Process conditions of gas production in the electrochemical treatment of fracturing flowback fluid process based on support vector machine method

Gas production during the electrochemical treatment of waste liquids has potential safety risks. Herein, hydrogen (H2) production in such treatment was taken as the control target. Electrode combinations with relatively low H2 production were selected for electrochemical treatment during a fracturing flowback fluid experiment, in which mesh-like titanium–based ordinary ruthenium–iridium–palladium-coated electrode was the anode and plate–like titanium–based ordinary ruthenium–iridium–palladium-coated electrode was the cathode. The effects of electrolysis time (t), electrolytic current (I), electrode spacing (D), and other factors on H2 production in the electrolytic process were investigated. On the basis of the experimental data, a model was established using the support vector machine (SVM) method. Firstly, the three parameters of the radial basis function kernel of the model were simultaneously optimized using the quantum–particle swarm optimization algorithm as follows: penalty parameter = 256, kernel parameter = 0.0097039, and loss parameter = 0.014928. Then, an SVM regression model was established according to the three optimal parameters. The correlation coefficient was 0.98291 (r = 0.9914), and the mean–square error was 1.1883. The regression model was used to predict the technological conditions (t, I, and D) for the maximum/minimum H2 production in the electrochemical treatment of fracturing flowback fluid. The values were t = 50 min, I = 1.5 A, and D = 2.0 cm for the maximum H2 production, and t = 30 min, I = 0.5 A, and D = 6.0 cm for the minimum H2 production. Under these process conditions, the predicted maximum/minimum H2 production volumes were 145.04 and 20.47 mL, respectively.