甘蔗收获机根部切割系统负载压力预测模型研究

Load Pressure Prediction Model for Sugarcane Harvester Base-cutting System

摘要:

为了提高甘蔗收获机切割深度控制系统的适用范围和准确度，针对当前参考压力设定无法根据土壤参数和机车参数自动调整的问题，建立了负载压力预测模型。通过正交试验方法对负载压力与入土切割深度、喂入量、土壤含水率、土壤坚实度之间的关系进行了数据采集，并将试验数据作为负载压力预测模型的训练样本和测试样本。根据训练样本建立极限学习机(ELM)和基于麻雀搜索算法优化的极限学习机(SSA-ELM)负载压力预测模型，并通过测试样本对预测模型进行性能评价。结果表明，与ELM模型相比，SSA-ELM预测模型平均绝对误差、平均相对误差和均方根误差在黄壤条件下降低50.00%、44.14%和44.44%，在红壤条件下降低58.33%、56.98%和57.14%。为了检验负载压力预测模型在实际收获过程中的适用性，在试验平台上模拟蔗地遇到的各种工况，将预测模型应用于现有控制系统进行试验。结果表明，当入土切割深度为20mm、作业速度为0.34m/s、刀盘转速为700r/min时，预测模型满足参考压力的设定要求，且切割深度与目标深度最大误差不大于5mm，满足甘蔗收获生产的实际要求。

Abstract:

Aiming to enhance the applicability and accuracy of the cutting depth control system for sugarcane harvesters, a load pressure prediction model was established to address the problem that the current reference pressure setting could not be automatically adjusted according to soil parameters and locomotive parameters. The relationship between the load pressure and the cutting depth into the soil, the feeding volume, the soil moisture content and the soil firmness was collected by orthogonal test methods, and the test data were used as the training samples and test samples of the load pressure prediction model. Based on the training samples, load pressure prediction models using extreme learning machine (ELM) and ELM based on sparrow search algorithm optimization (SSA-ELM)were established. Performance of the prediction model was evaluated by the test samples, and the results showed that compared with the ELM model, the mean absolute error, mean relative error and root-mean-square error of the SSA-ELM prediction model were reduced by 50.00%, 44.14% and 44.44% under the yellow soil condition, and reduced by 58.33%, 56.98% and 57.14% under red soil conditions. To verify the applicability of the load pressure prediction model in actual harvesting processes,various working conditions encountered in the cane field were simulated on the test platform, and the prediction model was applied to the existing control system for testing. The results showed that the prediction model met the setting requirements of the reference pressure when the cutting depth into the soil was 20mm, the operating speed was 0.34m/s, and the rotational speed of the cutter disc was 700r/min, and the maximum error between the cutting depth and the target depth was no more than 5mm, which met the actual requirements of sugarcane harvesting production.