?

01拉普拉斯變換定義

---

1.變換公式

2. 常見函數(shù)的Laplace變換

?

02 Laplace數(shù)值逆變換

---

根據(jù)拉普拉斯逆變換的公式，可以看到，f(t)可以變成如下的公式。

#!/usr/local/bin/python # -*- coding: gbk -*- #============================================================ # TEST1.PY -- by Dr. ZhuoQing 2020-07-10 # # Note: #============================================================from headm import *################################################################# # Function InvLap(t,omega,sigma,nint), numerically inverts a # # Laplace transform F(s) into f(t) using the Fast Fourier # # Transform (FFT) algorithm for a specific time "t", an # # upper frequency limit "omega", a real parameter "sigma" # # and the number of integration intervals "nint" . # # # # Function F(s) is defined in separate as Fs(s) (see code # # below). Fs(s) has to be changed accordingly everytime the # # user wants to invert a different function. # # # # I suggest to use omega>100 and nint=50*omega. The higher # # the values for omega, the more accurate the results will be # # in general, but at the expense of longer processing times. # # # # Sigma is a real number which must be a little bigger than # # the real part of rightmost pole of the function F(s). For # # example, F(s) = 1/s + 1/(s-2) + 1/(s+1) has poles for s=0, # # s=2 and s=-1. Hence, sigma must be made equal to, say, # # 2.05 so as to keep all poles at the left of this value. # # The analytical inverse for this simple function is # # f(t) = 1 + exp(-t) + exp(2t). For t=1.25, omega=200, # # nint=10000 and sigma=2.05, the numerical inversion yields # # f(1.25) ~= 13.456844516, or -0.09% away from the actual # # analytical result, 13.468998757 (results truncated to 9 # # decimal places). This is the example used in this code. # # # # Creator: Fausto Arinos de Almeida Barbuto (Calgary, Canada) # # Date: May 18, 2002 # # E-mail: fausto_barbuto@yahoo.ca # # # # Reference: # # Huddleston, T. and Byrne, P: "Numerical Inversion of # # Laplace Transforms", University of South Alabama, April # # 1999 (found at http://www.eng.usouthal.edu/huddleston/ # # SoftwareSupport/Download/Inversion99.doc) # # # # Usage: invoke InvLap(t,omega,sigma,nint), for t>0. # # # ################################################################# # We need cmath because F(s) is a function operating on the # complex argument s = a + bj from math import ceil from cmath import *# *** Driver InvLap function *** def InvLap(t,omega,sigma,nint): # Sanity check on some parameters.omega = ceil(omega)nint = ceil(nint)if omega <= 0:omega = 200if nint <= 0:nint = 10000return (trapezoid(t,omega,sigma,nint))# *** Function trapezoid computes the numerical inversion. *** def trapezoid(t,omega,sigma,nint):sum = 0.0delta = float(omega)/nintwi = 0.0# The for-loop below computes the FFT Inversion Algorithm. # It is in fact the trapezoidal rule for numerical integration.for i in range(1,(nint+1)):witi = complex(0,wi*t)wf = wi + deltawfti = complex(0,wf*t)fi = (exp(witi)*Fs(complex(sigma,wi))).realff = (exp(wfti)*Fs(complex(sigma,wf))).realsum = sum + 0.5*(wf-wi)*(fi+ff)wi = wfreturn ((sum*exp(sigma*t)/pi).real)# *** The Laplace function F(s) is defined here. *** def Fs(s): # return (1.0/s + 1.0/(s+1.0) + 1.0/(s-2.0))return 1.0/s# Function InvLap(t,omega,sigma,nint) is invoked. printf(InvLap(1.25, 200, 3.05, 10000))#------------------------------------------------------------ # END OF FILE : TEST1.PY #============================================================

?

03一些常見函數(shù)數(shù)值逆變換結(jié)果

---

• f(t)=sin(t)
  

def Fs(s):return 1/(s*s+1)

^{▲ sin(t)數(shù)值逆變換結(jié)果}

• f(t)=t
  
  

  ^{▲ 函數(shù)t的Laplace數(shù)值逆變換結(jié)果}

• f(t)=exp(-t)

^{▲ exp(-t)函數(shù)數(shù)值逆變換結(jié)果}

?

※ 結(jié)論

---

本文測試了 一個簡易版本的數(shù)值Laplace逆變換的程序。

這個算法需要人工給出積分的上下限，積分限的實部等。如果參數(shù)給定不合適，計算出來的數(shù)值誤差較大。

總結(jié)

以上是生活随笔為你收集整理的逆Laplace数值逆变换的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯，歡迎將生活随笔推薦給好友。