Jiri Vass

Fault Diagnosis of Air Handling Units

Abstract This paper presents an improved method for fault detection and diagnostics (FDD) of air handling units (AHUs). The fault detection module defines observable states of the AHU, where each state depends on current values of sensor data and control signals. The fault diagnostic module maps the observable states to the faults and then applies the cumulative sum chart (CUSUM) to define the size and development of each fault in time. The FDD method was tested on real datasets and its results were confirmed by the building technician. Finally, the method is compared with the standard APAR (AHU performance assessment rules) method developed by Schein et al.

Automatic setup of fault detection algorithms in building and home automation

The complexity and diversity of building automation systems (BAS) and the various faults that may happen in buildings make it very challenging to set up fault detection and diagnostics (FDD). This knowledge-intensive, expensive tasks urgently needs to be automated due to the limited time and budget available in this domain. A novel approach for computer-aided, automatic setup of FDD for BAS is described in this paper. Using the Web Ontology Language (OWL), the characteristics, requirements and necessary configuration of the FDD algorithms are formally specified in a machine-interpretable way. Also buildings and their BAS are described by OWL as building information model (BIM). Given these descriptions, the approach can match the specific characteristics and requirements of the FDD algorithms with the specific circumstances in the BAS in a formal evaluation process. All applicable FDD algorithms can be determined and subsequently configured to the individual characteristics of the BAS.

Vibration-based approach to lifetime prediction of electric motors for reuse

This paper is concerned with lifetime prediction of components in washing machines. Vibration signals were measured on electric motors during an accelerated lifetime test ranging from 26.7 to 38.5 simulated years. Loose bearings have initiated air-gap eccentricity and rotor-to-stator rubbing, which resulted in a motor breakdown. Significant frequency bands were identified using a spectral comparison based on the constant percentage bandwidth (CPB) spectrum. Increasing trends were extracted from several vibration indicators, such as envelope cepstrum (EC) and a weighted integral of CPB differences. The EC is computed as the real cepstrum of the envelope signal obtained by demodulating the band identified by the CPB comparison. Hence the EC is more sensitive as it employs a priori information provided by historical data. The fault was first detected 9.7 years in advance and confirmed 5.3 years before the breakdown. The indicators can be integrated with a recent methodology based on Weibull analysis and neural network modelling.