Font Size: a A A

Study On Steel Leakage Prediction In Continuous Casting And Research In Crystallizer Vibration Monitoring Software

Posted on:2018-08-24Degree:MasterType:Thesis
Country:ChinaCandidate:H J ZhengFull Text:PDF
GTID:2321330533463714Subject:Engineering
Abstract/Summary:PDF Full Text Request
With the development of continuous casting technology,high efficient continuous casting technology has been a main research direction in continuous casting field.The core of high efficient continuous casting is the high casting speed,but the increase of casting speed will raise the risk of steel leakage.The steel leakage of continuous casting will bring heavy losses to the enterprise once it occurs.Therefore,it is significance in theory and practice to developing effective prediction of steel leakage in continuous casting.The main work of this paper is as follows:First,the causes,formation process and testing methods of steel leakage,especially adhesive steel leakage,are analyzed.The prediction theory of continuous casting prediction based on thermocouple temperature measurement is selected as the theoretical basis of this paper.Secondly,a support vector machine with advantages of small sample training data is used as a steel leakage prediction model for continuous casting.The particle swarm optimization(PSO)algorithm is introduced into the training process of SVM,increasing the optimization speed of steel leakage prediction model by using the advantages of its less parameters and fast optimization speed.Besides,In order to solve the problem that particle swarm optimization is easy to fall into local extremum,the simulated annealing algorithm is introduced to evaluate the new position of the updated particles to judge whether the new position is accepted in the process of updating the particle swarm iterations The feasibility and effectiveness of this algorithm are proved by the test on history data of continuous casting of a steel plant.Thirdly,based on the existing PC monitoring software of crystallizer vibration system driven by electrohydraulic servo in the laborary,the upper computer monitoring interface of steel leakage prediction is designed,including real-time thermometry thermocouple module,algorithm module and alarm module,.Finally,on the shaking table of continuous casting mould driven by electro-hydraulic servo in laboratory,the simulation experiment research of testing real-time temperature monitoring function and whether can timely and accurately judge the leakage of steel and sending the the alarm information are carried out.
Keywords/Search Tags:continuous casting mold, support Vector Machine(SVM), steel leakage prediction, particle swarm algorithm, simulated annealing algorithm
PDF Full Text Request
Related items