大规格尖端铝合金锻件温度场协同解耦控制

Cooperative decoupling control of temperature field of large-size and cutting-edge aluminum alloy forgings

中文关键词  大型铝合金锻件  大尺度温度场  有限元法  外推法  智能解耦

英文关键词  large-size aluminum  large scale temperature field  finite element method  extrapolation  intelligent decou pling

中文摘要

      大型立式淬火炉炉内温度是决定大规格尖端铝合金锻件性能的主要因素.为解决传统淬火炉温度控制方 法无法克服多区耦合换热影响实现大尺度温度场均匀性控制目标的问题,本文首先构建了炉内温度场模型,提出时 间维有限元外推方法,利用有限元方法计算不同时间步长下的温度结果结合推得的时间维外推公式,实现铝合金锻 件温度高精度高时效预测;随后,提出特征向量快速聚类与竞争学习融合的变结构神经网络,结合预测温度,实时调 节控制器参数,实现多区协同解耦控制;最后,通过仿真实验,验证了所提方法温度控制误差低于±2?C.

英文摘要

      The temperature inside a large vertical quenching furnace is the primary factor determining the performance of large-scale advanced aluminum alloy forgings. To address the issue that traditional quenching furnace temperature con trol methods cannot overcome the effects of multi-zone coupled heat transfer, thereby failing to achieve the objective of uniform temperature field control over large scales, this paper first constructs a temperature field model inside the furnace. It proposes a time-domain finite element extrapolation method, which uses the finite element method to calculate tem perature results at different time steps. By combining these results with the derived time-domain extrapolation formula, it achieves high-precision, high-efficiency temperature prediction for aluminum alloy forgings. Subsequently, it intro duces a variable-structure neural network integrating fast clustering of characteristic vectors and competitive learning. This network, combined with the predicted temperatures, adjusts the controller parameters in real time to achieve multi-zone coordinated decoupling control. Finally, simulation experiments verify that the proposed method achieves a temperature control error of less than ±2?C.