Font Size: a A A

Temperature Soft Sensor Modeling For Forgings Based On EM&Gaussian Mixture Model

Posted on:2014-12-21Degree:MasterType:Thesis
Country:ChinaCandidate:Y SuFull Text:PDF
GTID:2251330425973082Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
High intension aluminum alloy is stress mechanism’s main m-aterial of spaceflight aerocraft’s coating,beam grider,wing.And the temp-erature setpoint of the quenching fuenace is the Activity Criticality Index. During the actual production process, it is too hard to mearure the forging’s tempareture directly for it’s hanging in the central of the studio. Therefore,it’s necessary to find a way to get the forging’s tempareture indirectly.In this paper,we establish the relationship between the forgings and the studio,then try to get the forging’s tempareture by measuring the studio’s tempareture.But it’s difficult to establish the quench furnace’s exact mathematical model for its thermal inertia is great,its delayed time is long,the coupling influence is severe.This paper discusses the component temperature soft measurement modeling method.According to quenching furnace’s characteristics and working condition of diversity, this paper first established the component temperature soft measurement model based on ARX model. Due to the local optimality of ARX model, while applying the model to the global temperature prediction,the prediction result is not ideal.And during practical applications, temperature setpoint of the quenching furnace is relevant to the forgings’ size, number and type.So different forgings need to set different temperature setpoint. And it is impossible to collect comprehensive temperature data of all heating process.In order to solve this problem, this paper presents temperature soft sensor modeling for forgings based on EM&Gaussian Mixture Mode. The model adopts the Gaussian mixture autoregressive model to approximate the model’s bayesian network.And use EM algorithm to identify the parameters in the Gaussian mixture autoregressive model.Simulation results show that the proposed Gaussian Mixture model based on EM algorithm has realized the accurate temperature measure-ment for the global structure, fully meet the practical application of the industrial field, and improve the performance of the aluminum products and quality.
Keywords/Search Tags:quenching furnace, temperature soft sensor, autoregressiveEM algorithm, Gaussian Mixture
PDF Full Text Request
Related items