• Energy Research

As an alternative to fossil fuels, biodiesel can be a source of clean and environmentally friendly energy source. However, its commercial application is limited by expensive feedstock and the slow nature of the pretreatment step-acid catalysis. The conventional approach to carry out this reaction uses stirred tank reactors. Recently, the lab-scale experiments using microbubble mediated mass transfer technology have demonstrated its potential use at commercial scale. However, all the studies conducted so far have been at a lab scale~100 mL of feedstock. To analyze the feasibility of microbubble technology, a larger pilot scale study is required. In this context, a kinetic study of microbubble technology at an intermediate scale is conducted (3 L of oil). Owing to the target for industrial application of the process, a commercial feedstock (Spirulina), microalgae oil (MO) and a commercial catalyst para-toluene sulfonic acid (PTSA) are used. Experiments to characterize the kinetics space (response surface, RSM) required for up-scaling are designed to develop a robust model. The model is compared with that developed by the gated recurrent unit (GRU) method. The maximum biodiesel conversion of 99.45 ± 1.3% is achieved by using these conditions: the molar ratio of MO to MeOH of 1:23.73 ratio, time of 60 min, and a catalyst loading of 3.3 wt% MO with an MO volume of 3 L. Furthermore, predicted models of RSM and GRU show proper fits to the experimental result. It was found that GRU produced a more accurate and robust model with correlation coefficient R2 = 0.9999 and root-mean-squared error (RSME) = 0.0515 in comparison with RSM model with R2 = 0.9844 and RMSE = 3.0832, respectively. Although RSM and GRU are fully empirical representations, they can be used for reactor up-scaling horizontally with microbubbles if the liquid layer height is held constant while the microbubble injection replicates along the floor of the reactor vessel—maintaining the tessellation pattern of the smaller vessel. This scaling approach maintains the local mixing profile, which is the major uncontrolled variable in conventional stirred tank reactor up-scaling.

As an alternative to fossil fuels, biodiesel can be a source of clean and environmentally friendly energy source. However, its commercial application is limited by expensive feedstock and the slow nature of the pretreatment step-acid catalysis. The conventional approach to carry out this reaction uses stirred tank reactors. Recently, the lab-scale experiments using microbubble mediated mass transfer technology have demonstrated its potential use at commercial scale. However, all the studies conducted so far have been at a lab scale~100 mL of feedstock. To analyze the feasibility of microbubble technology, a larger pilot scale study is required. In this context, a kinetic study of microbubble technology at an intermediate scale is conducted (3 L of oil). Owing to the target for industrial application of the process, a commercial feedstock (Spirulina), microalgae oil (MO) and a commercial catalyst para-toluene sulfonic acid (PTSA) are used. Experiments to characterize the kinetics space (response surface, RSM) required for up-scaling are designed to develop a robust model. The model is compared with that developed by the gated recurrent unit (GRU) method. The maximum biodiesel conversion of 99.45 ± 1.3% is achieved by using these conditions: the molar ratio of MO to MeOH of 1:23.73 ratio, time of 60 min, and a catalyst loading of 3.3 wt% MO with an MO volume of 3 L. Furthermore, predicted models of RSM and GRU show proper fits to the experimental result. It was found that GRU produced a more accurate and robust model with correlation coefficient R2 = 0.9999 and root-mean-squared error (RSME) = 0.0515 in comparison with RSM model with R2 = 0.9844 and RMSE = 3.0832, respectively. Although RSM and GRU are fully empirical representations, they can be used for reactor up-scaling horizontally with microbubbles if the liquid layer height is held constant while the microbubble injection replicates along the floor of the reactor vessel—maintaining the tessellation pattern of the smaller vessel. This scaling approach maintains the local mixing profile, which is the major uncontrolled variable in conventional stirred tank reactor up-scaling.

Coal-fired power plants have been used to meet the energy requirements in countries where coal reserves are abundant and are the key source of NOx emissions. Owing to the serious environmental and health concerns associated with NOx emissions, much work has been carried out to reduce NOx emissions. Sophisticated artificial intelligence (AI) techniques have been employed during the past few decades, such as least-squares support vector machine (LSSVM), artificial neural networks (ANN), long short-term memory (LSTM), and gated recurrent unit (GRU), to develop the NOx prediction model. Several studies have investigated deep neural networks (DNN) models for accurate NOx emission prediction. However, there is a need to investigate a DNN-based NOx prediction model that is accurate and computationally inexpensive. Recently, a new AI technique, convolutional neural network (CNN), has been introduced and proven superior for image class prediction accuracy. According to the best of the author’s knowledge, not much work has been done on the utilization of CNN on NOx emissions from coal-fired power plants. Therefore, this study investigated the prediction performance and computational time of one-dimensional CNN (1D-CNN) on NOx emissions data from a 500 MW coal-fired power plant. The variations of hyperparameters of LSTM, GRU, and 1D-CNN were investigated, and the performance metrics such as RMSE and computational time were recorded to obtain optimal hyperparameters. The obtained optimal values of hyperparameters of LSTM, GRU, and 1D-CNN were then employed for models’ development, and consequently, the models were tested on test data. The 1D-CNN NOx emission model improved the training efficiency in terms of RMSE by 70.6% and 60.1% compared to LSTM and GRU, respectively. Furthermore, the testing efficiency for 1D-CNN improved by 10.2% and 15.7% compared to LSTM and GRU, respectively. Moreover, 1D-CNN (26 s) reduced the training time by 83.8% and 50% compared to LSTM (160 s) and GRU (52 s), respectively. Results reveal that 1D-CNN is more accurate, more stable, and computationally inexpensive compared to LSTM and GRU on NOx emission data from the 500 MW power plant.

Coal-fired power plants have been used to meet the energy requirements in countries where coal reserves are abundant and are the key source of NOx emissions. Owing to the serious environmental and health concerns associated with NOx emissions, much work has been carried out to reduce NOx emissions. Sophisticated artificial intelligence (AI) techniques have been employed during the past few decades, such as least-squares support vector machine (LSSVM), artificial neural networks (ANN), long short-term memory (LSTM), and gated recurrent unit (GRU), to develop the NOx prediction model. Several studies have investigated deep neural networks (DNN) models for accurate NOx emission prediction. However, there is a need to investigate a DNN-based NOx prediction model that is accurate and computationally inexpensive. Recently, a new AI technique, convolutional neural network (CNN), has been introduced and proven superior for image class prediction accuracy. According to the best of the author’s knowledge, not much work has been done on the utilization of CNN on NOx emissions from coal-fired power plants. Therefore, this study investigated the prediction performance and computational time of one-dimensional CNN (1D-CNN) on NOx emissions data from a 500 MW coal-fired power plant. The variations of hyperparameters of LSTM, GRU, and 1D-CNN were investigated, and the performance metrics such as RMSE and computational time were recorded to obtain optimal hyperparameters. The obtained optimal values of hyperparameters of LSTM, GRU, and 1D-CNN were then employed for models’ development, and consequently, the models were tested on test data. The 1D-CNN NOx emission model improved the training efficiency in terms of RMSE by 70.6% and 60.1% compared to LSTM and GRU, respectively. Furthermore, the testing efficiency for 1D-CNN improved by 10.2% and 15.7% compared to LSTM and GRU, respectively. Moreover, 1D-CNN (26 s) reduced the training time by 83.8% and 50% compared to LSTM (160 s) and GRU (52 s), respectively. Results reveal that 1D-CNN is more accurate, more stable, and computationally inexpensive compared to LSTM and GRU on NOx emission data from the 500 MW power plant.