Chemical process simulation has been widely used in the petrochemical industry. This has been the main means of process optimization and aided design. The process parameters in a chemical process are diverse and complicated. It is difficult for traditional optimization methods to achieve global optimization by sensitivity analysis and optimization of a small number of key parameters. Therefore, an optimization method for simulating operating parameters of chemical processes based on particle swarm optimization algorithm is proposed in the present paper. The natural gas decarbonization process is chosen as the research object, the process simulation and optimization algorithm are coupled using Aspen HYSYS software. Combined with the knowledge of the process mechanism, the optimization of operation parameters of the natural gas decarbonization steadystate process simulation based on a particle swarm optimization algorithm has been achieved. Under the condition that the product meets the process requirements, and the controllable operation parameters that have a great influence on the process are used as the decision variables, the operation parameters of a 5.8×10⁶ m³/d natural gas purification unit are optimized by taking the maximum decarbonization rate, the minimum operation cost of the unit as the objective function. The optimization results show that fewer plates in the absorption tower and regeneration tower can meet the needs of acid gas removal requirements. Under the condition that the each tray is in a good operating state, the reflux ratio of the regeneration tower is reduced compared with the original process, and the gas-liquid phase load is also reduced to a certain extent, resulting in a decrease of the reboiler load. The temperature of the lean amine liquid into the absorption tower is lower than the original process, so that the positive reaction degree of CO₂ with the alcohol amine liquid is increased, and the increased absorption driving force slows down the corrosion of the equipment. The pressure in the absorption tower is increased compared with the original process, which increases the mass transfer driving force in the tower and the purification of the gas. Based on the particle swarm optimization algorithm for the natural gas decarbonization process, the carbon dioxide content in the purified gas is reduced from 0.16 mol% to 0.05 mol%, and the annual energy consumption cost is reduced by about 13%. The method proposed in the present work can find the global optimal operation scheme quickly and automatically without human involvement, and can be flexibly extended to the process optimization of various industrial processes.

Coking at the fractionating tower bottom and the decant oil circulation system disrupts the heat balance, leading to unplanned shutdown and destroying the long period stable operation of the Fluid Catalytic Cracking Unit (FCCU). The FCCU operates through interconnected subsystems, generating high-dimensional, nonlinear, and non-stationary data characterized by spatiotemporally correlated. The decant oil solid content is the crucial indicator for monitoring catalyst loss from the reactor-regenerator system and coking risk tendency at the fractionating tower bottom that relies on sampling and laboratory testing, which is lagging responsiveness and labor-intensive. Developing the online decant oil solid content soft sensor using industrial data to support operators in conducting predictive maintenance is essential. Therefore, this paper proposes a hybrid deep learning framework for soft sensor development that combines spatiotemporal pattern extraction with interpretability, enabling accurate risk identification in dynamic operational conditions. This framework employs a Filter-Wrapper method for dimensionality reduction, followed by a 2D Convolutional Neural Network (2DCNN) for extracting spatial patterns, and a Bidirectional Gated Recurrent Unit (BiGRU) for capturing long-term temporal dependencies, with an Attention Mechanism (AM) to highlight critical features adaptively. The integration of SHapley Additive exPlanations (SHAP), Complementary Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), 2DCNN, and expert knowledge precisely quantifies feature contributions and decomposes signals, significantly enhancing the practicality of risk identification. Applied to a China refinery with processing capacity of 2.80 × 10⁶ t/a, the soft sensor achieved the R² value of 0.93 and five-level risk identification accuracy of 96.42%. These results demonstrate the framework's accuracy, robustness, and suitability for complex industrial scenarios, advancing risk visualization and management.

Since chemical processes are highly non-linear and multiscale, it is vital to deeply mine the multiscale coupling relationships embedded in the massive process data for the prediction and anomaly tracing of crucial process parameters and production indicators. While the integrated method of adaptive signal decomposition combined with time series models could effectively predict process variables, it does have limitations in capturing the high-frequency detail of the operation state when applied to complex chemical processes. In light of this, a novel Multiscale Multi-radius Multi-step Convolutional Neural Network (MsrtNet) is proposed for mining spatiotemporal multiscale information. First, the industrial data from the Fluid Catalytic Cracking (FCC) process decomposition using Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) extract the multi-energy scale information of the feature subset. Then, convolution kernels with varying stride and padding structures are established to decouple the long-period operation process information encapsulated within the multi-energy scale data. Finally, a reconciliation network is trained to reconstruct the multiscale prediction results and obtain the final output. MsrtNet is initially assessed for its capability to untangle the spatiotemporal multiscale relationships among variables in the Tennessee Eastman Process (TEP). Subsequently, the performance of MsrtNet is evaluated in predicting product yield for a 2.80 × 10⁶ t/a FCC unit, taking diesel and gasoline yield as examples. In conclusion, MsrtNet can decouple and effectively extract spatiotemporal multiscale information from chemical process data and achieve a approximately reduction of 30% in prediction error compared to other time-series models. Furthermore, its robustness and transferability underscore its promising potential for broader applications.