A deep reinforcement learning approach to gasoline blending real-time optimization under uncertainty

The gasoline inline blending process has widely used real-time optimization techniques to achieve optimization objectives,such as minimizing the cost of production.However,the effectiveness of real-time optimization in gasoline blending relies on accurate blending models and is challenged by stochastic disturbances.Thus,we propose a real-time optimization algorithm based on the soft actor-critic(SAC)deep reinforcement learning strategy to optimize gasoline blending without relying on a single blending model and to be robust against disturbances.Our approach constructs the environment using nonlinear blending models and feedstocks with disturbances.The algorithm incorporates the Lagrange multiplier and path constraints in reward design to manage sparse product constraints.Carefully abstracted states facilitate algorithm convergence,and the normalized action vector in each optimization period allows the agent to generalize to some extent across different target production scenarios.Through these well-designed components,the algorithm based on the SAC outperforms real-time optimization methods based on either nonlinear or linear programming.It even demonstrates comparable performance with the time-horizon based real-time optimization method,which requires knowledge of uncertainty models,confirming its capability to handle uncertainty without accurate models.Our simulation illustrates a promising approach to free real-time optimization of the gasoline blending process from uncertainty models that are difficult to acquire in practice.

Blending Scheduling under Uncertainty Based on Particle Swarm Optimization Algorithm

Blending is an important unit operation in process industry. Blending scheduling is nonlinear optimiza- tion problem with constraints. It is difficult to obtain optimum solution by other general optimization methods. Particle swarm optimization (PSO) algorithm is developed for nonlinear optimization problems with both contin- uous and discrete variables. In order to obtain a global optimum solution quickly, PSO algorithm is applied to solve the problem of blending scheduling under uncertainty. The calculation results based on an example of gasoline blending agree satisfactory with the ideal values, which illustrates that the PSO algorithm is valid and effective in solving the blending scheduling problem.

Global Optimization of Nonlinear Blend-Scheduling Problems

The scheduling of gasoline-blending operations is an important problem in the oil refining industry. Thisproblem not only exhibits the combinatorial nature that is intrinsic to scheduling problems, but alsonon-convex nonlinear behavior, due to the blending of various materials with different quality properties.In this work, a global optimization algorithm is proposed to solve a previously published continuous-timemixed-integer nonlinear scheduling model for gasoline blending. The model includes blend recipe optimi-zation, the distribution problem, and several important operational features and constraints. The algorithmemploys piecewise McCormick relaxation （PMCR） and normalized multiparametric disaggregation tech-nique （NMDT） to compute estimates of the global optimum. These techniques partition the domain of oneof the variables in a bilinear term and generate convex relaxations for each partition. By increasing the num-ber of partitions and reducing the domain of the variables, the algorithm is able to refine the estimates ofthe global solution. The algorithm is compared to two commercial global solvers and two heuristic methodsby solving four examples from the literature. Results show that the proposed global optimization algorithmperforms on par with commercial solvers but is not as fast as heuristic approaches.

A New Sample-Selection and Modeling Method Based on Near-Infrared Spectroscopy and Its Industrial Application

Near-infrared( NIR) spectroscopy has been widely employed as a process analytical tool( PAT) in various fields; the most important reason for the use of this method is its ability to record spectra in real time to capture process properties. In quantitative online applications,the robustness of the established NIR model is often deteriorated by process condition variations,nonlinear of the properties or the high-dimensional of the NIR data set. To cope with such situation,a novel method based on principal component analysis( PCA) and artificial neural network( ANN) is proposed and a new sample-selection method is mentioned. The advantage of the presented approach is that it can select proper calibration samples and establish robust model effectively. The performance of the method was tested on a spectroscopic data set from a refinery process. Compared with traditional partial leastsquares( PLS),principal component regression( PCR) and several other modeling methods, the proposed approach was found to achieve good accuracy in the prediction of gasoline properties. An application of the proposed method is also reported.

Effects of different mixing ratios on emissions from passenger cars fueled with methanol/gasoline blends

Regulated and unregulated emissions from four passenger cars fueled with methanol/gasoline blends at different mixing ratios （M15,M20,M30,M50,M85 and M100） were tested over the New European Driving Cycle （NEDC）.Volatile organic compounds （VOCs） were sampled by Tenax TA and analyzed by thermal desorption-gas chromatograph/mass spectrometer （TD-GC/MS）.Carbonyls were trapped on dinitrophenylhydrazine （DNPH） cartridges and analyzed by high performance liquid chromatography （HPLC）.The results showed that total emissions of VOCs and BTEX （benzene,toluene,ethylbenzene,p,m,o-xylene） from all vehicles fueled with methanol/gasoline blends were lower than those from vehicles fueled with only gasoline.Compared to the baseline,the use of M85 decreased BTEX emissions by 97.4%,while the use of M15 decreased it by 19.7%.At low-to-middle mixing ratios （M15,M20,M30 and M50）,formaldehyde emissions showed a slight increase while those of high mixing ratios （M85 and M100） were three times compared with the baseline gasoline only.When the vehicles were retrofitted with new three-way catalytic converters （TWC）,emissions of carbon monoxide （CO）,total hydrocarbon （THC）,and nitrogen oxides （NOx） were decreased by 24%–50%,10%–35%,and 24%–58% respectively,compared with the cars using the original equipment manufacture （OEM） TWC.Using the new TWC,emissions of formaldehyde and BTEX were decreased,while those of other carbonyl increased.It is necessary that vehicles fueled with methanol/gasoline blends be retrofitted with a new TWC.In addition,the specific reactivity of emissions of vehicles fueled with M15 and retrofitted with the new TWC was reduced from 4.51 to 4.08 compared to the baseline vehicle.This indicates that the use of methanol/gasoline blend at a low mixing ratio may have lower effect on environment than gasoline.