A methodology to select the experimental plant instrumentation based on an a priori analysis of measurement errors and instrumentation cost