基于POD-αATS的油浸变压器瞬态温升降阶自适应变步长计算方法

Adaptive Variable Step Size Calculation Method for Transient Temperature Rise and Fall of Oil Immersed Transformer Based on POD-αATS

针对油浸式电力变压器瞬态温升计算效率过低的问题,该文提出本征正交分解-αATS(proper orthogonal decomposition-adaptive time stepping based onαfactor,POD-αATS)降阶自适应变步长瞬态计算方法。首先,推导变压器绕组瞬态温升计算的有限元离散方程;其次,采用POD降阶算法改善传统瞬态计算中存在的条件数过大及方程阶数过高的问题;同时对于瞬态计算中的时间步长选择问题,提出适用于非线性问题的αATS变步长策略;然后,为验证方法的有效性,基于110 kV油浸式电力变压器绕组的基本结构建立二维八分区数值计算模型,同时将计算结果与基于110 kV绕组的温升实验结果进行对比。数值计算及实验结果表明,所提算法与全阶定步长算法在流场和温度场中的精度几乎相同,且流场计算效率提升约45倍,温度场计算效率提升约38倍,计算速度得到显著提高。这一点在温升实验中同样得到验证,说明该文所提算法的准确性、高效性及一定的工程实用性。

In response to the problem of low efficiency in calculating transient temperature rise of oil immersed power transformers,this paper proposes POD-αATS reduced order adaptive variable step size transient calculation method.First,the article briefly derives the finite element discrete equation for calculating the transient temperature rise of transformer windings.Next,the proper orthogonal decomposition(POD)order reduction algorithm is adopted to improve the problems of excessive number of conditions and equation orders in traditional transient calculations.Meanwhile,for the time step selection problem in transient calculations,this paper proposes a method suitable for nonlinear problemsαATS(adaptive time stepping based onαfactor,αATS)variable step size strategy.Then,in order to verify the effectiveness of the method,this paper establishes a two-dimensional eight zone numerical calculation model based on the basic structure of the 110 kV oil-immersed power transformer winding,and compares the calculation results with the temperature rise experimental results based on the 110 kV winding.The numerical calculation and experimental results show that the accuracy of the proposed algorithm is almost the same as that of the full order fixed step algorithm in the flow field and temperature field,the efficiency of flow field calculation is improved by about 45 times,and the efficiency of temperature field calculation is improved by about 38 times,significantly improving the calculation speed.This has also been verified in temperature rise experiments,which fully demonstrates the accuracy,efficiency,and certain engineering practicality of the algorithm proposed in this paper.