冰水主機之性能提升及節能分析

Analysis of energy saving and improvement for chiller performance

江彩綺(Chai-Chee Kong)

指導教授 ： 張永宗

國立臺北科技大學/機電學院/能源與冷凍空調工程系碩士班/碩士(2013年)

摘要

本研究主要使用了迴歸分析、類神經網路及基因類神經網路三種方法，分別建立冰水主機在冷凝器清洗前的性能係數(C.O.P)模型。接著在冷凝器清洗後收集數據，使用該模型模擬冷凝器清洗前的性能係數(C.O.P)，並分析及比較三種方法在該基準下所模擬的結果及提升的效率。類神經網路使用倒傳遞類神經網路，而基因類神經網路是根據倒傳遞類神經網路所模擬的結果，設計適當的適應函數，來搜尋其最佳的加權值及偏壓值。
本研究以兩種不同的案例做模擬比較，結果顯示案例一的冰水主機性能係數(C.O.P)平均提升了3.82%,而案例二的冰水主機性能係數(C.O.P) 平均提升了3.78%。一般而言，使用類神經網路來進行模擬其準確度已非常高，基因類神經網路只是根據所設計的條件來尋找最佳的加權值及偏壓值，達到最佳化的模擬效果。

關鍵字

模擬 ； 性能係數 ； 迴歸 ； 倒傳遞神經網路 ； 基因類神經網路

並列摘要

There are three methods applied in this study: linear regression, neural network and genetic neural network. The C.O.P models, before cleaning the chiller’s condenser, are established by using these three methods. We collect data of the chiller’s condenser after cleaning, then simulate the C.O.P of the chiller’s condenser before cleaning by using these three models. After that, analysis and compare the simulate results and improvement of performance by using three methods under the same baseline. Backpropagation network is used for neural network and the genetic neural network is used according the results of neural network, to devise a suitable fitness function and to search the optimal weights and biases of the backpropagation network.
Two different cases are presented for simulate simulation and to compare. The results show that the C.O.P of case 1 can be improved up to 3.82%, and the C.O.P of case 2 can be improved up to 3.78%. Generally, using neural network to do simulation will get high accuracy. Genetic neural network is used based under the devise conditions to search the optimal weights and biases of the neural network to reach the goal of optimizing simulation.

並列關鍵字

simulation ； coefficient of performance(C.O.P) ； linear regression ； backpropagation network ； genetic neural network

參考文獻

[15] A.Azadeh, S.F. Ghaderi, S. Tarverdian, M.Saberi, “Integration of artificial network and genetic algorithm to predict electrical energy consumption”, Applied Mathematics and Computation, vol. 186, 2007, pp. 1731-1741.

[18] Leonard D. Tijing, Hang Young Kim, Dong Hwan Lee, Cheol Sang Kim, Young. l.Cho, “Physical water treatment using RF electric fields for the mitigation of CaCo3 fouling in cooling water”, International Journal of Heat and Mass Transfer,vol.53,2010,pp. 1426-1437.

[19] Liu Ke, Guo Wenyan, Shen Xiaoliu, Tan Zhongfu,a, “Research of the forcast model of electricity power industry loan based on GA-BP Neural Network”, Energy Procedia, vol.14,2012,pp.1918-1924.

[20] 段春雷，變頻氣冷式冰水主機節能耗電分析，碩士論文，國立臺北科技大學能源與冷凍空調工程系，2012。

[25] 李信蔚，應用模糊進化規劃法於空調負載預測，碩士論文，國立台北科技大學能源與冷凍空調工程系，2010。