Instrumentation and estimation for high temperature control

Within a variety of industrially relevant high temperature production processes such as welding, heat treatment and metal deposition, the quality of the manufactured component is largely affected by how well parameters can be controlled during processing. These parameters might be, in the case of metal deposition, power input, material feed, or a parameter which is common for all of the aforementioned processes: material temperature. The ability to correctly measure, or in other ways estimate process parameters is vital in order to successfully control high temperature processes such as above 700 degrees Celsius. In this work, instrumentation and estimation solutions adapted to high temperature control are proposed and implemented with a focus on the laser metal wire deposition process. Special attention is given to temperature measurements on specimens with varying emissivity as commonly found in high temperature processes. A calibration procedure for a single-wavelength pyrometer is also presented together with a general discussion on limitations of such a system for measurands with varying emissivity. A new method for non-contact emissivity compensated temperature estimations using a spectrometer is presented. Simulations and industrially relevant experiments have been carried out validating the method. The theoretical framework for the developed method will be further investigated in the future together with additional experimental validation.In addition to temperature measurements, a method for real-time process control of laser metal wire deposition has been developed and implemented with good results. This control scheme estimates and controls the tool-to-workpiece distance based on resistance measurements. Such measurements allow for placement of instruments outside of the processing chamber and easy integration into existing equipment. Future work will be directed towards incorporation of resistance measurements into an iterative learning control scheme. Also, improvement on the resistance-distance model and further investigation into suitable signal processing methods for the resistance signal will be pursued