Modeliranje in zaznavanje napak v klimatskih sistemih

STRŽINAR, ŽIGA (author), Dovžan, Dejan (mentor), Škrjanc, Igor (co-mentor)

V delu je predstavljeno modeliranje ter zaznavanje napak v klimatskem sistemu. Predstavljenih je več modelov pomembnejših elementov klimatskih sistemov: grelca, hladilnika, rekuperatorja, vlažilca, sobe, ventilatorjev. Modeli so uporabljeni za gradnjo skupnega modela klimatskega sistema, ki posnema realen sistem. Najdeni so parametri modela, doseženo je zadovoljivo ujemanje med zgrajenim modelom in realnim sistemom. Generirani so učni simulacijski teki, ki ponazarjajo delovanje sistema ob pojavih štirinajstih različnih napak. Učni teki so uporabljeni za učenje detektorjev napak. Detektorji napak delujejo na osnovi dinamične metode glavnih komponent. Prikazano je delovanje detektorjev na tesnih simulacijskih setih. Predstavljen način zaznavanja napak za večino predvidenih napak deluje z zadovoljivo uspešnostjo.

• Language: Slovenian
• Keywords: modeliranje, klimatski sistem, Simscape, zaznavanje napak, prepoznavanje napak, metoda glavnih komponent, zaznavanje reda
• Work type: Master's thesis/paper
• Organization: FE - Faculty of Electrical Engineering
• Year: 2017
• PID: 20.500.12556/RUL-99028
• Publication date in RUL: 21.12.2017

Secondary language

• Language: English
• Title: Modeling and fault detection of heating, ventilation, and air conditioning system

Abstract

This work presents modelling and fault detection in heating, ventilation and air conditioning (HVAC) systems. Several models of the most important elements are presented. The modelled elements include: a heating element (heat exchanger), a cooling element, a recouperator, a humidifier, a room and fans. The models are used to construct an unified HVAC system model able to match the behaviour of a real system. Model parameters are determined and a sufficient fit between the model and the real device is achieved. Several train simulation runs are generated representing system behaviour with fourteen different faults. These simulations are used to train fault detectors. Fault detection is based on dynamic principle component analysis (DPCA). Detectors are tested on test simulation runs and the results are presented. The achieved fault detection rates are satisfactory.

• Keywords: modelling, HVAC, Simscape, fault detection, fault classification, PCA, order detection