For effective monitoring and control of the fermentation process, an accurate real-time measurement of important variables is necessary. These variables are very hard to measure in real-time due to constraints such as the time-varying, nonlinearity, strong coupling, and complex mechanism of the fermentation process. Constructing soft sensors with outstanding performance and robustness has become a core issue in industrial procedures. In this paper, a comprehensive review of existing data pre-processing approaches, variable selection methods, data-driven (black-box) soft-sensing modeling methods and optimization techniques was carried out. The data-driven methods used for the soft-sensing modeling such as support vector machine, multiple least square support vector machine, neural network, deep learning, fuzzy logic, probabilistic latent variable models are reviewed in detail. The optimization techniques used for the estimation of model parameters such as particle swarm optimization algorithm, ant colony optimization, artificial bee colony, cuckoo search algorithm, and genetic algorithm, are also discussed. A comprehensive analysis of various soft-sensing models is presented in tabular form which highlights the important methods used in the field of fermentation. More than 70 research publications on soft-sensing modeling methods for the estimation of variables have been examined and listed for quick reference. This review paper may be regarded as a useful source as a reference point for researchers to explore the opportunities for further enhancement in the field of soft-sensing modeling.

L-Lysine is produced by a complex non-linear fermentation process. A non-linear model predictive control (NMPC) scheme is proposed to control product concentration in real time for enhancing production. However, product concentration cannot be directly measured in real time. Least-square support vector machine (LSSVM) is used to predict product concentration in real time. Grey-Wolf Optimization (GWO) algorithm is used to optimize the key model parameters (penalty factor and kernel width) of LSSVM for increasing its prediction accuracy (GWO-LSSVM). The proposed optimal prediction model is used as a process model in the non-linear model predictive control to predict product concentration. GWO is also used to solve the non-convex optimization problem in non-linear model predictive control (GWO-NMPC) for calculating optimal future inputs. The proposed GWO-based prediction model (GWO-LSSVM) and non-linear model predictive control (GWO-NMPC) are compared with the Particle Swarm Optimization (PSO)-based prediction model (PSO-LSSVM) and non-linear model predictive control (PSO-NMPC) to validate their effectiveness. The comparative results show that the prediction accuracy, adaptability, real-time tracking ability, overall error and control precision of GWO-based predictive control is better compared to PSO-based predictive control.

Due to the characteristics of uncertainty, multivariate, strong coupling and non-linearity in the marine alkaline protease MP fermentation process, the realization of high-performance decoupling control has always been the goal. In this thesis, a nonlinear robust decoupling control method based on artificial bee colony algorithm and multiple least squares support vector machine inversion is proposed, which combines the inverse system method with support vector machine theory. Firstly, the "gray box" dynamics model of marine alkaline protease MP fermentation process is developed based on the material balance relationship of the fermentation process, and the existence of the inverse system is analyzed. Secondly, the inverse model is constructed off-line using least squares support vector machine, and the inverse model is corrected on-line by artificial bee colony algorithm according to the model deviation. Finally, the inverse model is connected in series with the original system to form a combining pseudo-linear system, and according to the characteristics of the pseudo-linear system, the internal model controller is introduced to perform closed-loop control on the system. The simulation results show that the control method can realize dynamic decoupling control of marine alkaline protease fermentation process, and the system has strong parameter robustness and good anti-interference ability.

Due to the high degree of strong coupling and nonlinearity of marine lysozyme fermentation process, it is difficult to accurately model the mechanism. In order to achieve real‐time online measurement and effective control of bacterial concentration during fermentation, a generalized predictive control method based on least squares support vector machines is proposed. The particle swarm optimization least squares support vector machine (PSO‐LS‐SVM) model of lysozyme concentration is established by optimizing the regularization parameters and the kernel parameters of the least squares support vector machine by particle swarm optimization. To avoid the nonlinear problems in predictive control, the model is linearized at each sampling point and the generalized predictive algorithm is used to predict the bacteria concentration of lysozyme. The experimental simulation shows that the least squares support vector machine model with particle swarm optimization can achieve good prediction effect. The linearized model performs generalized predictive control, which makes the total activity of the enzyme increased from 60% to 80% and the yield improved by 30%.

The problems that the key biomass variables in Pichia pastoris fermentation process are difficult measure in real time; this paper mainly proposes a multi-model soft sensor modeling method based on the piecewise affine (PWA) modeling method, which is optimized by particle swarm optimization (PSO) with an improved compression factor (ICF). Firstly, the false nearest neighbor method was used to determine the order of the PWA model. Secondly, the ICF-PSO algorithm was proposed to cooperatively optimize the number of PWA models and the parameters of each local model. Finally, a least squares support vector machine was adopted to determine the scope of action of each local model. Simulation results show that the proposed ICF-PSO-PWA multi-model soft sensor modeling method accurately approximated the nonlinear features of Pichia pastoris fermentation, and the model prediction accuracy is improved by 4.4884% compared with the weighted least squares vector regression model optimized by PSO.

For Pichia pastoris fermentation process with multi-operating conditions, it is difficult to predict the cell concentration under the new operating conditions by the soft sensor model established under the specific operating conditions. Inspired by the idea of transfer learning, a method based on an improved balanced distribution adaptive regularization extreme learning machine (IBDA-RELM) was proposed to solve the problem. The domain adaptation (DA) method in transfer learning is developed to reduce distribution distance by transforming data. However, the joint distribution adaptation (JDA) and the balanced distribution adaptation (BDA) in DA cannot be directly applied to regression problems. The fuzzy sets (FSs) method was proposed to solve this issue. Finally, a soft sensor model of Pichia pastoris cell concentration was realized by inputting the converted data to the RELM model. Simulation verification was carried out with three operating conditions at the scene of fermentation. The transfer effects of three DA methods, including transfer component analysis (TCA), improved joint distribution adaptation (IJDA) as well as IBDA, were compared. The predicted results show that IBDA-RELM had a better performance in the soft sensor of Pichia pastoris cell concentration under multi-operating conditions.