Study On Phase Analysis Of Iron And Copper In Geological Samples

There are two properties of all elements in nature, which are quantity andquality. The property of quantity of an element shows that how many amounts of theelement there is, generally speaking, it is called the abundance of the element. Theanother one depicts that what kind existent state the element could be. Ingeochemical samples, the different existent state of the element indicates thedifference of the behavior for the element in the course of the geochemical action.This difference provides scientific basis for study the pattern of the elementdistribution and evolution in the geochemical system. With the development of thescience and technology, only to know the total content of unknown elements cannotmeet the practical demands distantly. Furthermore, we must know the different modeof occurrence and distribution of the elements. So it is very important to study onphase analysis of elements in geological exploration, mineral deposit evaluation,beneficiation process, minerals comprehensive utilization and environmentalconservation.At present, the main method used in chemical phase analysis is to determine theelement content of different phases after a selective dissolution for the geologicalsamples. There is a large error because of the inevitable crossing phase in the courseof dissolution. The relative error is allowed to be 50% probably in phase analysis.On top of this, the procedure of chemical extraction is cockamamie, time taking andexigent technically. GA is independent of the calibration model and prior knowledgeabout the system, and it is a high efficient global optimization method. It has beenused to calculate equilibrium constant of complex in analytical chemistry, to selectoptimal reaction condition and to determine multi-component simultaneously.There is no report about that GA was applied to phase analysis till to now. In order to improve these defects, Multi-objective Restriction GeneticAlgorithms (MGA) is introduced to determine five phases of iron in geologicalsamples simultaneously (1,2,, and ) and four phases ofcopper (,,1 and 2). In this article, a new method of phaseanalysis of geological samples will be developed, which is simple, quick, accurate,but also practical. Below are the achievements of this method. 1. Optimal extracting conditions for five phases of iron with six infusions andfour phases of copper with five infusions have been found. After a series ofexperiments, for example, concentration of solvent, extracting time, temperature,proportion of solid to liquid, different weight of sample, artificial mixed single phasesample and experiment repeatedly, the extracting ratio constants for all phases indifferent infusions are gained, which is steady and reliable.In above mathematical model, Pij and Ti are known, but Xj is phase concentrationthat we want to know. Furthermore, this model is fit for GA for optimization, bywhich we can work out Xj. Thus, the solution of question can be gotten. 3. In GA, The binary sub-string of a certain length will replace each phasecontent respectively. These binary sub-strings will be connected to a longer binarystring named an individual. There is a certain coordinate relation between a binarysub-string and the content of a phase. By the multi-objective restriction function, wecan judge whether the individual is better or worse, which supplies the evidence for"survival of the fittest". Therefore, the phase analysis is converted into the problemof multi-target restriction optimization by MGA. The multi-objective restrictionfunction is founded as follows:Here, Tobs,i is the total dissolved content of sample in i solvent by experiment. Tcal,i isthe calculating sum of all dissolved phases by MGA. When the differences betweenTobs,i and Tcal,i are all small enough, the individual is the best solution of the question. 4. The initial hunting range can be narrowed gradually, which makes thecalculating result more accurate, and which is favor for hunting efficiency. Theobjective function was improved. The multi-objective is composed of euclideandistance and F-test function, and it is effective for optimization. The effects of GAparameters on calculation speed and result are discussed, such as, initial valueranges, population size, space division, mating probability, mutation probability andobjective function, Initial value ranges is unnecessary to have prior knowledge forsystem, and initial ranges are easy to be decided. The optimal population size is 16,with the list going on, space division 212, mating probability 0.8, and mutationprobability 0.02. 5. The accurate solution can be gotten when gaussian elimination is used tosolve the linear equation group of phase analysis. The accurate solution is comparedwith the calculating result by MGA. It is found that the later is better than the formergreatly, due to the fact that accurate solution by gaussian elimination is zero error.But zero error is ideal situation and it is difficult to reach in practice. So all we can...