Anaerobic digestion is associated with various crucial variables, such as biogas yield, chemical oxygen demand, and volatile fatty acid concentration. Real-time monitoring of these variables can not only reflect the process of anaerobic digestion directly but also accelerate the efficiency of resource conversion and improve the stability of the reaction process. However, the current real-time monitoring equipment on the market cannot be widely used in the industrial production process due to its defects such as expensive equipment, low accuracy, and lagging analysis. Therefore, it is essential to conduct soft sensor modeling for unmeasurable variables and use auxiliary variables to realize real-time monitoring, optimization, and control of the an-aerobic digestion process. In this paper, the basic principle and process flow of anaerobic digestion are first briefly introduced. Subsequently, the development history of the traditional soft sensor is systematically reviewed, the latest development of soft sensors was detailed, and the obstacles of the soft sensor in the industrial production process are discussed. Finally, the future development trend of deep learning in soft sensors is deeply discussed, and future research directions are provided.

Anaerobic digestion technology is an effective way to solve the problem of urban kitchen waste. Volatile fatty acid (VFA) is an essential intermediate product in the anaerobic digestion process. Real-time monitoring of the concentration of VFA can not only directly reflect the anaerobic digestion process and improve resource conversion efficiency but also effectively avoid the reactor operation failure caused by acidification. Traditional soft sensors of VFA cannot adapt to the dynamic characteristics of industrial process. In this paper, a dynamic soft sensor based on spatiotemporal graph convolution network is developed. The spatial and temporal information of the anaerobic digestion process is extracted through graph convolutional network (GCN) and gated recurrent unit (GRU) respectively. With the purpose of further improving the prediction accuracy and generalization ability of the model, the adaptive adjacency matrix and graph attention mechanism are introduced into the model to solve the over-smoothing problem of GCN, and the semi-supervised learning mechanism based on manifold regularization is introduced to fully mine data information of unlabeled samples. Then, the gated unit is used to realize the information fusion of different dimensional features and the selection of model depth. At last, a dynamic soft sensor based on the semi-supervised gated spatiotemporal graph attention network is established to estimate the concentration of VFA in real time. Compared with baseline models such as GCN and GRU, the root mean square error of this model is reduced by 17.34% and 15.54%, respectively.

Anaerobic digestion technology is the most environmentally friendly approach to treat kitchen waste. Volatile fatty acid (VFA) is an essential quality monitoring indicator in the anaerobic digestion process of treating kitchen waste. In this paper, a soft measurement method of volatile fatty acid (VFA) concentration in the anaerobic digestion process is established based on stacked supervised auto-encoder combine kernel extreme learning machine algorithm (SSAE-KELM) to improve the real-time monitoring level and resource conversion efficiency of the anaerobic digestion process. Given the problems of poor feature extraction and low accuracy and efficiency of the model, a stack supervised autoencoder is proposed to realize nonlinear and deep feature extraction of process data. Simultaneously, using the idea of the extreme learning machine to train the network significantly improves the efficiency of the model. Then, the kernel extreme learning machine is used to realize regression modelling. Besides, a combined feature selection algorithm is presented to select auxiliary variables more accurately. The simulation results demonstrate that the soft sensor model can predict the concentration of volatile fatty acids (VFA) more efficiently and accurately.

Anaerobic digestion is a new method for treating kitchen waste, which can reduce waste, protect environment, and create clean energy. Because the concentration of volatile fatty acids (VFA) in kitchen waste anaerobic digestion process can not be measured in real time online, a soft measurement method based on deep belief network (DBN) is applied to the measurement of VFA. In this paper, extreme learning machine (ELM) is applied to the training of deep belief network. The adaptive learning rate is introduced to increase the convergence speed of the network, which are different from traditional deep belief network. The data from a real plant is classified and decomposed using Gaussian mixture model (GMM) and ensemble empirical mode analysis (EEMD) before training the network firstly. A deep belief network is used to perform numerical analysis on original data to extract features. Then the extracted features are input into extreme learning machine for training to obtain a soft measurement model. Experimental verification shows that this method is more precise than tradition methods and pure deep belief network model.

Kitchen waste is both a pollutant and an available resource. The anaerobic digestion (AD) technology can effectively treat waste and generate biogas. The concentration of volatile fatty acids (VFA) is one of the essential quality monitoring indicators in the anaerobic digestion process of kitchen waste and is difficult to measure online. In this paper, the hierarchical extreme learning machine (HELM) is applied to the soft sensor for the concentration of VFA considering the shortcomings of traditional soft sensor. The combined feature information is extracted by the multi-layer extreme learning machine-autoencoder (ELM-AE) structure in the HELM. Besides, the last layer is the combined manifold regularization extreme learning machine (ELM) and used to predict and mine the data structure of labeled data and unlabeled data. Considering that the characteristics of system operating conditions are prone to change in the actual production process of anaerobic digestion, the domain space transfer hierarchical extreme learning machine (DSTHELM) is used to solve the data drift problem caused by different initial conditions in the actual data; this is different from the traditional hierarchical extreme learning machine. This method increases the accuracy by 14.31% and 9.67% in training set and test set compared with HELM, indicating that the model possesses higher prediction accuracy and better generalization performance.