Application of Grey Neural Network to Energy Saving Verification for Air-Conditioning Systems
碩士 === 國立臺北科技大學 === 能源與冷凍空調工程系碩士班 === 104 === This study investigated the central air-conditioning system of a hospital in New Taipei City. It collected the actual data from the central monitoring system established by the air-conditioning manufacturer and adopted regression analysis to eliminate un...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Online Access: | http://ndltd.ncl.edu.tw/handle/uykp26 |
id |
ndltd-TW-104TIT05703060 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-104TIT057030602019-05-15T22:54:23Z http://ndltd.ncl.edu.tw/handle/uykp26 Application of Grey Neural Network to Energy Saving Verification for Air-Conditioning Systems 應用灰色神經網路於空調系統之節能驗證 Dai-Lun Wu 吳岱倫 碩士 國立臺北科技大學 能源與冷凍空調工程系碩士班 104 This study investigated the central air-conditioning system of a hospital in New Taipei City. It collected the actual data from the central monitoring system established by the air-conditioning manufacturer and adopted regression analysis to eliminate unreasonable data, then used grey neural network to construct the coefficient of performance model, using the patented "Energy Saving Verification of Air-Conditioning System" from the Intellectual Property Office, MOEA, to analyze the energy consumption situation of chiller unit under different operation modes. According to the results of this study the average C.O.P of the constant speed, spiral model is 3.99, total power consumption is about 12625kW; average C.O.P for variable speed, active magnetic bearing compressor reaches 6.30, total power consumption is about 7991kW, the calculated energy saving on power consumption differs by about 4634 kW, the energy efficiency for the latter is as high as 37%. In order to verify grey neural network’s ability in extrapolation, this study conducted modeling via the three methods of grey system, back propagation neural network and grey neural network respectively and estimated the data outside the known data range of the two chillers. The result of this study shows that its mean absolute percentage error (MAPE) is approximately 3 to 4%, showing that grey neural network has good predictive ability. Furthermore, using grey neural network to carry out small amount data modeling and prediction, apart from its characteristics of grey system small amount data modeling, it also improves the shortcoming of long duration in neural network modeling, thus proves grey neural network’s accuracy and characteristics in handling extrapolation. 張永宗 學位論文 ; thesis 0 |
collection |
NDLTD |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立臺北科技大學 === 能源與冷凍空調工程系碩士班 === 104 === This study investigated the central air-conditioning system of a hospital in New Taipei City. It collected the actual data from the central monitoring system established by the air-conditioning manufacturer and adopted regression analysis to eliminate unreasonable data, then used grey neural network to construct the coefficient of performance model, using the patented "Energy Saving Verification of Air-Conditioning System" from the Intellectual Property Office, MOEA, to analyze the energy consumption situation of chiller unit under different operation modes. According to the results of this study the average C.O.P of the constant speed, spiral model is 3.99, total power consumption is about 12625kW; average C.O.P for variable speed, active magnetic bearing compressor reaches 6.30, total power consumption is about 7991kW, the calculated energy saving on power consumption differs by about 4634 kW, the energy efficiency for the latter is as high as 37%.
In order to verify grey neural network’s ability in extrapolation, this study conducted modeling via the three methods of grey system, back propagation neural network and grey neural network respectively and estimated the data outside the known data range of the two chillers. The result of this study shows that its mean absolute percentage error (MAPE) is approximately 3 to 4%, showing that grey neural network has good predictive ability. Furthermore, using grey neural network to carry out small amount data modeling and prediction, apart from its characteristics of grey system small amount data modeling, it also improves the shortcoming of long duration in neural network modeling, thus proves grey neural network’s accuracy and characteristics in handling extrapolation.
|
author2 |
張永宗 |
author_facet |
張永宗 Dai-Lun Wu 吳岱倫 |
author |
Dai-Lun Wu 吳岱倫 |
spellingShingle |
Dai-Lun Wu 吳岱倫 Application of Grey Neural Network to Energy Saving Verification for Air-Conditioning Systems |
author_sort |
Dai-Lun Wu |
title |
Application of Grey Neural Network to Energy Saving Verification for Air-Conditioning Systems |
title_short |
Application of Grey Neural Network to Energy Saving Verification for Air-Conditioning Systems |
title_full |
Application of Grey Neural Network to Energy Saving Verification for Air-Conditioning Systems |
title_fullStr |
Application of Grey Neural Network to Energy Saving Verification for Air-Conditioning Systems |
title_full_unstemmed |
Application of Grey Neural Network to Energy Saving Verification for Air-Conditioning Systems |
title_sort |
application of grey neural network to energy saving verification for air-conditioning systems |
url |
http://ndltd.ncl.edu.tw/handle/uykp26 |
work_keys_str_mv |
AT dailunwu applicationofgreyneuralnetworktoenergysavingverificationforairconditioningsystems AT wúdàilún applicationofgreyneuralnetworktoenergysavingverificationforairconditioningsystems AT dailunwu yīngyònghuīsèshénjīngwǎnglùyúkōngdiàoxìtǒngzhījiénéngyànzhèng AT wúdàilún yīngyònghuīsèshénjīngwǎnglùyúkōngdiàoxìtǒngzhījiénéngyànzhèng |
_version_ |
1719137966417248256 |