Method for controlling a plant of separation and treatment industrial processes without chemical reaction
The present invention refers to a method for controlling a plant of separation and treatment industrial processes without chemical reaction using artificial intelligence and machine learning, aiming at improving revenues and profits obtained, as well as the performance of the system, and the technique can be applied in steps of conceptual design for a unit in operation, comprising the steps of: defining objectives and gains of the plant; delimiting the plant; evaluation in steady state of the plant; evaluation in dynamic state of the plant; and performing non-linear dynamic simulation of the plant.
1 . A method for controlling a separation and treatment plant, the method comprising:
defining an objective function of a separation and treatment plant for industrial processes without chemical reaction;
delimiting the separation and treatment plant by:
defining the boundary conditions of the objective function;
identifying models of a process for a dynamic and static evaluation of the process based on the objective function; and
identifying processes to use neural networks or deep learning networks to create a behavior model of each process;
evaluating a steady state of the separation and treatment plant;
evaluating a dynamic state of the separation and treatment plant; and
performing a non-linear dynamic simulation of the separation and treatment plant,
wherein evaluating the dynamic state of the separation and treatment plant comprises:
evaluating a stability of one or more control loops according to a Routh-Hurwitz or Ultimate-Gain criteria;
evaluating a performance of the one or more control loops based on an analysis criteria of error integrals and deviation in relation to plant restrictions;
generating and evaluating one or more candidate setpoints for maximum production points using genetic algorithms;
modifying necessary controls to operate in ranges defined by the genetic algorithms; and
reviewing the one or more control loops and implementing the one or more control loops.
2 . The method of claim 1 , wherein the objective function of the separation and treatment plant comprises characteristics including one or more of processes, equipment, products and inputs of the separation and treatment plant.
3 . The method of claim 2 , wherein the objective function is a production quantity.
4 . The method of claim 1 , wherein the boundary conditions comprise one or more of: degrees of freedom, restrictions, limits, and expected conditions of the separation and treatment plant.
5 . The method of claim 1 , further comprising obtaining the boundary conditions of the objective function from handbooks of plant equipment manufacturers and from technical information of similar plants in operation.
6 . The method of claim 4 , further comprising determining the degrees of freedom as a function of process models and variables obtained from the process of the separation and treatment plant.
7 . The method of claim 1 , wherein evaluating the steady state of the separation and treatment plant comprises:
defining one or more variables to be controlled; and
evaluating a productivity and a sensitivity of the one or more variables to be controlled against a production performed for each condition foreseen in the process.
8 . The method of claim 7 , wherein evaluating the steady state of the separation and treatment plant further comprises evaluating the sensitivity of the one or more variables to be controlled by implementation of the one or more control loops.
9 . The method of claim 1 , wherein performing the non-linear dynamic simulation of the separation and treatment plant comprises evaluating the one or more candidate setpoints for maximizing the objective function without reaching a limits of the process.