Important sensors for chiller fault detection and diagnosis (FDD) from the perspective of feature selection and machine learning

Han, Hongbin; Gu, Bin; Wang, T.; Li, Z.R.

doi:10.1016/j.ijrefrig.2010.08.011

Cited by 109 publications

(41 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This study introduces a correction function g to compensate for the systematic error in the measurements, as shown in Eq. (5). The correction function of one sensor is formulated with offsetting constants and its measurement M. The specific equation is based on a characteristic of the systematic sensor error.…”

Section: Benchmark and Correction Function In Virtual In-situ Calibramentioning

confidence: 99%

“…To address the challenges in the building sector, a comprehensive solution package, including continuous fine-tuning of building automation systems, automated analytical optimization, and automated fault detection, diagnostics and repair, is needed. Research into these challenges has been conducted in order to help mitigate increasing energy use in buildings [3][4][5][6][7][8]. Most of the proposed approaches will be effective only if the data obtained from sensors are reliable and accurate [3].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Extended virtual in-situ calibration method in building systems using Bayesian inference

Yoon

2017

Automation in Construction

View full text Add to dashboard Cite

Measurements from sensors and knowledge of key parameters are of great importance in the operation of modern building systems. Accurate and reliable information as these serves as the base for ensuring the desired performance of control algorithms, fault detection and diagnostics rules, analytical optimization strategies. They are also crucial for developing trustworthy building models. However, unlike mass produced industrial devices, building systems are generally one of a kind and sparsely instrumented. Despite the indispensable need, dense deployment of sensors or a periodic manual calibration for ensuring the quality of thousands variables in building systems is not practical. To address the challenge, we extend our virtual in-situ calibration method by marrying it with Bayesian inference, which has a better capability in handling uncertainties. Strategies, including local, global, and combined calibration, are evaluated in a case with various sensor errors and uncertain parameters. The detailed procedure and results are presented.

show abstract

Section: Benchmark and Correction Function In Virtual In-situ Calibramentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Extended virtual in-situ calibration method in building systems using Bayesian inference

Yoon

2017

Automation in Construction

View full text Add to dashboard Cite

show abstract

“…Zhou et al (2009) developed a modelbased FDD algorithm for chillers based on a semi-empirical chiller model from Ma et al (2008). Han et al (2011) developed a machine-learning-based FDD algorithm using data from Comstock and Braun (1999b). Zhao et al (2012) developed a virtual condenser fouling sensor for chillers.…”

Section: Introductionmentioning

confidence: 99%

“…Although not included in the four methods described by Reddy et al (2003), a Kalman-filter based chiller model was used in Navarro-Esbrí et al (2006) to create a chiller fault detection algorithm. Furthermore, machine-learning based FDD algorithms such as Han et al (2011) typically use machine-learning algorithms to create a model of faulted chillers to be used in the chiller FDD. DOE-2 (DOE-2 2014) and EnergyPlus (EnergyPlus 2015) are commonly used building simulation programs that employ the linear empirical chiller model originally developed for DOE-2.1 (DOE 1981).…”

Section: Introductionmentioning

confidence: 99%

Empirical modeling of the impacts of faults on water-cooled chiller power consumption for use in building simulation programs

Cheung

Braun²

2016

Applied Thermal Engineering

View full text Add to dashboard Cite

Empirical models of four chiller faults that can be applied within existing building models to study overall impacts are developed in this paper. The faults include overcharging, excess oil, noncondensables in refrigerant and water-side condenser fouling. A single generalized model structured was developed for these faults that forces predicted fault impacts to be zero with no fault and increase with increasing fault level. The models were trained and tested using available laboratory data for a water-cooled centrifugal chiller where all four faults were artificially introduced. The fault model behavior was studied and then they were integrated in hospital models from DOE commercial reference building models (Deru et al. 2011) and simulations were performed in different climates. The simulation results showed maximum increases of building electricity consumption, electricity peak demand and water consumption of the hospitals due to faults of 4.7%, 7.8% and 1.8% respectively. The fault impacts were found to be more severe in hotter and more humid climates.

show abstract

“…Under the conventional HVAC control structure, the majority of sensor fault detection and diagnosis (SFDD) methods are based on a centralized algorithm [2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19]. To the best of our knowledge, these methods have a series of disadvantages, such as poor adaptability and instantaneity.…”

Section: Introductionmentioning

confidence: 99%

Decentralized Optimization for a Novel Control Structure of HVAC System

Wang

Xing

Jiang

et al. 2016

Mathematical Problems in Engineering

View full text Add to dashboard Cite

A decentralized control structure is introduced into the heating, ventilation, and air conditioning (HVAC) system to solve the high maintenance and labor cost problem in actual engineering. Based on this new control system, a decentralized optimization method is presented for sensor fault repair and optimal group control of HVAC equipment. Convergence property of the novel method is theoretically analyzed considering both convex and nonconvex systems with constraints. In this decentralized control system, traditional device is fitted with a control chip such that it becomes a smart device. The smart device can communicate and operate collaboratively with the other devices to accomplish some designated tasks. The effectiveness of the presented method is verified by simulations and hardware tests.

show abstract

Important sensors for chiller fault detection and diagnosis (FDD) from the perspective of feature selection and machine learning

Cited by 109 publications

References 18 publications

Extended virtual in-situ calibration method in building systems using Bayesian inference

Extended virtual in-situ calibration method in building systems using Bayesian inference

Empirical modeling of the impacts of faults on water-cooled chiller power consumption for use in building simulation programs

Decentralized Optimization for a Novel Control Structure of HVAC System

Contact Info

Product

Resources

About