| CPC G06N 3/042 (2023.01) [B01D 19/0005 (2013.01); B01D 19/0063 (2013.01); C02F 1/008 (2013.01); C02F 9/00 (2013.01); C09K 8/845 (2013.01); E21B 41/00 (2013.01); G06N 3/10 (2013.01); C02F 1/001 (2013.01); C02F 1/20 (2013.01); C02F 1/66 (2013.01); C02F 2101/10 (2013.01); C02F 2103/08 (2013.01); C02F 2209/001 (2013.01); C02F 2209/006 (2013.01); C02F 2209/02 (2013.01); C02F 2209/06 (2013.01); C02F 2209/38 (2013.01); C02F 2209/40 (2013.01); C02F 2209/42 (2013.01)] | 20 Claims |

|
1. A computer-implemented method, comprising:
generating a model configured to predict a power of hydrogen (pH) of treated seawater output by a deaerator of a seawater processing plant, wherein generating the model comprises correlating process parameter values and historical data of the seawater processing plant of oil and gas reservoirs, the seawater processing plant comprising a temporary PH analyzer used for developing the model;
receiving, by a soft sensor pHI predictor installed at the seawater processing plant, upstream parameters of the seawater processing plant, the seawater processing plant having the temporary PH analyzer removed after developing the model;
predicting, using the model and a neural network software of the soft sensor pH predictor, the pH of treated seawater processed by the seawater processing plant;
removing dissolved oxygen (DO) from seawater to adjust the pH of treated seawater processed by the seawater processing plant to be within a target range pH of seawater; and
injecting, by a flow controller, the treated seawater into an oil and gas reservoir to maintain a pressure of the oil and gas reservoir.
|