Intelligence Method Research In Deep Shaft Freezing Design

In the artificial freezing shaft sinking project,the freeze design is a core technology,which relates to the entire freezing project success or failure.At present,the theory about freezing wall and the freezing programme above 450m has been grasped.But ultra surpasses 450m in deep alluvium,the multi-coil pipes freezing programme has been carried on,the existing freezing wall computation theory has not been suitable and the multi-coil pipes freezing programme also has many problems.Therefore,some new topics have been proposed in the theory, technology and craft. From the theory and the macroscopic analysis,the freezing design is a multi-inputs and multi-objectives nonlinear system model.Based on the solution of the multi-dimensional and non-linear complex model, the artificial intelligence method has the extremely prominent merit.Therefore,the research about intelligence method for freezing design has been carried out in the paper.Base on freezing wall design practice at present in deep alluvium,the paper has proposed the multi-inputs and multi-bojecttives mathematical function model about the freezing wall , the multi-coil pipes freezing plan and the freezing plan optimal design;Studied the intelligent network algorithm and its realization flow,which the artificial neural networks(ANN) and the genetic algorithms(GA) unifies;Constructed corresponding intelligence networks and their study samples according to the above mathematical model;Through the network training,obtained the solutions of the freezing walk the multi-coil pipes freezing programme and its optimized computation model;As the sub-functions with these model solutions,compiled a set of deep shaft freezing design intellectual software using the MATLAB 7.0 platforms. This software has realized the following function: Can carry on the deep shaft freezing wall design, computate freezing wall thickness and temperature;Can carry on the multi-coil pipes freezing programme,realize intellectualization and visibitity;According to the optimal goal, can computate key parameters of freezing wall such as freezing wall thickness, temperature, shaft-side temperature and so on, from this,carry on intelligent design of the multi-coil pipes freeze proposal optimal design, obtain the optimal freezing programme.The paper has also developed the influence (sensitivity) acted from inputs to outputs based on the ANN and some influencing factors analysis about multi-coil pipes freezingwall temperature field,obtained the influence reacted from freezing distance to point temperature and each influence factors sensitivity of multi- coil pipes freeze temperature field.These conclution has provided the reference for the freezing design and the multi-coil pipes freezing wall temperature field research.The achievements in this paper have provided theory and practice reference for the research of freezing design,dressed a certain lessons for further and thoroughly artificial intelligence application with freezing shaft sinking in deep alluvium.