Determination Of Archie’s Parameters By Micro-genetic Algorithm For Analysis Of Water Saturation From Well-logging Data In The Area Of Zhang Wu

In general there are some quantitative parameters in logging interpretation of equations, such as the Archie formula for calculating the formation water saturation and the estimation of oil and gas reservoir. However the parameters of a、m、n in the formula is difficult to accurately determine at the same time in practice, and usually these parameters will be determined by a single or two together, then when these obtained parameters are going to get the real into the calculation in actual production, the results are often unsatisfactory, and whether these parameters can be acquired correctly or not has close relationship with the accuracy of the final interpretation. Genetic algorithm, a kind of new optimization method, has the characteristic about global optimization. It is suitable to solve nonlinear optimization problems, and in recent years genetic algorithm has been successfully applied in the field of complex function optimization for the solution of the structure optimization design, adaptive control, system control, pattern recognition and so on. 1. Collect nine Temple group well logging data in Zhang-Wu area, including water saturation, porosity, resistivity, and core measured water saturation data involved in Archie’s formula from the same stratum in different locations of logs; 2. To determine the approximate range of Archie parameters combinations; 3. According to the value of the range of parameters, to establish the coding of formula of the parameter a、m、n from decimal to binary code; 4. To determine the fitness function of the algorithm that is, establish the function for the minimum value of the absolute of D-value for the water saturation and water saturation core measured; 5. Randomly generated 50 groups of Archie parameters combinations made up binary; 6. The determine of the genetic operators about the crossover probability, mutation probability values; 7. The use of genetic operator for genetic manipulation of Archie parameters, and the termination of the process for iteration process of algorithm by the prior set termination of generation to select the optimal combination of the parameters of Archie; 8. Decoding selected parameters combinations to decimal system which we know. And at the end we can finish the optimization process of the algorithm. However general genetic algorithms often have to be trapped in local optimal solution, this paper presents the algorithm improvement steps which are called Micro-genetic algorithm, namely in step 6, when after the completion of a genetic operator, to recalculate individuals in a population of fitness, and then select the individuals ranking the first 70% in population fitness value, using 30% of the randomly generated individuals to replace those individuals remaining in the population. The purpose of doing such step is to ensure the efficiency of genetic algorithm at the same time, promptly introduce a new individual, and avoid algorithm to converge to a local optimum due to some of the advantages of local optimal solution occupying the population. The modified genetic algorithm has greatly improved the accuracy of the algorithm.The use of the algorithm can obtain Archie formula a、m、n values at the same time, according to other known parameters in the formula to calculate water saturation belonging to the strata in the area of water saturation. By comparing the water saturation calculated by these parameters and water saturation by parameters obtained from conventional chart method and the result is that Micro genetic algorithm is more accurate, on the other hand, compared with the traditional genetic algorithm and found that total genetic generations restrained to the optimal solution of improved algorithm was significantly smaller than total genetic generations in the traditional genetic algorithm, and it has a lower relative error than traditional genetic algorithm. The improved genetic algorithm is good for solving the a、m、n values of the Archie formula, and the accuracy is good and the feasibility is high. The modified genetic algorithm is suitable for prediction of water saturation in the region.