Comparative assessment on ammonia nitrogen adsorption onto a saline soil-groundwater environment: distribution, multi-factor interaction, and optimization using response surface methodology and artificial neural network

The adsorption of soil can reduce the leaching of NH₄⁺-N from the external environment into groundwater. The adsorption of NH₄⁺-N is affected by many factors. It is critical to use statistical model to quantitatively describe the effects of interaction between two or more factors on the system response. In this study, HJ-Biplot was used to analyze the correlation characteristics of soil water, salt, and nitrogen, and the response surface methodology and artificial neural network were used to statistically visualize the interaction between factors, including concentration, total dissolved solids (TDS), temperature, and pH. The results showed that the study soil was a typical saline soil, with maximum soil NH₄⁺-N content of 85.45 mg/kg. For the adsorption experiments of NH₄⁺-N on saline soils, the effects of factors on the adsorption capacity were assessed using the RSM model. The RSM model was coupled with an ANN to predict the adsorption of NH₄⁺-N by saline soils. The NH₄⁺-N concentration and water pH were both significant at a linear level (p < 0.0001). The interaction between NH₄⁺-N concentration and pH was also more significant (p < 0.01). Under optimal conditions (concentration: 800 mg/L; temperature: 24 °C; TDS: 637 mg/L; pH: 7.83), the NH₄⁺-N adsorption capacity was 1650.2 ug/g, which was in general agreement with the calculated values from the Box-Behnken and RSM model. In addition, a statistical error criterion for the model showed that the RSM-ANN model had greater predictive ability than RSM model.

Nonlinear autoregressive-moving average-L2 (NARMA-L2) controller for multivariable ball mill plant

Mineral processing facilities concern an enormous amount of dynamically complex unit operations (due to nonlinearities), for instance ball mill system. Normally, these processes need multivariable controllers to smooth actions by designing for plant constraints such as deadtimes and dynamics interactions. The present work presents a comparison between a classical PI and nonlinear moving average autoregressive-linearization level 2 (NARMA-L2) controllers based on artificial neural network (ANN) for a ball mill system. The manipulated variables of this plant are the rotation velocity (Vr) and the feeding weight (Wf), while the controlled parameters are the hold up (HU) and the mass fraction under 45 μm (P45). The simulation was built in the MATLAB software (Simulink), comparing the actions of PI and NARMA-L2 controllers in the face of operational changes in specific regions (constraints). The performance of proposed controllers was verified by the integral of absolute error (IAE), integral of squared error (ISE), or the integral of time-weighted absolute error (ITAE). The results of simulation showed the validity of the model obtained and the control technique proposed in this paper, which contributes to studies of multivariate controller designs for ball mills with significant applications. Additionally, this paper brings a first hybrid approach (PI/NARMA-L2) with successful implementation described in the literature.