An electric agricultural machine working condition recognition and energy management method and system

Abstract

The present application relates to the technical field of intelligent control of electric agricultural equipment, in particular to an electric agricultural machine operation working condition recognition and energy management method and system, the method comprising: constructing multi-modal time series data; generating operation working condition semantic information; generating continuous adjustment strategy of power system; and adaptively coordinating and controlling output torque of driving motor of operation execution mechanism and walking speed of whole machine. The present application realizes accurate recognition of complex agricultural operation working conditions, improves reliability and robustness of working condition recognition; the present application realizes continuous coordination and adjustment of torque of operation execution mechanism and walking speed of whole machine, avoiding energy waste caused by traditional discrete control strategy; and the present application constructs a closed-loop control structure of operation state recognition-decision optimization-execution feedback, so that the system can dynamically and adaptively respond to load mutation and operation environment change, improving safety and stability of system operation.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for electric agricultural equipment, specifically to a method and system for identifying the operating conditions and managing the energy of electric agricultural machinery. Background Technology

[0002] With the continuous improvement of electrification and intelligence in agricultural machinery, electric agricultural machinery, represented by electric rotary tillers, electric tractors, electric lawnmowers, and intelligent plant protection machinery, is gradually being widely used in typical agricultural operation scenarios such as greenhouses, hilly areas, and small plots, thanks to its advantages such as zero emissions, low noise, compact structure, and strong low-speed, high-torque output capability. Compared with traditional fuel-powered agricultural machinery, electric agricultural machinery places higher demands on the precision of power system control, energy utilization efficiency, and adaptability to complex operating environments.

[0003] In actual operation, electric agricultural machinery faces significant randomness and time-varying characteristics in terms of the physical properties of the work objects and the operational behaviors. For example, the soil entry and stable tillage of electric rotary tillers under different soil moisture contents, the resistance fluctuations of electric tractors when traction is performed on heterogeneous surfaces, and the changes in the blade load of electric lawnmowers under different vegetation densities all cause the operating mechanisms to exhibit strong nonlinearity, strong time-varying characteristics, and abrupt load changes. How to accurately identify the current operating conditions of agricultural machinery and adjust the power output accordingly is a key issue affecting its operating efficiency, operational safety, and range.

[0004] Existing methods for identifying the operating conditions of electric agricultural machinery mostly rely on single or a few sensor signals (such as motor current or speed) set with fixed thresholds for judgment. However, in complex agricultural operating environments, a single signal is insufficient to accurately distinguish between operating conditions such as sudden load changes, local overload, and operational stagnation. Relying solely on threshold judgments can easily lead to identification delays or misjudgments. Furthermore, the mechanical vibrations, electromagnetic interference, and inconsistent sampling frequencies of multiple sensors commonly found in agricultural operations make it difficult for simple statistical features to fully reflect the true evolution of the load.

[0005] In terms of energy management, existing equipment mostly adopts power regulation methods based on rules or proportional-integral-derivative (PID) control. The output power control commands exhibit discrete or hierarchical characteristics, making it difficult to achieve continuous and precise dynamic regulation. This approach is prone to power redundancy and energy waste under light loads, while under heavy loads or abnormal operating conditions, it is prone to insufficient response or overload losses, reducing the overall energy utilization efficiency of the machine and affecting the service life of the power system.

[0006] Therefore, there is an urgent need for an energy management method and system that can accurately identify various operating conditions of electric agricultural machinery in complex environments and effectively transform the identification results into a continuous power regulation strategy. Summary of the Invention

[0007] To address the problems of low accuracy in identifying the operating conditions of electric agricultural machinery, insufficient connection between the identification results and energy management strategies, and discrete power adjustment methods and lag in response in existing technologies, this invention proposes a method and system for identifying the operating conditions and managing the energy of electric agricultural machinery.

[0008] The technical problem to be solved by this invention is achieved by the following technical solution: A method for identifying the operating conditions and managing the energy of electric agricultural machinery includes the following steps: Step S101: Collect multi-source operation data reflecting the operation status of electric agricultural machinery, preprocess the multi-source operation data, normalize and align the time axis of the preprocessed multi-source operation data, and construct multimodal time series data that can reflect the synchronous change characteristics during the operation of electric agricultural machinery. Step S102: Based on multimodal time series data, analyze the time series characteristics of load changes of the work execution mechanism through the work condition identification method, and generate work condition semantic information to characterize different work states. Step S103: Integrate the semantic information of the working condition with the machine state parameters that represent the current operating state of the electric agricultural machinery to construct a system state vector, and input the system state vector into the continuous action space decision model to generate a continuous adjustment strategy for the power system. Step S104: Based on the continuous adjustment strategy, adaptive and coordinated control is performed on the output torque of the drive motor of the work execution mechanism and the overall walking speed to achieve dynamic optimization management of energy during the operation.

[0009] As a further improvement of the present invention, the multi-source job data in step S101 includes: Electrical characteristic data, used to characterize power output characteristics, includes the current and voltage of the drive motor; Dynamic data, used to characterize changes in work load, includes the rotational speed and vibration of the working components; Motion parameter data is used to characterize the overall machine operating status. Motion parameter data includes the overall machine travel speed and working depth.

[0010] As a further improvement of the present invention, the preprocessing of multi-source operation data in step S101 includes: detecting and removing abnormal data, and interpolating and repairing missing or interrupted data.

[0011] As a further improvement of the present invention, the work condition identification method in step S102 includes: Based on multimodal time series data, local time series features reflecting load changes of the operation execution mechanism are extracted to obtain the first time series feature characterizing the transient change characteristics during the operation. A time-series correlation analysis was performed on the first time-series feature to obtain a second time-series feature that reflects the load change trend of the work execution mechanism. Based on the contribution of different time-series features to the judgment of work status, key time-series features that reflect sudden or abnormal load changes are given higher weights.

[0012] As a further improvement of the present invention, the expression of the work condition semantic information in step S102 is as follows: ; In the formula, s ( t () represents semantic information about the working conditions; F (·) represents the time-series feature mapping function used to extract the dynamic evolution characteristics of the job load; f 1 ( t ) is the first time-series feature; f 2 ( t ) is the second time series feature; α and β These are the corresponding weight coefficients, and they satisfy... α + β =1.

[0013] As a further improvement of the present invention, the semantic information of the working condition in step S102 is used to identify the current working state of the electric agricultural machinery. The identification results include no-load working condition, steady-state working condition, transient load working condition, abnormal stagnation working condition and overload working condition.

[0014] An electric agricultural machinery operation condition identification and energy management system is applied to the above-mentioned electric agricultural machinery operation condition identification and energy management method, including an operation condition identification system and an energy management system; The work condition identification system includes: The multimodal data acquisition module is used to collect multi-source operation data that reflects the operating status of electric agricultural machinery in real time; The data processing module is used to perform abnormal data removal, data repair, scale unification and time axis alignment on the acquired multi-source operation data to construct multimodal time series data; The working condition identification module is used to analyze multimodal time series data based on a deep temporal feature extraction network and generate semantic information of working conditions. The energy management system includes: The decision optimization module is used to input the semantic information of the working conditions and the machine state parameters into the continuous motion space decision model to generate the power system adjustment strategy. The execution control module is used to coordinate and control the output torque of the work actuator and the overall travel speed of the machine according to the power system adjustment strategy.

[0015] The beneficial effects of this invention are: This invention achieves accurate identification of complex agricultural operation conditions by performing in-depth time-series feature analysis on multi-source operation data, thereby improving the reliability and robustness of operation condition identification. This invention, based on a dynamic optimization decision model in continuous motion space, realizes continuous coordinated adjustment of the torque of the work execution mechanism and the overall travel speed, avoiding the energy waste caused by traditional discrete control strategies. This invention constructs a closed-loop control structure of job status identification, decision optimization, and execution feedback, enabling the system to dynamically and adaptively respond to sudden load changes and changes in the working environment, thereby improving the safety and stability of system operation. Attached Figure Description

[0016] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the system structure of the present invention.

[0017] In the picture: 100. Operation condition recognition system; 110. Multimodal data acquisition module; 120. Data processing module; 130. Operation condition recognition module; 200. Energy Management System; 210. Decision Optimization Module; 220. Execution Control Module; 300. Electric rotary tiller; 310. Rotary tiller motor; 320. Walking drive system. Detailed Implementation

[0018] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0019] The present invention proposes a method and system for identifying the operating conditions and managing the energy of electric agricultural machinery, the application of which is not limited to a specific type of agricultural machinery. The electric agricultural machinery includes, but is not limited to, agricultural machinery devices with electric drive and operation execution mechanisms, such as electric rotary tillers, electric tractors, electric lawn mowers, electric sprayers, and electric harvesters.

[0020] These different types of electric agricultural machinery all face similar load fluctuations and energy management needs in actual operation. For example, the resistance changes of electric tractors during traction operations under different geological conditions, the blade load fluctuations of electric lawnmowers under different vegetation densities, and the power distribution of electric sprayers in complex terrain can all be optimized for energy using the working condition identification and continuous adaptive adjustment method described in this invention.

[0021] like Figure 1 As shown, this embodiment uses an electric rotary tiller as a specific application of electric agricultural machinery, and provides a method for identifying the operating conditions and managing the energy of electric agricultural machinery. It includes the following steps: Step S101: Collect multi-source operation data reflecting the operation status of the electric rotary tiller using various sensors installed on the electric rotary tiller. Preprocess the multi-source operation data, then normalize and align the preprocessed data along the time axis to eliminate sampling differences between different operation signals, constructing multimodal time-series data that reflects the synchronous changes during the operation of the electric rotary tiller.

[0022] The multi-source operational data mentioned above includes electrical characteristic data, dynamic data, and motion parameter data. The electrical characteristic data characterizes the power output characteristics and includes at least the current and voltage information of the drive motor. The dynamic data characterizes changes in the operational load and includes at least the rotational speed and vibration information of the working components. The motion parameter data characterizes the overall machine's operational status and includes at least the machine's travel speed and working depth information.

[0023] The preprocessing of multi-source operation data specifically involves: detecting and removing abnormal data, and interpolating and repairing missing or interrupted data to improve the continuity and integrity of the data.

[0024] Furthermore, in order to reduce the impact of environmental interference on data quality, the dynamic data and electrical characteristic data are subjected to noise reduction processing.

[0025] Multi-source job data was obtained by normalizing the data. , The expression is as follows: ; in, These are the normalized eigenvalues. x max and x min These represent the maximum and minimum values ​​of the corresponding features in the sample database, respectively.

[0026] right Time axis alignment is performed to construct multimodal time-series data that can reflect the synchronous changes of various physical quantities during rotary tillage operations.

[0027] Step S102: Based on multimodal time series data, analyze the time series characteristics of load changes of the work execution mechanism through the work condition identification method, and generate work condition semantic information to characterize different work states.

[0028] Specifically, the methods for identifying operating conditions include: Based on multimodal time series data, local time series features reflecting load changes of the operation execution mechanism are extracted to obtain the first time series feature characterizing the transient change characteristics during the operation. The first time series feature is used to reflect short-term operation state changes such as rotary tillage into the soil and sudden load changes.

[0029] A time-series correlation analysis was performed on the first time-series feature to obtain a second time-series feature that reflects the load change trend of the operation execution mechanism. The second time-series feature is used to characterize the steady-state tillage stage and the continuous impact of changes in the operating environment on the rotary tillage load.

[0030] Furthermore, in the process of generating semantic information about work conditions, key temporal features reflecting sudden or abnormal load changes are assigned higher weights based on their contribution to work status discrimination. This enhances the representation of typical work states such as abnormal stalls and overloads. The mathematical expression for this calculation is shown below: ; In the formula, s ( t () represents semantic information about the working conditions; F (·) represents the time-series feature mapping function used to extract the dynamic evolution characteristics of the job load; f 1 ( t ) is the first time-series feature; f 2 ( t ) is the second time series feature; α and β These are the corresponding weight coefficients, and they satisfy... α + β =1. The weighting coefficient is set or adaptively adjusted according to the characteristics of changes in the work status.

[0031] The high-level state representation, which is composed of the first temporal feature, the second temporal feature and their weight relationship, is defined as the semantic information of the working condition, which is used to comprehensively characterize the dynamic evolution of the rotary tillage load in the time dimension.

[0032] The aforementioned processes of extracting local temporal features, performing temporal correlation analysis, and assigning feature weights can be achieved through a deep temporal feature extraction model that combines convolutional neural networks, bidirectional recurrent neural networks, and attention mechanisms.

[0033] Step S103: Integrate the semantic information of the working condition with the machine state parameters that represent the current operating state of the electric agricultural machinery to construct a system state vector, and input the system state vector into the continuous action space decision model to generate a continuous adjustment strategy for the power system.

[0034] Specifically, based on the semantic information of the operating conditions, the current operating conditions of the electric rotary tiller are identified. These operating conditions include at least no-load conditions, steady-state operating conditions, transient load conditions, abnormal stall conditions, and overload conditions. After identifying the operating conditions, the semantic information is used as the primary basis for power adjustment and is fused with machine state parameters characterizing the operating state of the power system and the overall machine operating state to construct a system state vector.

[0035] The system state vector is input into a continuous motion space decision model. Based on the comprehensive optimization objectives of optimal energy consumption, maximized work efficiency, and system safety, a continuous adjustment strategy for the rotary tiller motor output torque and the overall machine travel speed is generated. The continuous motion space decision model can be implemented using continuous decision-making methods, including reinforcement learning methods based on policy learning. The cost function for the comprehensive optimization objective is... J As shown below: ; in, J For the target result, P total The total power consumption of the system is η This is an indicator of work efficiency. S safe As a security constraint penalty item, w 1, w 2, w 3 represents the optimization weight coefficients for each objective.

[0036] In abnormal stagnation or overload conditions, the continuous action space decision model increases the weight of safety-related constraints in the decision-making process, causing the generated adjustment strategy to tend to reduce power output or limit walking speed, so as to avoid the power system being in an overloaded operating state.

[0037] Step S104: Based on the continuous adjustment strategy, adaptive and coordinated control is performed on the output torque of the drive motor of the work execution mechanism and the overall walking speed to achieve dynamic optimization management of energy during the operation.

[0038] Specifically, according to the continuous adjustment strategy, the output torque of the rotary tiller motor 310 and the speed of the walking drive system 320 are coordinated and controlled, so that the power output can be continuously adjusted according to the changes in the working conditions, realizing dynamic optimization management of energy during the rotary tillage operation. The mapping formula of the actuator control command is defined as follows: in, A ( t ) is the action vector. T m ( t () represents the motor torque command.V f ( t () represents the walking speed command. G ( s ( t ), m ( t )) is the nonlinear mapping function of the decision model. m ( t ) represents the current machine status parameters.

[0039] The target torque adjustment and target travel speed adjustment generated by the continuous adjustment strategy are used by the execution control unit to adjust the rotary tiller motor and travel system in real time. After execution control, new operation status data are collected again and fed back to the operation condition identification and decision-making module, forming a closed-loop control process of data acquisition—operation condition identification—strategy generation—execution feedback, enabling the system to dynamically respond to sudden load changes and changes in the operating environment during rotary tillage operations.

[0040] like Figure 2 As shown, an electric agricultural machinery operation condition identification and energy management system is applied to the above-mentioned electric agricultural machinery operation condition identification and energy management method, including an operation condition identification system 100 and an energy management system 200.

[0041] The work condition identification system 100 is used to perform in-depth processing and temporal feature analysis on the work status data S1 to generate high-level work condition semantic information S2 that represents the current work status.

[0042] The work condition identification system 100 includes a multimodal data acquisition module 110, a data processing module 120, and a work condition identification module 130.

[0043] The multimodal data acquisition module 110 is used to uniformly collect operational status data S1 from different sensing devices and provide standardized basic data input for subsequent processes.

[0044] The data processing module 120 is used to clean and synchronize the collected data, specifically including: detecting and removing abnormal data, interpolating and repairing missing data, unifying the scale of different physical quantities, and aligning the time axis of multi-source data. The processed data yields multimodal operational data that synchronously reflects the dynamic characteristics of the system.

[0045] The working condition identification module 130 is used to analyze multimodal time series data based on a deep temporal feature extraction network to generate working condition semantic information.

[0046] By comprehensively analyzing the local temporal characteristics and their evolution trends in the operation data, the operation condition identification module 130 can generate operation condition semantic information S2 to describe the current operation state of the electric rotary tiller. The operation condition semantic information S2 reflects the operation condition type of the electric rotary tiller and its changing trend. The generated operation condition semantic information S2 is output to the energy management system 200.

[0047] The energy management system 200 is responsible for converting the perceived semantic state into specific power regulation strategies to achieve dynamic optimization of energy utilization. The energy management system 200 includes a decision optimization module 210 and an execution control module 220.

[0048] The decision optimization module 210 is used to receive the working condition semantic information S2, and input the working condition semantic information S2 and the machine status parameters fed back in real time by the electric rotary tiller 300 into the continuous motion space decision model, with the goals of optimal energy consumption, maximum working efficiency and system safety, to generate a power system adjustment strategy.

[0049] The execution control module 220 is used to convert the adjustment strategy into specific control commands S3 and send them to the actuator. The control commands S3 include at least: a torque adjustment command acting on the rotary tiller motor 310 and a speed adjustment command acting on the walking drive system 320.

[0050] Through the above structure, this embodiment forms a closed-loop control process with the main line of operation status data S1-operation condition semantic information S2-control command S3, realizing: real-time identification of the operation condition of the electric rotary tiller; effective transmission of operation condition information between the perception layer and the energy management layer; and dynamic optimization and adjustment of the power system output according to the operation condition.

[0051] This embodiment can improve the energy utilization efficiency of electric rotary tillers in complex agricultural operating environments and reduce the risk of power system load under abnormal operating conditions.

[0052] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely prisms of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for identifying the operating conditions and managing the energy of electric agricultural machinery, characterized in that: Includes the following steps: Step S101: Collect multi-source operation data reflecting the operation status of electric agricultural machinery, preprocess the multi-source operation data, normalize and align the time axis of the preprocessed multi-source operation data, and construct multimodal time series data that can reflect the synchronous change characteristics during the operation of electric agricultural machinery. Step S102: Based on multimodal time series data, analyze the time series characteristics of load changes of the work execution mechanism through the work condition identification method, and generate work condition semantic information to characterize different work states. Step S103: Integrate the semantic information of the working condition with the machine state parameters that represent the current operating state of the electric agricultural machinery to construct a system state vector, and input the system state vector into the continuous action space decision model to generate a continuous adjustment strategy for the power system. Step S104: Based on the continuous adjustment strategy, adaptive and coordinated control is performed on the output torque of the drive motor of the work execution mechanism and the overall walking speed to achieve dynamic optimization management of energy during the operation.

2. The method for identifying the operating conditions and managing the energy of electric agricultural machinery according to claim 1, characterized in that: The multi-source job data in step S101 includes: Electrical characteristic data, used to characterize power output characteristics, includes the current and voltage of the drive motor; Dynamic data, used to characterize changes in work load, includes the rotational speed and vibration of the working components; Motion parameter data is used to characterize the overall machine operating status. Motion parameter data includes the overall machine travel speed and working depth.

3. The method for identifying the operating conditions and managing the energy of electric agricultural machinery according to claim 1, characterized in that: The preprocessing of multi-source operation data in step S101 includes: detecting and removing abnormal data, and interpolating and repairing missing or interrupted data.

4. The method for identifying the operating conditions and managing the energy of electric agricultural machinery according to claim 1, characterized in that: The job condition identification method in step S102 includes: Based on multimodal time series data, local time series features reflecting load changes of the operation execution mechanism are extracted to obtain the first time series feature characterizing the transient change characteristics during the operation. A time-series correlation analysis was performed on the first time-series feature to obtain a second time-series feature that reflects the load change trend of the work execution mechanism. Based on the contribution of different time-series features to the judgment of work status, key time-series features that reflect sudden or abnormal load changes are given higher weights.

5. The method for identifying the operating conditions and managing the energy of electric agricultural machinery according to claim 1, characterized in that: The expression for the semantic information of the work condition in step S102 is as follows: ； In the formula, s ( t () represents semantic information about the working conditions; F (·) represents the time-series feature mapping function used to extract the dynamic evolution characteristics of the job load; f 1 ( t ) is the first time-series feature; f 2 ( t ) is the second time series feature; α and β These are the corresponding weight coefficients, and they satisfy... α + β =1.

6. The method for identifying the operating conditions and managing the energy of electric agricultural machinery according to claim 1, characterized in that: The semantic information of the working condition in step S102 is used to identify the current working state of the electric agricultural machinery. The identification results are no-load working condition, steady-state working condition, transient load working condition, abnormal stagnation working condition and overload working condition.

7. An electric agricultural machinery operating condition identification and energy management system, characterized in that: The method for identifying the operating conditions and managing the energy of an electric agricultural machine, as described in any one of claims 1 to 6, includes an operating condition identification system (100) and an energy management system (200). The work condition identification system (100) includes: The multimodal data acquisition module (110) is used to collect multi-source operation data reflecting the operation status of electric agricultural machinery in real time; The data processing module (120) is used to perform abnormal data removal, data repair, scale unification and time axis alignment on the acquired multi-source operation data to construct multimodal time series data; The working condition identification module (130) is used to analyze multimodal time series data based on a deep temporal feature extraction network and generate working condition semantic information. The energy management system (200) includes: The decision optimization module (210) is used to input the semantic information of the working condition and the machine state parameters into the continuous motion space decision model to generate the power system adjustment strategy; The execution control module (220) is used to coordinate and control the output torque of the work actuator and the overall travel speed according to the power system adjustment strategy.

Images