Development, implementation, and evaluation of a fault detection and diagnostics system based on integrated virtual sensors and fault impact models

The primary goal of this work was to develop, demonstrate, and evaluate a fault detection and diagnostics (FDD) system that incorporates integrated virtual sensors and fault impact models for decision support. FDD systems have the potential for improving energy efficiency along with reducing service costs. To achieve this goal, virtual sensors and fault impact models that only require low cost measurements are implemented to detect the cause of the fault, diagnose the severity of a fault and assess the performance reduction. The integrated virtual sensors can handle multiple simultaneous faults to isolate a specific fault from other faults. The fault impact models can decouple performance impacts of individual faults under multiple fault conditions. A complete diagnostic FDD system was implemented and demonstrated for a rooftop air conditioner. The test conditions included a wide range of fault and operating conditions with multiple simultaneous fault situations. Laboratory tests were performed to define reasonable thresholds for the FDD system with multiple simultaneous faults to assess whether RTU service should be performed. The FDD system provided the following diagnostic outputs for and impacts of: 1) loss of compressor performance, 2) low or high refrigerant charge, 3) fouled condenser, 4) evaporator filter, and 5) faulty expansion device. (C) 2020 Elsevier B.V. All rights reserved.