Development of Demand Prediction Simulator based on Multiple Water Resources

Transcription

1 Available online at ScienceDirect APCBEE Procedia 10 (2014 ) ICESD 2014: February 19-21, Singapore Development of Demand Prediction Simulator based on Multiple Water Resources Ji Young Kim, Jin Hong Jung, Young H. Yoon, Ju-Suk An, Weon Jae Kim and Hyun Je Oh Korea Institute of Construction Technology(, 283, Goyangdae-Ro, Ilsanseo-Gu, Goyang-Si, Gyeonggi-Do, , Korea Abstract One-way water supply systems which do not take into consideration the required demand are becoming an issue in terms of their efficient use of water. In this study, a demand prediction simulator was developed used to present the measures for increasing the water reuse ratio and reducing energy consumption. For its development, by using a multi-layer perception (MLP), a demand prediction module was developed, and an operating simulation module for the supply facilities of multiple water resources were implemented Published by Elsevier by Elsevier B.V. This B.V. is an Selection open access and/or article peer under review the CC under BY-NC-ND responsibility license of Asia-Pacific ( Chemical, Biological & Environmental Engineering Society Selection and peer review under responsibility of Asia-Pacific Chemical, Biological & Environmental Engineering Society Keywords: Multiple water resources, demand prediction, simulator 1. Introduction In the water supply process (water intake, water purification, and water drainage), the supply system which is irrelevant to required demand has water use efficiency problems such as an over-supply of water, increases in water leakage ratios, and the supply cost increases. Therefore, a flexible water securing and supply system is needed considering these problems. The goal of this study is to analyze water usage data per hour, and to predict the demand, and then, based on these, to develop a simulator for an hourly operating plan to achieve the water reuse target value and the Corresponding author. Tel.: ; fax: address: hjoh@kict.re.kr Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license ( Selection and peer review under responsibility of Asia-Pacific Chemical, Biological & Environmental Engineering Society doi: /j.apcbee

2 Ji Young Kim et al. / APCBEE Procedia 10 ( 2014 ) energy savings target value. 2. Methods The Research Institute at which we are currently working was selected as a test-bed. Also, the measured amount of water inflow included mineral water, water supplied from tap, groundwater inflow, and the seasonal flow of rainwater. The program was developed to be directly executed if there is Microsoft's DOT.NET Framework 4.0installed, without requiring a separate installation program. For the neural network models to predict the output data, multi-layer perception (MLP) was used. 3. Results 3.1. Development of the Demand Prediction module The demand prediction module can present learning data and the verification data through a linear regression model and the neural network model. For this to happen, a model optimizing algorithm through learning data was developed. The problems with water supply can be reviewed through the predicted output item data which was generated from the actual output item and the optimized model. For the learning data and the verification data, the value of the coefficient of determination should be at least 0.6 to be considered as providing satisfactory performance, wherein the higher the coefficient value of determination, the higher the accuracy of the model. The demand prediction module is shown in Fig.1. Fig. 1. Modeling results of learning data and verification data

3 226 Ji Young Kim et al. / APCBEE Procedia 10 ( 2014 ) Development of operating simulation module for supply facilities of multiple water resources Based on the operating data of the current facilities, the capacity of the facilities and the countermeasures for the water reuse target value and energy savings target values are analyzed and presented Setting the target value of the water reuse ratio The priorities for the simulation were classified into three categories A, B, and C, depending on intended use and water quality, as shown in Fig.2. The amount of reuse could be calculated by setting the maximum available amount to use for each type of demand. However, in case of the 3 rd priority C demand, it is not required to set maximum amounts, because it is assumed that the remaining portions would be used after using the 1 st and 2 nd priorities. Fig. 2. Modeling results of learning data and verification data for water reuse ratio Setting target values for energy savings In order to perform the analysis of energy savings, the average operating hours of current facilities, the amount of power consumption per hour, and off-peak / normal power billing rate information is required. For energy saving measures, the following two cases are presented, which are considered to be applicable conditions in the Research Institute. Energy saving measure to use midnight power Energy saving measure to use solar energy facilities Queries about analysis results The analysis results as shown in Fig.3 are displayed as in two cases; one is the analysis of the operating data based on the maximum use amount (CASE 1), and the other is that which is based on the average use amount (CASE 2). For each case, the available supply amounts depend upon demands A, B, and C which are calculated respectively. For each case, reviewing facility capacity, reviewing facility operation, and reviewing water quality improvement can be carried out. And by using these features, the expansion or reduction of the facilities can also be presented. In addition, power consumption amounts in accordance with the capacity variations of facilities and variations to the operating hours and measures to achieve energy savings targets can be provided accordingly.

4 Ji Young Kim et al. / APCBEE Procedia 10 ( 2014 ) Fig. 3. Analysis results from the operating simulation module for multiple water resources 4. Discussion First, by using the simulator, I could learn that the test-bed, the Research Institute building where we are employed has the possibility to apply water reuse. Research should be further continued by obtaining largely accumulated amounts of data and developing an optimized simulator. It is important to obtain a model that has optimized performance by reducing the number of hidden layers and nodes through demand prediction using neural network models. That s because it reduces the complexity of the model. It is necessary to select items having data variations and to exclude items whose values are unchanging, or predicting items (e.g. flux) whose variations are irrelevant at all, from the entry of the model. In addition, as much as possible, measures should be sought that consider event type factors such as rainfall. 5. Conclusions This simulator presents measures to increase the water reuse ratio and save energy. And therefore, the research is believed to be necessary in terms of not only environmental aspects, but also economic aspects. In the future, an expanded configuration of this simulator which can be applicable to both the buildings and cities will contribute to the establishment of the water supply system, which can solve problems of imbalance in the water supply and provide an emergency water supply system as well.

5 228 Ji Young Kim et al. / APCBEE Procedia 10 ( 2014 ) Acknowledgements This subject is supported by Korea Ministry of Environment as The Eco-Innovation project (Global Top project). (GT-SWS ) This research was supported by a grant from a Strategic Research Project ( ) funded by the Korea Institute of Construction Technology. References [1] Gee S. Choi, Young H. Choi, Sung K. Choi, Sang S. Lee, Myung G. Chun, Short-term Water Demand Forecasting with NN- TWI, Journal of KIIT. 7(5), 9-16, [2] Gu, Wenquan, Dongguo Shao, Yufang Jiang, Water demand simulation and water shortage risk analysis for sources area of water diversion engineering. Shuili Fadian Xuebao(Journal of Hydroelectric Engineering), 31(5), 23-28, [3] Hall, J. W., Watts, G., Keil, M., de Vial, L., Street, R., Conlan, K., O'Connell, P. E., Beven, K. J., Kilsby, C. G., Urban water demand forecasting and uncertainty assessment using ensemble wavelet-bootstrap-neural network models. Water and Environment Journal, 26 (1), , [4] Herrera, M., Torgo, L., Izquierdo, J., Pérez-García, R., Predictive models for forecasting hourly urban water demand. Journal of Hydrology, 387(1-2), , [5] Jain, A., Varshney, A. K., Joshi, U. C., Short-Term Water Demand Forecast Modelling at IIT Kanpur Using Artificial Neural Networks. Water Resources Management, 15(5), , [6] Kanta, L., Zechman, E., A Complex Adaptive Systems Framework to Assess Supply-side and Demand-side Management for Urban Water Resources. Journal of Water Resources Planning and Management / (ASCE) WR [7] Lily A. House-Peters, Heejun Chang, Urban water demand modeling: Review of concepts, methods, and organizing principles. Water Resources Research, 47(5), [8] Mark W. Rosegrant & Ximing Cai, Global Water Demand and Supply Projections. Water International, 27(2), , [9] Raskin, P., Hansen, E., Zhu, Z., Stavisky, D., Simulation of Water Supply and Demand in the Aral Sea Region. Water International, 17(2), 55-67, [10] Salvatore Campisi-Pinto, Jan Adamowski, Gideon Oron, Forecasting Urban Water Demand Via Wavelet-Denoising and Neural Network Models. Case Study: City of Syracuse, Italy. Water Resour Manage, 26 (12), , [11] Zhou, S.L., McMahon, T.A., Walton, A., Lewis, J., Forecasting daily urban water demand: a case study of Melbourne. Journal of Hydrology, 236(3-4), , [12] Zhou, S.L., McMahonv, T.A., Walton, A., Lewis, J., Forecasting operational demand for an urban water supply zone. Journal of Hydrology, 259(1-4), , 2002.