A Developed Method for CO 2 Hydrate Prediction Conditions SeyyedMohammadreza Hesami 1 , Zakarya Kamali, Majid Zendedel Siuki, Seyed Borhan Armand, Emad Jamshidi 1 Department of Chemical engineering, Petroleum University of Technology Smr92hesami@gmail.com Abstract Gas hydrates are easily formed during the transportation of oil and gas when it contains a certain amount of water. This paper describes the development and application of Artificial Neural Networks (ANN) that optimized by imperialist competitive algorithm (ICA) for the prediction of natural gas hydrates formation Temperature. In this work after training so ICA-ANN models we find an optimum one that have been used to accurately simulating carbon dioxide hydrate dissociation Temperature. Hence, reliable data of Co 2 were used as input in obtained ICA-ANN model. The obtained results are in the minimum error in compare to experimental. Keywords: Gas Hydrate, dissociation temperature, ICA-ANN, modeling, optimization Introduction Gas hydrates are a group of non-stoichiometric, ice-like crystalline compounds formed through a combination of water and suitably sized guest molecules under low-temperatures and elevated pressures[1]. Hydrate formation can cause pipeline blockages and other operational problems in petroleum industry[2]. To inhibit formation of gas hydrates, glcols, and alcohols are normally used [3]. Several studies have been performed on the prediction of hydrate formation conditions for various gas mixtures and inhibitors [4-7]; in this paper, the ability of the artificial intelligence in establishing and predicting hydrate formation condition is to be investigated. Artificial intelligence have been widely used and are gaining attention in petroleum engineering because of their ability to solve problems that previously were difficult or even impossible to solve[8]. Artificial neural network (ANN) could be used as soft sensor building approach[9]. Diferent algorithms can evolve neural network at various levels: weight training, architecture adaptation (including number of hidden layers, number of hidden neurons and node transfer functions) and learning rules [10]. In this paper, we propose an evolutionary algorithm for optimizing the weights of feed-forward ANNs called Imperialist Competitive Algorithm (ICA). This optimization algorithm is inspired by imperialistic competition [10]. In order to create ICA-ANN model, version 7.10.0.499 (R2010a) of MATLAB was used. Optimization ICA is a new optimization algorithm whose efficiency for solving some of benchmark cost functions has been proved. This algorithm, which was first introduced by ز ایرانفت و گاش ملی و همای اولیه آبان1392