An Agent-Based Hybrid Intelligent System for Petroleum Reservoir Characterisation

C. Li,Zili Zhang,Fanhua Shang
2005-01-01
Computational Intelligence
Abstract:Petroleum reservoir characterisation is a new practical technology for synthetically studying and evaluating petroleum reservoirs. Since the complexity of the underground geology, a single intellige,nt technique can not solve the complicated and elaborate geologic problems. It is necessary that those geologic problems are synthetically studied by combining multiple intelligent techniques. We discuss an agent-based hybrid intelligent system for solving petroleum reservoir characterisation. The system starts with the digitisation of well logs parameter graphs based on an expert system. The identification of lithology and the prediction of porosity and permeability can be conducted by using hybrid intelligent techniques. Four intelligent technique agents (the complicated lithology identification with parthenogenetic algorithm, the porosity prediction with neural network, the permeability estimation with jitzzy neural network, and the well logs curve-digitizing with expert system) have been integrated The prototype of the petroleum reservoir characterisation system has been implemented by using C language and Socket technique. The legacy resources can be used in the system by means of middleware agents. The robustness of the system is facilitated through the use of the redundant middle agents with ring-based architectural model.
What problem does this paper attempt to address?