Sensor validation and reconstruction for building central chilling systems based on principal component analysis

Abstract

This paper presents an automatic and online FDD (fault detection and diagnose) and sensor reconstruction scheme that can be used in a building energy management and control system to detect and diagnose sensor faults and reconstruct faulty sensors in building central chilling systems. The scheme is based on principal component analysis to build a model that captures the correlation among the flow meters and temperature sensors installed in the chilling systems. The model is employed to reconstruct an assumed faulty sensor. The square prediction error based on the model and the sensor validity index based on the construction are employed, respectively, to detect the sensor fault and identify the faulty sensor. An existing building central chilling system is simulated to test and validate the developed scheme online. Four types of sensor faults are introduced, respectively, during the simulation to validate the sensor FDD and sensor reconstruction scheme.