Home / Current Issue / Paper 1714659
IoT-Powered Real-Time Demand Forecasting to Optimize Fuel & Material Supply Chains for Power Plants
Subject area: Science,Engineering and Technology · Area of research: ICT
Abstract
The power generation industry faces critical challenges in optimizing fuel and material supply chains due to fluctuating demand patterns, seasonal variations, and the increasing complexity of multi-source energy systems. Traditional forecasting methods struggle with the dynamic nature of power plant operations, leading to either excess inventory costs or critical fuel shortages that compromise generation capacity. This research investigates the integration of Internet of Things (IoT) sensors with advanced neural network forecasting models to enable real-time demand prediction and supply chain optimization for power plant operations. The study employs a hybrid forecasting framework combining IoT-enabled data collection systems with Recurrent Neural Networks (RNN), specifically leveraging Elman and Jordan network architectures, to predict fuel consumption patterns and material requirements. Data was collected from three coal-fired power plants and two combined-cycle gas turbine facilities over an 18-month period, capturing 2.4 million real-time sensor readings including fuel flow rates, combustion chamber temperatures, load demands, weather conditions, and maintenance schedules. The neural network models were trained using genetic algorithms and compared against traditional statistical methods including Multiple Discriminant Analysis (MDA) and conventional time-series forecasting. Results demonstrate that the IoT-powered RNN forecasting system achieved 94.3% accuracy in predicting daily fuel requirements, representing a 23.7% improvement over traditional methods. The system reduced fuel inventory holding costs by 31.2%, decreased stockout incidents by 47.8%, and improved supply chain responsiveness by enabling real-time adjustments to procurement schedules. For intermittent demand items such as spare parts and specialized materials, the Croston method integrated with neural networks achieved 87.6% forecast accuracy compared to 62.4% for conventional approaches. The IoT infrastructure enabled predictive maintenance scheduling, reducing unplanned outages by 38.4% through early detection of equipment degradation patterns. Economic analysis reveals annual cost savings of $3.2 million per 500MW facility through optimized inventory management, reduced emergency procurement, and improved operational efficiency. The research contributes theoretically by extending supply chain forecasting literature to the power generation context and demonstrating the synergistic benefits of IoT-neural network integration for handling lumpy and intermittent demand patterns characteristic of power plant operations.
Keywords
Internet of Things (IoT), Demand forecasting, Neural networks, Supply chain optimization, Power plant operations, Recurrent neural networks, Genetic algorithms, Fuel management, Predictive maintenance, Smart sensors
References
[1] W. Kong, Z. Y. Dong, Y. Jia, D. J. Hill, Y. Xu and Y. Zhang, "Short-Term Residential Load Forecasting Based on LSTM Recurrent Neural Network," in IEEE Transactions on Smart Grid, vol. 10, no. 1, pp. 841-851, Jan. 2018, doi: 10.1109/TSG.2017.2753802
[2] Gutierrez, R.S., Solis, A.O., & Mukhopadhyay, S. (2008). Lumpy demand forecasting using neural networks. International Journal of Production Economics, 111(2), 409-420.
[3] Han, S.-Y. (2004). Analysis of two-dimensional warranty data using artificial neural network. KAIST.
[4] Helms, M.M., Ettkin, L.P., & Chapman, S. (2000). Supply chain forecasting: Collaborative forecasting supports supply chain management. Business Process Management Journal, 6(5), 392-407.
[5] Hong, C.-Y., Park, J.-H., Yoon, T.-S., & Park, J.-B. (2004). A study on the Bayesian recurrent neural network for time series prediction. Institute of Control, Robotics and Systems, 10(12), 1295-1304.
[6] Hong, S.-G. (2000). Development of the forecasting model for parts in an automobile. Seoul National University.
[7] Hwang, Y.-S. (2012). A hybrid forecasting framework based on case-based reasoning and artificial neural network. Journal of Intelligence and Information Systems, 18(4), 43-57.
[8] Gubbi, J., Buyya, R., Marusic, S., & Palaniswami, M. (2013). Internet of Things (IoT): A vision, architectural elements, and future directions. Future Generation Computer Systems, 29(7), 1645–1660. https://doi.org/10.1016/j.future.2013.01.010
[9] Ghosh, A., & Das, S. K. (2008). Coverage and connectivity issues in wireless sensor networks: A survey. Pervasive and Mobile Computing, 4(3), 303–334. https://doi.org/10.1016/j.pmcj.2008.02.001
[10] Kim, Y.-T., & Kim, M.-H. (2001). The comparison of predict MDA with artificial neural networks for credit valuation of bonds: Focus on NASDAQ. Korean Accounting Journal, 10(4), 1-24.
[11] Kobayashi, I. (2008). Basics of artificial intelligence. Dream Media.
[12] Cai, S., & Jun, M. (2017). A qualitative study of the internalization of ISO 9000 standards: The linkages among firms’ motivations, internalization processes, and performance. International Journal of Production Economics, 196, 248–260. https://doi.org/10.1016/j.ijpe.2017.12.001
[13] Lee, K.-J. (2000). Online evolution of complex robot behavior using neurogenetic programming. Seoul University.
[14] Y. Bengio, A. Courville and P. Vincent, "Representation Learning: A Review and New Perspectives," in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 35, no. 8, pp. 1798-1828, Aug. 2013, doi: 10.1109/TPAMI.2013.50.
[15] Atzori, L., Iera, A., & Morabito, G. (2010). The Internet of Things: A survey. Computer Networks, 54(15), 2787–2805. https://doi.org/10.1016/j.comnet.2010.05.010
[16] Pham, D.T., & Karaboga, D. (1999). Training Elman and Jordan networks for system identification using genetic algorithms. Artificial Intelligence in Engineering, 13(2), 107-117.
[17] Ryu, I.-H. (2006). A comparative study of time series forecasting by artificial neural networks. Yonsei University.
[18] Simchi-Levi, D., Simchi-Levi, E., & Kaminsky, P. (2001). Designing and managing the supply chain: Concepts, strategies, and case studies. Journal of Business Logistics, 22(1), 1-30.
[19] Sun, J.-G., Park, J.-S., Khan, N.T., & Kim, Y. (2016). A study on developing forecasting methodology for parts including intermittent demands. Conference in Korean Institute of Industrial Engineers.
How to cite this paper
@article{1714659,
author = {Olusegun Gbolade, Daniel Ekwunife, Mayowa Jimoh, Samuel Ojo},
title = {IoT-Powered Real-Time Demand Forecasting to Optimize Fuel & Material Supply Chains for Power Plants},
journal = {Iconic Research And Engineering Journals},
year = {2018},
volume = {2},
number = {2},
pages = {187-206},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1714659.pdf},
abstract = {The power generation industry faces critical challenges in optimizing fuel and material supply chains due to fluctuating demand patterns, seasonal variations, and the increasing complexity of multi-source energy systems. Traditional forecasting methods struggle with the dynamic nature of power plant operations, leading to either excess inventory costs or critical fuel shortages that compromise generation capacity. This research investigates the integration of Internet of Things (IoT) sensors with advanced neural network forecasting models to enable real-time demand prediction and supply chain optimization for power plant operations. The study employs a hybrid forecasting framework combining IoT-enabled data collection systems with Recurrent Neural Networks (RNN), specifically leveraging Elman and Jordan network architectures, to predict fuel consumption patterns and material requirements. Data was collected from three coal-fired power plants and two combined-cycle gas turbine facilities over an 18-month period, capturing 2.4 million real-time sensor readings including fuel flow rates, combustion chamber temperatures, load demands, weather conditions, and maintenance schedules. The neural network models were trained using genetic algorithms and compared against traditional statistical methods including Multiple Discriminant Analysis (MDA) and conventional time-series forecasting. Results demonstrate that the IoT-powered RNN forecasting system achieved 94.3% accuracy in predicting daily fuel requirements, representing a 23.7% improvement over traditional methods. The system reduced fuel inventory holding costs by 31.2%, decreased stockout incidents by 47.8%, and improved supply chain responsiveness by enabling real-time adjustments to procurement schedules. For intermittent demand items such as spare parts and specialized materials, the Croston method integrated with neural networks achieved 87.6% forecast accuracy compared to 62.4% for conventional approaches. The IoT infrastructure enabled predictive maintenance scheduling, reducing unplanned outages by 38.4% through early detection of equipment degradation patterns. Economic analysis reveals annual cost savings of $3.2 million per 500MW facility through optimized inventory management, reduced emergency procurement, and improved operational efficiency. The research contributes theoretically by extending supply chain forecasting literature to the power generation context and demonstrating the synergistic benefits of IoT-neural network integration for handling lumpy and intermittent demand patterns characteristic of power plant operations.},
keywords = {Internet of Things (IoT), Demand forecasting, Neural networks, Supply chain optimization, Power plant operations, Recurrent neural networks, Genetic algorithms, Fuel management, Predictive maintenance, Smart sensors},
month = {August},
doi = {https://doi.org/10.64388/IREV2I2-1714659}
}