电力系统保护与控制
電力繫統保護與控製
전력계통보호여공제
POWER SYSTM PROTECTION AND CONTROL
2013年
12期
60-64
,共5页
丛伟%孙允%路庆东%哈恒旭
叢偉%孫允%路慶東%哈恆旭
총위%손윤%로경동%합항욱
电磁式 PT%励磁特性曲线%拟合%正交函数系%复合解法
電磁式 PT%勵磁特性麯線%擬閤%正交函數繫%複閤解法
전자식 PT%려자특성곡선%의합%정교함수계%복합해법
electromagnetic PT%excitation characteristic curve%fitting%orthogonal function system%comprehensive solving
电磁式 PT 铁芯的磁通-电流特性曲线通常基于试验测得的伏安特性数据,对待拟合参数进行最小二乘估计获得。由于试验样本数据较多,采用传统的牛顿法求解存在雅克比矩阵规模较大、受初始值影响较大、迭代计算可能无法收敛等问题。提出一种基于一般多项式和正交函数系的待拟合参数复合求解方法,使雅克比矩阵规模仅与待拟合参数的数量有关,与试验样本数据的数量无关,可有效控制求解迭代过程中雅克比矩阵的规模,从而提高了计算效率,保证了迭代的收敛性,使整个计算过程更加稳定。算例仿真验证了该复合求解方法能够简化计算过程,且具有较高的拟合精度,在试验数据量较多的情况下更能突出其优越性。
電磁式 PT 鐵芯的磁通-電流特性麯線通常基于試驗測得的伏安特性數據,對待擬閤參數進行最小二乘估計穫得。由于試驗樣本數據較多,採用傳統的牛頓法求解存在雅剋比矩陣規模較大、受初始值影響較大、迭代計算可能無法收斂等問題。提齣一種基于一般多項式和正交函數繫的待擬閤參數複閤求解方法,使雅剋比矩陣規模僅與待擬閤參數的數量有關,與試驗樣本數據的數量無關,可有效控製求解迭代過程中雅剋比矩陣的規模,從而提高瞭計算效率,保證瞭迭代的收斂性,使整箇計算過程更加穩定。算例倣真驗證瞭該複閤求解方法能夠簡化計算過程,且具有較高的擬閤精度,在試驗數據量較多的情況下更能突齣其優越性。
전자식 PT 철심적자통-전류특성곡선통상기우시험측득적복안특성수거,대대의합삼수진행최소이승고계획득。유우시험양본수거교다,채용전통적우돈법구해존재아극비구진규모교대、수초시치영향교대、질대계산가능무법수렴등문제。제출일충기우일반다항식화정교함수계적대의합삼수복합구해방법,사아극비구진규모부여대의합삼수적수량유관,여시험양본수거적수량무관,가유효공제구해질대과정중아극비구진적규모,종이제고료계산효솔,보증료질대적수렴성,사정개계산과정경가은정。산례방진험증료해복합구해방법능구간화계산과정,차구유교고적의합정도,재시험수거량교다적정황하경능돌출기우월성。
The flux-current characteristic curve of electromagnetic PT is usually obtained by estimating the fitting parameters based on the Least Square method and experimental U I- characteristic data. When the equation set is solved with traditional Newton method, the Jacobian matrix has a larger scale and is badly affected by the initial value of fitting parameters due to large amount of experimental data. The iteration is probably not convergent. A comprehensive fitting parameters’ solving method based on the combination of the general polynomial and orthogonal function system is presented. This method makes the scale of Jacobian matrix downsized and confined by the number of fitting parameters rather than that of experimental data. Therefore, the calculation efficiency is enhanced and the convergence of iteration is guaranteed. And the whole computational process is more stable. Simulation verifies that this method can simplify the overall calculation and has higher fitting accuracy. In case of large amount of experimental data, the superiority of this method can be better projected.