Font Size: a A A

The Cutting Force Modeling And Control Of SiC Monocrystal Wafer Processing

Posted on:2015-08-07Degree:MasterType:Thesis
Country:ChinaCandidate:S M DuFull Text:PDF
GTID:2381330596979768Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
Having good physical and mechanical properties,SiC monocrystal is widely used in high-power devices and IC industry.However,due to high hardness and brittleness of the material,it becomes difficult to process.Diamond wire saw cutting technology with its narrow kerf,environment-friendly and high quality wafer are widely used in the process of cutting hard and brittle materials.Cutting as the first procedure has significant influence on subsequent processing.Because of the dynamic changes of the cutting force,wafer and wire saw break happened frequently,and the wafer surface quality are damaged heavily.The SiC monocrystal cutting force are analyzed to establish a model for the dynamic control of the cutting force of the process,and two controllers are designed to keep the cutting force stay at a constant value by changing the part feed rate throughout the cutting process and compared by experiments.Part feed rate is chosen as the factor affect cutting force most to simplify the system after analyzed the parameters(wire saw velocity,part feed rate,wire tension and part rotation)influence on cutting force.The normal force is chosen as the final output after analyzed the generate of cutting force in reciprocating diamond wire saw cutting system and the characteristic of tangential force and normal force which are two division of cutting force are analyzed.A single input single output system is build consists of the part feed rate and normal force.According to the F-test theory,analyze the system output to the system input of inverse M sequence,calculate the specify statistic to determine the system order.Forgetting factor recursive least square with forgetting factor is 0.98 is selected to identify the system parameters after compare the performance of general recursive least squares and forgetting factor recursive least square.The differential equation between normal force and part feed rate is developed.Two different controllers are designed using minimum variance control and generalized predictive control,respectively.MATLAB simulation show that two controllers both can meet the requirement of control the cutting force.However the predictive cutting force controller hasbetter performance because it can't relay on system parameters to calculate the control variable and it also can avoid the error caused by mismatch of estimated parameters and system actual parameters.An automatic part feeding system with diamond wire saw cutting is built to test the performance of the two controllers by tracing reference force,square wave signal,triangle wave signal and sine wave signal.The experiments results show that two different force controllers can both track the reference output well,with good robustness.Compared with constant part feed rate,normal force processing can achieve higher surface quality and productivity and provide a basis for multi-variable control.
Keywords/Search Tags:SiC monocrystal wafer, System identification, F-test, Adaptive Control, Predictive control
PDF Full Text Request
Related items