本研究主要目的在利用類神經演算法應用於輻射測溫，以預測金屬表面放射率與溫度。針對六種鋼材試件AISI 420、AISI 630、AISI A2、AISI A6、AISI H10、AISI H13推測出在三種溫度700K、800K、900K下的放射率，並預測試件表面溫度。
以類神經網路原理為基礎，藉由紅外線光譜儀對實驗溫度下試件所量測的輻射強度作為修正資料，利用數值方法推導出修正後的神經結構以推測表面放射率及溫度。另外透過實驗檢測此輻射測溫法之準確性並與本實驗室之前研究多光譜輻射測溫法結果比較，了解是否此法有更佳的溫度預測準確性。
研究結果顯示在神經網路結構與參數的分析上：(1)選取隱藏層神經元數目為3時將有最佳預測結果；(2)學習速率的選用將會影響推論結果與模擬時間，研究發現所用學習速率為0.0005下時有最佳表現；(3)選取最大波長數將可以得到更佳的預測結果。
對推測溫度而言: (1)於最佳結構與參數設定下，不同類型鋼材所推測表面溫度，其誤差絕大部分於20K內；(2)所推論之放射率越接近真實放射率時，越能準確地預測出真實表面溫度；(3)不銹鋼的溫度預測誤差隨溫度增高而增加。而熱作工具鋼與冷作工具鋼，隨不同溫度變化下亦能準確預測出表面溫度；(4)與多光譜輻射測溫法比較，ANNs有較佳之預測結果。

This research is mainly aimed at the analysis of applying artificial neural networks (ANNs) on radiation thermometry to predict the surface emissivity and temperature for six different steels (AISI 420、AISI 630、AISI A2、AISI A6、AISI H10、AISI H13) at three different temperatures (700K、800K、900K).
Based on ANNs principle, the spectral intensity measured by spectrometer was used as revised data to modify the network structure in numerical simulation and the emissivity and temperature were eventually obtained. On the other hand, experiments were conducted to examine the accuracy of this radiation thermometry. ANNs is then compared to MRT to see whether this method has better temperature prediction or not.
The results of neural network structure and parameters show that: (1) the best temperature prediction can be achieved when the hidden layer is three neurons; (2) learning rate is an important factor to simulation results and converge time. The best result was obtained when learning rate is 0.0005; (3) increasing number of wavelength will get better results.
For the temperature predictions: (1) with the best structure and parameters settings, most results show the temperature errors are under 20K; (2) the closer the emissivity value to real one, the more accurate surface temperature prediction; (3) hot-work steel and cold-work steel could achieve accurate prediction with the variation of temperature. However, for stainless steel the temperature error increases due to the surface melt at high temperature; (4) compared with MRT, ANNs has better temperature prediction.

1.Chunling Yang, Dai Jing-Min, and Hu Yan, “Optimum Identi_cations of Spectral Emissivity and Temperature for Multi-Wavelength Pyrometry”, CHIN. PHYS. LETT. Vol. 20, No. 10 pp.1685-1688.,2003.

2.Chunling Yang, Yong Yu, Dongyang Zhao, and Guoliang Zhao, “Study on Modeling of Multispectral Emissivity and Optimization Algorithm”, IEEE TRANSACTIONS ON NEURAL NETWORKS, VOL. 17, NO. 1, 2006.

3.SUN Xiao-gang, DAI Jing-min,” Multispectral thermometry based on neural network”, JOURNAL OF HARBIN INSTITUTE OF TECHNOLOGY(NEW SERIES), TH811.2, 2003.

4.Herve ́ Abdi, “Neural Networks”, The University of Texas at Dallas , 2003.

5.Robert A. Jacobs, “Increased Rates of Convergence Through Learning Rate Adaption”, Neural Networks, Vol. 1 . pp. 295-307, 1988.

6.R.B. Boozarjomehry, W.Y. Svrcek, ” Automatic design of neural network structures”, Computers and Chemical Engineering 25 pp.1075–1088 ,2001.

7.Touloukian, Y. S. and DeWitt, D.P., “Thermal Radiative Properties, Metallic Elements and Alloys,” Thermophysical Properties of Matter, Vol. 7, Plenum Corp., New York. 1970.

8.Peacock, G. R., “A Review of Non-contact Process Temperature Measurement in Steel Manufacturing,” Part of the SPIE Conference on Thermosense XXI, pp. 171-189, April 1999.

9.Pellerin, M. A., “Multispectral Radiation Thermometry for Industrial Application,” PhD Thesis, Purdue University, school of Mechanical Engineering, Dec 1999.

10.Pellerin, M. A., “Multispectral Radiation Thermometry: Emissivity Compensation Algorithm,” M.S. Thesis, Purdue University, school of Mechanical Engineering, 1990.

11.Haugh M. J., “Radiation Thermometry in the Aluminum Industry,” Theory and Practice of Radiation Thermometry, D. P. DeWitt and G. D. Nutter, ed., John Wiley and Sons, New York, pp. 905-971, 1988.

12.Kanayama, K., Apparent Directional Emittance of V-Groove and Circular-Groove Rough Surface, Heat Transfer Japn. Res. 1, 11, 1972.

13.呂建財, “鋼材放射特徵研究與多光譜輻射測溫法之應用,” 碩士論文,國立成功大學機械系, 2007.