US 12,281,029 B2
Model predictive control system and method using neural network to control dissolved oxygen and PH in seawater
Prafulchandra Bhavanbhai Sonara, Abqaiq (SA); and Abdulrahman S. Johar, Dammam (SA)
Assigned to Saudi Arabian Oil Company, Dhahran (SA)
Filed by Saudi Arabian Oil Company, Dhahran (SA)
Filed on Mar. 21, 2022, as Appl. No. 17/655,631.
Prior Publication US 2023/0295009 A1, Sep. 21, 2023
Int. Cl. C02F 1/00 (2023.01); C02F 1/20 (2023.01); C02F 1/68 (2023.01); C02F 101/10 (2006.01); C02F 103/08 (2006.01); G05B 13/02 (2006.01); G05B 13/04 (2006.01); G06N 20/00 (2019.01)
CPC C02F 1/008 (2013.01) [C02F 1/20 (2013.01); C02F 1/685 (2013.01); G05B 13/027 (2013.01); G05B 13/048 (2013.01); G06N 20/00 (2019.01); C02F 2101/10 (2013.01); C02F 2103/08 (2013.01); C02F 2209/06 (2013.01); C02F 2209/22 (2013.01); C02F 2305/00 (2013.01)] 17 Claims
OG exemplary drawing
 
1. A computer implemented method for model predictive control of a seawater treatment system for removing dissolved oxygen (DO) from seawater to produce treated seawater having less than a prescribed DO concentration and a prescribed pH, wherein an inlet provides untreated seawater into a vessel of the treatment system at an inflow rate which is controllable by a feed flow control valve (FCV), a N2 feed line delivers N2 gas into the vessel at an N2 flow rate which is controllable by an N2 FCV, an SO2 feed line delivers an SO2 solution into the vessel at an SO2 flow rate which is controllable by an SO2 FCV, and wherein the treated seawater is output from the vessel through an outlet, the method comprising the steps of:
receiving, at a processor of a controller computing device, wherein the processor is executing program code that is in the form of software modules and stored in a non-transitory storage medium, values of operational input parameters of the seawater treatment system including, the inflow rate, the N2 flow rate, the SO2 flow rate, a DO concentration of the treated seawater, and a pH of the treated seawater, wherein the inflow rate, the N2 flow rate and the SO2 flow rate are controllable operational input parameters, and wherein the software modules comprise a predictive machine learning model, and a model predictive control algorithm;
calculating, with the processor using the predictive machine learning model and based on the values of the operational input parameters, a predicted DO concentration and a predicted pH of the treated seawater for a future point in time;
determining, with the processor using the model predictive control algorithm and based on the predicted DO concentration and predicted pH and one or more of the operational input parameters, a setting for adjusting one of the controllable operational input parameters to change or maintain the DO concentration or pH of the treated seawater;
controlling, with the processor, operation of the seawater treatment system according to the setting, wherein the controlling step comprises sending, by the processor to the seawater treatment system over a communication interface, a control signal configured to adjust one or more of the feed FCV, N2 FCV and the SO2 FCV as a function of the setting;
monitoring, with the processor, at least the DO concentration and the pH of the treated seawater during operation of the seawater treatment system, wherein the DO concentration and the pH are measured using one or more sensors exposed to the treated seawater; and
based on the monitoring step, dynamically performing, with the processor during operation of the seawater treatment system, the calculating, determining and controlling steps such that the treated seawater output has less than the prescribed DO concentration and has less than the prescribed pH.