Uncertainty analysis for pump test based on maximum entropy and Monte Carlo method

Entropy analysis is a key factor for ensuring the accuracy of measured data in pump test. But a unified and widely accepted evaluation method has not been formed. In this paper, an improved method based on the combination of maximum entropy and Monte Carlo method is proposed. Based on maximum entropy concept, statistical model of random distribution variables was presented, and a parameter estimation method of the distribution function was proposed also. In order to improve the convergence of the numerical estimation method, swarm algorithm is applied to generate initial value. Then, a set of random numbers are generated which are used to estimate the uncertainty of components in pump test. Case study verified the effectiveness and higher accuracy of proposed method.