半導體製造是用水密集型產業，晶圓製造中含有多道晶圓清洗步驟，其中會產生大量之廢水。大多數晶圓清洗之用的溶液含過氧化氫 (H2O2)，因此從製程排放出廢水含有高濃度的過氧化氫。高濃度的過氧化氫會降低廢水處理系統之整體效率並增加環境污染的風險。目前半導體廠普遍在廢水處理廠預處理步驟以多個幫浦決定過氧化氫酶的投放量，作為過氧化氫去除策略。由於預處理桶槽所含的廢水過於複雜。因此，過氧化氫濃度監測傳感器只能安裝於預處理的後端出口的管路，使得過氧化氫酶劑量的準確和優化設置變得困難。
若過氧化氫酶的投放量過量或不足，會導致不穩定的過氧化氫去除過程和環境問題。因此本研究建構紫式決策分析架構，整合深度學習模型與最佳化數學模型，以節省過氧化氫酶的用量並提升系統的穩定性。本研究架構分爲兩個階段：（1） 建構多步過氧化氫濃度的預測（direct multi-step ahead LSTM-NN），以結合長短期記憶(Long Short-Term Memory, LSTM) 及深度神經網絡（Deep Neural Network）模型預測未來20分鐘後的過氧化氫的濃度；（2）建構最佳化過氧化氫酶幫浦調度之優化模型。本研究以台灣某半導體製程廢水處理廠進行實證並檢驗模型效度。 結果顯示，本研究架構可以節省68.89%之過氧化氫酶消耗，同時增進半導體綠色生產與永續資源利用。

Semiconductor manufacturing is a water-intensive industry that generates tremendous amounts of wastewater during the wafer cleaning process. Since the majority of the solutions utilize in wafer cleaning process includes hydrogen peroxide (H2O2), the effluent discharged from the manufacturing site to the wastewater treatment plant comprises a high concentration of hydrogen peroxide. The high concentration of hydrogen peroxide will reduce the overall effectiveness of wastewater treatment process and raise the risk of environmental pollution. Catalase dosing via multiple pump is widely adopted as hydrogen peroxide removal strategy in pretreatment step of wastewater treatment plant. Due to, the wastewater contains in the pretreatment tank is too complex. Thus the hydrogen peroxide concentration monitoring sensor can only be mounted at the back end of the pretreatment pipeline, making the accurate and optimum setup for the catalase dosing difficult. Consequently, the activation of the catalase pump is not optimum, resulting in excess of catalase being utilized, an unstable hydrogen peroxide removal process, and environmental concerns. Focusing on the realistic needs, this study proposes the UNISON framework which incorporates the direct multi-step ahead prediction based on combination of long short-term memory and deep neural network (LSTM-NN) and a decision model for the catalase dosage pump to generate the precise and optimal setting for the catalase pump operation. An empirical investigation was carried out in the leading semiconductor industry. The result demonstrate the framework practical viability by successfully reducing 68.89% of the catalase consumption in the semiconductor wastewater treatment facility to achieve semiconductor green production and sustainable resource utilization.

Table of Contents----------------------------------------------------i
List of Tables-----------------------------------------------------iii
List of Figures-----------------------------------------------------iv
Notation Table-------------------------------------------------------v
Chapter 1 Introduction-----------------------------------------------1
1.1 Research Background and Motivation---------------------------1
1.2 Research Objective-------------------------------------------4
1.3 Thesis Organization------------------------------------------4
Chapter 2 Literature Review------------------------------------------5
2.1 Wastewater Treatment in Semiconductor Industry---------------5
2.2 Multi-step Ahead Forecasting of effluent concentration-------7
2.3 Pump scheduling for wastewater treatment---------------------9
Chapter 3 Research Framework----------------------------------------11
3.1 Understand and Define Problem-------------------------------15
3.2 Identify the Niche for Decision Quality Improvement---------15
3.3 Structure influence relation--------------------------------16
3.3.1 Data Preparation--------------------------------------16
3.3.2 LSTM-NN model for hydrogen peroxide prediction--------16
3.3.2.1 Making the data temporal----------------------------16
3.3.2.2 LSTM-NN architecture--------------------------------17
3.4 Sense and describe expecte outcomes-------------------------19
3.5 Overall judgements and value assessments--------------------20
3.5.1 Catalase Pump Optimization----------------------------20
3.6 Trade-off and decision--------------------------------------21
Chapter 4 Research Framework----------------------------------------22
4.1 Understand and Define Problem-------------------------------22
4.2 Identify the Niche for Decision Quality Improvement---------24
4.3 Structure Influence Relation--------------------------------24
4.3.1 Modeling----------------------------------------------24
4.3.2 Research Design---------------------------------------25
4.3.3 Data Preparation for Empirical Study------------------26
4.3.4 LSTM-NN forecasting-----------------------------------27
4.4 Sense and Describe the Outcomes-----------------------------30
4.5 Overall Judgement and Value Assessments---------------------31
4.5.1 Catalase Pump Optimization----------------------------31
4.6 Trade-off and Decision--------------------------------------35
Chapter 5 Conclusion------------------------------------------------37
References----------------------------------------------------------39

Abadi, M., P. Barham, J. Chen, Z. Chen, A. Davis, J. Dean, M. Devin, S. Ghemawat, G. Irving and M. Isard (2016). Tensorflow: A system for large-scale machine learning. 12th {USENIX} symposium on operating systems design and implementation ({OSDI} 16).
Abba, S., G. Elkiran and V. Nourani (2019). Non-linear Ensemble Modeling for Multi-step Ahead Prediction of Treated COD in Wastewater Treatment Plant. International Conference on Theory and Application of Soft Computing, Computing with Words and Perceptions, Springer.
Alameer, Z., A. Fathalla, K. Li, H. Ye and Z. Jianhua (2020). "Multistep-ahead forecasting of coal prices using a hybrid deep learning model." Resources Policy 65: 101588.
Aoudj, S., A. Khelifa and N. Drouiche (2017). "Removal of fluoride, SDS, ammonia and turbidity from semiconductor wastewater by combined electrocoagulation–electroflotation." Chemosphere 180: 379-387.
Assembly, U. G. (2015). Transforming our world: the 2030 agenda for sustainable development, resolution adopted by the general assembly, A/70/L. 1. New York: UN General Assembly.
Aviso, K. B., C.-F. Chien, M. K. Lim, R. R. Tan and M.-L. Tseng (2021). "Taiwan Drought was a Microcosm of Climate Change Adaptation Challenges in Complex Island Economies." Process Integration and Optimization for Sustainability 5(3): 317-318.
Candelieri, A., R. Perego and F. Archetti (2018). "Bayesian optimization of pump operations in water distribution systems." Journal of Global Optimization 71(1): 213-235.
Cao, J., Z. Li and J. Li (2019). "Financial time series forecasting model based on CEEMDAN and LSTM." Physica A: Statistical Mechanics and its Applications 519: 127-139.
Chandra, R., S. Goyal and R. Gupta (2021). "Evaluation of deep learning models for multi-step ahead time series prediction." IEEE Access 9: 83105-83123.
Chien, C.-F. and C.-C. Chen (2020). "Data-driven framework for tool health monitoring and maintenance strategy for smart manufacturing." IEEE Transactions on Semiconductor Manufacturing 33(4): 644-652.
Chien, C.-F., Y.-J. Chen, Y.-T. Han and Y.-C. Wu (2021). "Industry 3.5 for optimizing chiller configuration for energy saving and an empirical study for semiconductor manufacturing." Resources, Conservation and Recycling 168: 105247.
Chien, C.-F., Y.-J. Chen, C.-Y. Hsu and H.-K. Wang (2013). "Overlay error compensation using advanced process control with dynamically adjusted proportional-integral R2R controller." IEEE Transactions on Automation Science and Engineering 11(2): 473-484.
Chien, C.-F., Y.-S. Lin and S.-K. Lin (2020). "Deep reinforcement learning for selecting demand forecast models to empower Industry 3.5 and an empirical study for a semiconductor component distributor." International Journal of Production Research 58(9): 2784-2804.
Chien, C.-F., J.-T. Peng and H.-C. Yu (2016). "Building energy saving performance indices for cleaner semiconductor manufacturing and an empirical study." Computers & Industrial Engineering 99: 448-457.
Chien, C.-F., M.-L. Tseng, R. G. Tan, K. Tan and O. Velek (2021). Industry 3.5 for Sustainable Migration and Total Resource Management, Elsevier. 169: 105505.
Chung, S., J. Chung and C. Chung (2020). "Enhanced electrochemical oxidation process with hydrogen peroxide pretreatment for removal of high strength ammonia from semiconductor wastewater." Journal of Water Process Engineering 37: 101425.
Ci, Y., G. Xiu and L. Wu (2019). A Short-Term Traffic Flow Prediction Method Based on Long Short-Term Memory Network, Singapore, Springer Singapore.
Eng, C. Y., D. Yan, N. Withanage, Q. Liang and Y. Zhou (2019). "Wastewater treatment and recycle from a semiconductor industry: A demo-plant study." Water Practice and Technology 14(2): 371-379.
Fu, W. and C.-F. Chien (2019). "UNISON data-driven intermittent demand forecast framework to empower supply chain resilience and an empirical study in electronics distribution." Computers & Industrial Engineering 135: 940-949.
Goodfellow, I., Y. Bengio and A. Courville (2016). Deep learning, MIT press.
Griffiths, D. V. and I. M. Smith (2006). Numerical methods for engineers, CRC press.
Guo, H., K. Jeong, J. Lim, J. Jo, Y. M. Kim, J.-p. Park, J. H. Kim and K. H. Cho (2015). "Prediction of effluent concentration in a wastewater treatment plant using machine learning models." Journal of Environmental Sciences 32: 90-101.
Gurobi Optimization, L. (2021). "Gurobi Optimizer Reference Manual."
Hadi, M., F. Yakub, A. Fakhrurradzi, C. Hui, A. Najiha, N. Fakharulrazi, A. Harun, Z. Rahim and A. Azizan (2020). Designing Early Warning Flood Detection and Monitoring System via IoT. IOP Conference Series: Earth and Environmental Science, IOP Publishing.
Han, H.-G., H.-J. Zhang, Z. Liu and J.-F. Qiao (2020). "Data-driven decision-making for wastewater treatment process." Control Engineering Practice 96: 104305.
Han, H.-G. and S. Zhang (2017). An early warning system for MBR based on multi-step prediction and deep belief network classifier. 2017 Chinese Automation Congress (CAC), IEEE.
Hu, Y.-F., J.-L. Hou and C.-F. Chien (2019). "A UNISON framework for knowledge management of university–industry collaboration and an illustration." Computers & Industrial Engineering 129: 31-43.
Huang, C., B. Yang, K. Chen, C. Chang and C. Kao (2011). "Application of membrane technology on semiconductor wastewater reclamation: A pilot-scale study." Desalination 278(1-3): 203-210.
Hunter, J. D. (2007). "Matplotlib: A 2D graphics environment." Computing in science & engineering 9(03): 90-95.
Ikeda, S., K. Nemoto, M. Funabashi, T. Uchino, H. Yamamoto, N. Yabuoshi, Y. Sasaki, K. Komori, N. Suzuki and S. Nishihara (2003). "Process integration of single-wafer technology in a 300-mm fab, realizing drastic cycle time reduction with high yield and excellent reliability." IEEE Transactions on Semiconductor Manufacturing 16(2): 102-110.
Kagermann, H., J. Helbig, A. Hellinger and W. Wahlster (2013). Recommendations for implementing the strategic initiative INDUSTRIE 4.0: Securing the future of German manufacturing industry; final report of the Industrie 4.0 Working Group, Forschungsunion.
Kim, T.-K., T. Kim, I. Lee, K. Choi and K.-D. Zoh (2021). "Removal of tetramethylammonium hydroxide (TMAH) in semiconductor wastewater using the nano-ozone H2O2 process." Journal of Hazardous Materials 409: 123759.
Ku, C.-C., C.-F. Chien and K.-T. Ma (2020). "Digital transformation to empower smart production for Industry 3.5 and an empirical study for textile dyeing." Computers & Industrial Engineering 142: 106297.
Kuo, W. (1992). "Decolorizing dye wastewater with Fenton's reagent." Water Research 26(7): 881-886.
Lim, B. and S. Zohren (2021). "Time-series forecasting with deep learning: a survey." Philosophical Transactions of the Royal Society A 379(2194): 20200209.
Lin, K.-Y., C.-F. Chien and R. Kerh (2016). "UNISON framework of data-driven innovation for extracting user experience of product design of wearable devices." Computers & Industrial Engineering 99: 487-502.
Lin, S. H. and C. R. Yang (2004). "Chemical and physical treatments of chemical mechanical polishing wastewater from semiconductor fabrication." Journal of hazardous materials 108(1-2): 103-109.
McKinney, W. (2010). Data structures for statistical computing in python. Proceedings of the 9th Python in Science Conference, Austin, TX.
Moazeni, F. and J. Khazaei (2021). "Co-optimization of wastewater treatment plants interconnected with smart grids." Applied Energy 298: 117150.
Moore, G. E. (1965). Cramming more components onto integrated circuits, McGraw-Hill New York.
Niu, G., X. Yi, C. Chen, X. Li, D. Han, B. Yan, M. Huang and G. Ying (2020). "A novel effluent quality predicting model based on genetic-deep belief network algorithm for cleaner production in a full-scale paper-making wastewater treatment." Journal of Cleaner Production: 121787.
Oliphant, T. E. (2006). A guide to NumPy, Trelgol Publishing USA.
Pedregosa, F., G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss and V. Dubourg (2011). "Scikit-learn: Machine learning in Python." the Journal of machine Learning research 12: 2825-2830.
Pisa, I., I. Santín, J. L. Vicario, A. Morell and R. Vilanova (2019). "ANN-based soft sensor to predict effluent violations in wastewater treatment plants." Sensors 19(6): 1280.
Reinhardt, K. and W. Kern (2018). Handbook of silicon wafer cleaning technology, William Andrew.
Song, K., K. Sawayanagi, T. Numano, Y. Taniichi, T. Kikuchi, T. Takeda, H. Kanou, S. Riya, T. Hori and M. Hosomi (2019). "High-rate partial nitrification of semiconductor wastewater: Implications of online monitoring and microbial community structure." Biochemical Engineering Journal 143: 34-40.
Tseng, M. L., A. S. Chiu, W. Ashton and V. Moreau (2019). "Sustainable management of natural resources toward sustainable development goals." Resources Conservation And Recycling 145(ARTICLE): 419-421.
UNESCO, U. N. E. S. a. C. O. (2017). "Wastewater: the Untapped Resource. Durbane (South-Africa).".
Verhaverbeke, S., J. W. Parker and C. F. McConnell (1997). "The Role of HO2− in SC-1 Cleaning Solutions." MRS Online Proceedings Library (OPL) 477.
Xiao, Y., T. Chen, Y. Hu, D. Wang, Y. Han, Y. Lin and X. Wang (2014). "Advanced treatment of semiconductor wastewater by combined MBR–RO technology." Desalination 336: 168-178.
Yaqub, M., H. Asif, S. Kim and W. Lee (2020). "Modeling of a full-scale sewage treatment plant to predict the nutrient removal efficiency using a long short-term memory (LSTM) neural network." Journal of Water Process Engineering 37: 101388.
Yoon, D., K. Won, Y. Kim, B. Song, S. Kim, S.-J. Moon and B. Kim (2007). "Continuous removal of hydrogen peroxide with immobilised catalase for wastewater reuse." Water science and technology 55(1-2): 27-33.
Yu, C.-M., C.-F. Chien and C.-J. Kuo (2017). "Exploit the value of production data to discover opportunities for saving power consumption of production tools." IEEE Transactions on Semiconductor Manufacturing 30(4): 345-350.
Yunpeng, L., H. Di, B. Junpeng and Q. Yong (2017). Multi-step ahead time series forecasting for different data patterns based on LSTM recurrent neural network. 2017 14th web information systems and applications conference (WISA), IEEE.
Zhang, Z., A. Kusiak, Y. Zeng and X. Wei (2016). "Modeling and optimization of a wastewater pumping system with data-mining methods." Applied Energy 164: 303-311.
Zhou, Y., F.-J. Chang, L.-C. Chang, I.-F. Kao and Y.-S. Wang (2019). "Explore a deep learning multi-output neural network for regional multi-step-ahead air quality forecasts." Journal of Cleaner Production 209: 134-145