunctional working machine under corresponding working conditions, including pesticide spraying, rotary tillage and mowing. According to the invention, by comprehensively utilizing the sensor array, the intelligent recognition and the machine learning method, the operation mode can be rapidly and accurately recognized under complex and changeable environmental conditions, high reliability can be maintained in severe environments such as high humidity, dustiness, strong electromagnetic interference and the like, misjudgment or delay caused by the fact that a traditional single sensor is easy to be interfered is effectively avoided, and the energy management instruction is ensured to timely and accurately respond to the actual operation requirement.

Inventors

Liu Haolu
ZHAO SHANHU
ZHANG YANHUA
Cao Guangqiao
SHEN CHENG
HU LIANGLONG

Assignees

农业农村部南京农业机械化研究所

Dates

Publication Date: 20260505
Application Date: 20260209

Claims (10)

1. The gardening multifunctional working machine working condition self-adaptive energy management system based on the multi-source information is characterized by comprising the following modules: The multisource perception fusion module is used for integrating an electronic identification interface and a multisource field sensor array, constructing a working condition identification system for multisource information fusion, and accurately identifying working modes of the gardening multifunctional working machine under corresponding working conditions, including pesticide spraying, rotary tillage and mowing; the dynamic load analysis module is used for carrying out on-line reconstruction and self-adaptive decomposition on a load spectrum by fusing real-time topographic information and load data, describing the nonlinear distortion characteristic of the load caused by topographic fluctuation and providing a load spectrum reference matched with a power distribution strategy; The mode self-adaptive control module is used for calling corresponding control strategies according to the identified different operation modes and optimizing control parameters of all the control strategies according to the real-time system state by utilizing the reinforcement learning agent; the thermal-electric cooperative management module is used for establishing an equivalent thermal model of the super capacitor, estimating the internal temperature rise of the super capacitor in real time, and optimizing a charge-discharge strategy by combining a drosophila algorithm to realize the cooperation of thermal management and energy distribution; And the collaborative optimization decision module is used for fusing NSGA-II with a drosophila algorithm, optimizing control parameters and energy distribution thresholds on line, and dynamically adjusting strategy weights by combining reinforcement learning to predict the health states of the battery and the super capacitor.
2. The working condition self-adaptive energy management system of the gardening multifunctional working machine based on the multi-source information according to claim 1, wherein the multi-source perception fusion module comprises an intelligent mounting identification unit and an environment adaptation perception unit; The intelligent mounting identification unit realizes plug and play and identity automatic identification of farm tools based on a standard communication protocol, optimizes an identification signal transmission process by introducing a drosophila algorithm, and dynamically adjusts communication parameters to resist instantaneous interference; The environment adaptation sensing unit is used for collecting the running state and external environment data of the gardening multifunctional operation machine in real time through the deployed multi-physical-field sensor array, and performing feature extraction and real-time mode identification on the sensor data by adopting an operation mode classification model to be used as effective supplement and verification of intelligent mounting identification.
3. The system for adaptively managing working conditions of a multifunctional gardening working machine based on multi-source information according to claim 2, wherein the intelligent mounting identification unit comprises the following steps: An ISO 11783 standard communication protocol based on a CAN bus is used for deploying a standard electronic identification interface at a farm tool mounting point, and when an intelligent farm tool is hung, the type and parameters of the intelligent farm tool are actively reported to a main controller, so that the automatic reporting of the plug and play and identity information of the farm tool is realized; A drosophila algorithm is introduced in the process of identifying signal transmission, the communication baud rate, the data frame retransmission interval and a verification mechanism are dynamically optimized, instantaneous electromagnetic interference and signal attenuation are resisted, and communication parameters are fed back and adjusted in real time by continuously monitoring signal integrity and error rate.
4. The system for managing the working condition self-adaptive energy of the multifunctional gardening working machine based on the multi-source information according to claim 2, wherein the execution step of the environment adaptation sensing unit comprises the following steps: disposing a multi-physical field sensor array on the machine body, and collecting current, vibration, inclination angle, humidity and temperature data in real time to form multi-source information input; Real-time judgment is carried out by utilizing a pre-trained operation mode classification model, and identification and classification of operation modes are carried out, namely operation modes including spraying medicine, rotary tillage and mowing are distinguished, wherein a lightweight convolutional neural network is adopted to extract time sequence characteristics of original sensor data, and total power requirement is selected Rate of change of power And working shaft torque standard deviation As a characteristic vector, carrying out operation mode classification and identification through a support vector machine; and cross-verifying the classification result and the identification result of the intelligent mounting identification unit, and if the classification result and the identification result are inconsistent, starting a redundancy decision mechanism, so as to output an accurate operation mode.
5. The multi-source information-based gardening multifunctional working machine working condition self-adaptive energy management system according to claim 2, wherein the dynamic load analysis module comprises a frequency spectrum self-adaptive decomposition unit and a terrain coupling modeling unit; The frequency spectrum self-adaptive decomposition unit is used for optimizing cut-off frequency band selection by adopting wavelet packet decomposition and reinforcement learning, dynamically adjusting a frequency domain segmentation strategy and adapting to frequency spectrum structural variation caused by soil hardness, grass density and gradient change in the operation modes of spraying, rotary tillage and mowing in real time; The terrain coupling modeling unit establishes a load-terrain coupling model based on real-time gestures and terrain information, predicts the variation trend and fluctuation intensity of load power under specific terrain in real time, quantizes the terrain factors into adjustment coefficients for load spectrums, and performs feedforward compensation analysis on load characteristics.
6. The system for managing working condition-adaptive energy of a multifunctional gardening working machine based on multi-source information according to claim 5, wherein the executing step of the spectrum adaptive decomposition unit comprises the following steps: Based on a load power signal acquired in real time, decomposing the load power signal into a plurality of sub-band signals by adopting a wavelet packet decomposition method, and establishing a time-frequency representation of a load frequency spectrum; combining the reinforcement learning intelligent agent, dynamically optimizing cutoff frequency band parameters and layer number setting of wavelet packet decomposition according to the current operation mode and terrain feedback information; and reconstructing the low-frequency load component and the high-frequency load component in real time to form a dynamic spectrum reference matched with the current working condition.
7. The system for adaptively managing working conditions of a multifunctional gardening working machine based on multi-source information according to claim 5, wherein the executing step of the terrain coupling modeling unit comprises the following steps: acquiring real-time attitude and topographic relief data through an inertial measurement unit and a gradient sensor, and establishing a load-topographic coupling model of a topographic-load coupling relation; taking the terrain gradient and the ground unevenness as input variables, predicting the variation trend and the fluctuation intensity of the load power by combining a load-terrain coupling model, and calculating a load frequency spectrum adjustment coefficient; the load spectrum adjusting coefficient is fed forward to a spectrum self-adaptive decomposition process, and the load spectrum is compensated and reconstructed in real time so as to reflect the load distortion characteristic caused by actual topography.
8. The system for managing working condition adaptive energy of a multifunctional gardening working machine based on multi-source information according to claim 5, wherein the executing step of the mode adaptive control module comprises the following steps: According to the working condition identification result, invoking a power distribution control algorithm corresponding to the rotary tillage, mowing or spraying modes from a strategy library; Executing a frequency domain power splitting strategy based on wavelet packet decomposition in a rotary tillage mode, adopting a mixed power supply strategy based on a fuzzy self-adaptive threshold in a mowing mode, and implementing a super capacitor dominant power filter strategy in a pesticide spraying mode, wherein the function of the super capacitor is defined as a main power supply and a power filter in the mode; And the reinforcement learning agent is utilized to monitor the state variables of the system in real time, dynamically optimize key parameters in each control strategy and continuously match the control behavior and the load characteristic.
9. The system for adaptively managing working conditions of a multifunctional gardening working machine based on multi-source information according to claim 8, wherein the executing step of the thermo-electric cooperative management module comprises the following steps: Establishing an equivalent thermal model of a super capacitor carried by the gardening multifunctional working machine, and estimating the internal temperature rise and the temperature change rate of the super capacitor based on real-time current, voltage and environmental temperature; taking the internal temperature rise and the temperature change rate of the super capacitor as the optimizing input of a drosophila algorithm, and dynamically optimizing the limit value of charge and discharge current, the working voltage window and the power distribution proportion; On the premise of meeting the instantaneous power demand, a power distribution scheme with the smallest thermal stress is preferentially selected, and the thermal management and energy distribution are cooperatively optimized.
10. The system for adaptively managing working conditions of a multifunctional gardening working machine based on multi-source information according to claim 9, wherein the executing step of the collaborative optimization decision-making module comprises the following steps: Adopting a multi-objective optimization mechanism fused by NSGA-II and a drosophila algorithm to iteratively optimize control parameters and energy distribution thresholds of each operation mode on line; Constructing a health state prediction model by combining reinforcement learning, predicting the health states of the battery and the super capacitor, and dynamically adjusting the weight coefficients of the service life and the energy efficiency in the optimization target; based on the real-time system state and element degradation trend, strategy decision logic is continuously updated, and performance balance and self-adaptive evolution in the whole life cycle are realized.

Description

Gardening multifunctional working machine working condition self-adaptive energy management system based on multi-source information Technical Field The invention relates to the technical field of intelligent agricultural machinery equipment, in particular to a working condition self-adaptive energy management system of a gardening multifunctional working machine based on multi-source information. Background The application of gardening multifunctional working machines in the fields of agriculture and gardening is increasingly wide, along with the diversification of working requirements and continuous progress of equipment technology, working condition self-adaptive energy management becomes an important direction for improving the efficiency and performance of the multifunctional working machines, currently, the energy utilization efficiency of traditional working machines under different working conditions is often limited, the multifunctional working machines cannot effectively adapt to environmental changes, working types and load changes, so that energy waste and working effects are poor, and the energy output and the actual requirements of the working machines are dynamically adjusted through real-time monitoring of working environments, load conditions and equipment performances, so that the optimal energy configuration can be realized. In actual operation of a gardening multifunctional operation machine, particularly in a high-humidity and dusty complex environment, the traditional working condition identification method based on a fixed sensor is easy to be interfered by the environment, so that identification delay or misjudgment is caused, the instantaneity and accuracy of energy management are affected, in the fluctuation topography of a mountain area, a sloping field and the like, nonlinear distortion can occur to a load frequency spectrum under the same operation mode due to the change of the topography, so that a power distribution strategy based on fixed frequency domain segmentation is difficult to adapt to actual load characteristics, the dynamic performance of a system is affected, meanwhile, in a continuous long-time operation scene of a remote park, the super capacitor is easy to generate temperature rise and capacity attenuation due to continuous high-frequency charge and discharge, the energy distribution strategy is gradually disabled in the later stage, and the system stability and energy efficiency are continuously reduced. Disclosure of Invention In order to solve the technical problems, the invention is realized by the following technical scheme that the working condition self-adaptive energy management system of the gardening multifunctional working machine based on multi-source information comprises the following modules: The multisource perception fusion module is used for integrating an electronic identification interface and a multisource information fusion sensor array, constructing a multisource information fusion working condition identification system, improving identification robustness and instantaneity in a high-humidity and dusty environment, accurately identifying operation modes of the gardening multifunctional operation machine under corresponding working conditions, including spraying medicines, rotary tillage and mowing, remarkably improving working condition identification accuracy in a complex environment, and guaranteeing system response instantaneity and control reliability; The dynamic load analysis module is used for carrying out on-line reconstruction and self-adaptive decomposition on a load spectrum by fusing real-time topographic information and load data, describing the nonlinear distortion characteristic of the load caused by topographic fluctuation, providing a load spectrum reference matched with a power distribution strategy and ensuring that the energy management strategy is highly consistent with an actual physical process; The mode self-adaptive control module is used for calling corresponding control strategies according to the identified different operation modes, optimizing control parameters of all the control strategies according to the real-time system state by utilizing the reinforcement learning agent, realizing on-line self-optimization of the control parameters, and improving the energy efficiency of the system and the service life of elements under each operation mode; the thermal-electric cooperative management module is used for establishing an equivalent thermal model of the super capacitor, estimating the internal temperature rise of the super capacitor in real time, optimizing a charge-discharge strategy by combining a drosophila algorithm, realizing the cooperation of thermal management and energy distribution, and prolonging the service life of the element; The collaborative optimization decision module is used for fusing NSGA-II and a drosophila algorithm, optimizing control parameters and energy distribution threshold values