Research On Multi-objective Optimization Of Drilling Parameters Based On Modified Particle Swarm Optimization

Drilling parameters are the important factors that affect the penetration rate, the cost and quality of drilling in the process of drilling. Drilling parameters optimization can improve the mechanical drilling rate, reduce the bit wear and the drilling costs. Original drilling parameters optimization theory regards drilling rate as the core,to reduce the cost as the single purpose, ignoring other factors, but now, not only to improve the penetration rate of drilling engineering, more comprehensive consideration of factors such as the bit life and the bit energy ratio, should be taken into consideration for multi-objective optimization. In addition, in recent years the application of intelligent optimization algorithms to effectively solve the real-time optimization has become one of the key technology of optimizing drilling parameters, introducting the new theory of optimization algorithm to solve the model of drilling becomes a inevitable trend.Aiming at the limitation of traditional single objective optimization of drilling parameters, multi-objective optimization model of drilling parameters, which included three goals: the mechanical drilling speed, the bit life and the bit energy ratio was put forward satified certain constraints. Due to the present optimization method is easy to trap into a local optimization and has a low efficiency, a modified chaotic multi-objective particle swarm optimization algorithm was proposed. Standard test function results showes that the improved algorithm has good convergence and distribution, and a better ability to jump out of local optimum.At last, the modified algorithm was introduced to solve multi-objective optimization model of drilling parameters according to an oil field practical example. Right value of the algorithm’s parameters was discussed, and coefficients of the drilling rate mode were determined using Regression analysis. The simulation test obtaines the evenly distributed pareto front, which proves the effectiveness of the model and the algorithm, provided effective evidence for the selection of optimization scheme.