J. Cult. Heritage 1 (2000) S303–S309

Jong-Myoung Lee, Ken G. Watkins

Department of Engineering, University of Liverpool, Brownlow Street, Liverpool L69 3GH, UK

During laser cleaning, surface damage on the substrate could easily occur by overexposure to a laser pulse of high fluence. The damage is especially serious in art conservation where recently the laser has found a successful field of application. Successful cleaning without surface damage can be achieved by skilled expert operators with long experience and good technique. This paper presents a fuzzy rule-based expert system to predict surface damage during laser cleaning like a human expert. In this work, a fuzzy rule base was used to embed the acoustic information including an indication for the progress in cleaning and the result. An inference process was conducted to predict whether and how much surface damage would be induced on the substrate. In order to detect the acoustic waves a wide-band microphone was utilised. Tests of the performance of the fuzzy expert system showed that the prediction of surface damage is well correlated with the actual results independent of initial surface conditions. Finally, a process control algorithm for laser cleaning has been developed on the basis of the surface damage prediction system