An approach to solving the problem associated with the calculation of the optimal parameters of the technological process in the field oil treatment plants is considered. The calculation algorithm is based on the dynamic programming method that implements the Bellman optimality principle with an additive optimality criterion. The generalized criterion of the problem is formed as a function of the sum of local reduced costs for the stages of oil treatment. The proposed method of solving the problem allows to predict the optimal parameters of the process regime and transfer them to the control system as tasks to operators and automatic controllers.
Keywords: field oil treatment, facility, technological regime, model, optimization, operative management
The method of managing the recipe of dry magnesia cement mixtures (DMCM) during their production, intended for cement slurries for cementing oil and gas wells, is characterized in that the recipe of the mixture is selected based on the requirements for the characteristics of the cement slurry and the quality of the cement stone, which are determined in the order in relation to the mining and geological conditions of a particular well. The method is based on solving multiobjective optimization problem in which as criteria protrude partial deflection characteristics of the resulting solution of the SMTS from predetermined values in order for the production batch. Generalized criterion problem is formed as an additive objective function of the weighted partial criteria.
Keywords: oil and gas wells, rheological characteristics of cement slurry, dry magnesia mixture, experiment planning, regression models of communication, optimization of mixture formulation, generalized criterion
Results of research on control object identification based on neural network processes modeling are given. Model of object with control system is represented with a dynamic neural network and regulator model. Regulator function is known. Neural network is trained on the data of control object operating. The resulting model simulate the behavior of the system and lets us find the system’s output, including outputs for periodic test influences. By the resulting complex frequency response we find the parameters of the channel. Observed objects represent technological processes with continuous production. We show an example of identification for laboratory control object channel.
Keywords: Object with control system, identification, neural network, modeling, complex frequency response, transfer function