Hybrid Intelligent Model Based on Least Squares Support Vector Regression and Artificial Bee Colony Optimization for Time-Series Modeling and Forecasting Horizontal Displacement of Hydropower Dam

This chapter presents a hybrid intelligent model for time-series modeling and forecasting horizontal displacement of hydropower dams, named as ABC-LSSVR. In the proposed hybrid approach, least-squares support vector regression (LSSVR) was used to create the displacement model. Furthermore, the model was optimized using the Artificial Bee Colony (ABC) algorithm. The time-series monitoring data at the Hoa Binh hydropower dam of Vietnam were used as a case study. Accordingly, reservoir water lever, air temperature, ageing, and observed horizontal displacement, which are time-series monitoring data, were used to establish and verify the model. The experiment result showed that the ABC-LSSVR model has high accuracy with both the training and forecasting data and the ABC algorithm is capable of finding optimized parameters for the model autonomously. The effectiveness of the model is further assessed by comparing with results produced from two benchmark models, the support vector regression and the multilayer perceptron neural networks. Since the ABC-LSSVR model outperforms the two benchmark models, we concluded that the ABC-LSSVR model is a promising alternative tool that is highly recommended for modeling and forecasting the horizontal displacement of hydropower dams. © 2017 Elsevier Inc. All rights reserved.