Pressure prediction under the condition of vacuum gauge failure in ultra high vacuum systems using regression model

Abstract

Particle accelerators are complex and large facilities requiring appropriate maintenance and operation condition control. Despite accelerator facility staff following all necessary operation regulations and the high-quality equipment usage, it is impossible to avoid some minor faults causing the pressure to arise inside the facility vacuum vessel. However, it is possible that during particle accelerator operation there is a gauge fault but vacuum pipeline pressure does not exceed the estimated critical value. In this case, operation should not be stopped, and the gauge can be easily replaced at the next planned maintenance session. In this article, the combination of Monte Carlo pressure simulation and a linear regression model to predict the pressure profile along the pipeline axis is shown. As a result, a trained AI-model based on a linear regression algorithm is able to predict the maximal pressure in the vacuum pipeline and define the pressure on a broken gauge according to other gauges’ measurements.