Forecasting daily natural gas load accurately is difficult because it is affected by various factors. A large number of redundant factors existing in the original dataset will increase computational complexity and decrease the accuracy of forecasting models. This study aims to provide accurate forecasting of natural gas load using a deep learning (DL)-based hybrid model, which combines principal component correlation analysis (PCCA) and (LSTM) network. PCCA is an improved principal component analysis (PCA) and is first proposed here in this paper. Considering the correlation between components in the eigenspace, PCCA can not only extract the components that affect natural gas load but also remove the redundant components. LSTM is a famous DL network, and it was used to predict daily natural gas load in our work. The proposed model was validated by using recent natural gas load data from Xi’an (China) and Athens (Greece). Additionally, 14 weather factors were introduced into the input dataset of the forecasting model. The results showed that PCCA–LSTM demonstrated better performance compared with LSTM, PCA–LSTM, back propagation neural network (BPNN), and support vector regression (SVR). The lowest mean absolute percentage errors of PCCA–LSTM were 3.22% and 7.29% for Xi’an and Athens, respectively. On these bases, the proposed model can be regarded as an accurate and robust model for daily natural gas load forecasting.

This study developed a theoretical model for predicting the downstream temperatures of high pressure high temperature condensate gas flowing through chokes. The model is composed of three parts: the iso-enthalpy choke model derived from continuity equation and energy conservation equation; the liquid-vapor equilibrium model based on the SRK equation of state (EoS); and the enthalpy model based on the Lee-Kesler EoS. Pseudocritical properties of mixtures, which are obtained by mixing rules, are very important in the enthalpy model, so the Lee-Kesler, Plocker-Knapp, Wong-Sandler and Prausnitz-Gunn mixing rules were all researched, and the combination mixing rules with satisfactory accuracy for high pressure high temperature condensate gases were proposed. The temperature prediction model is valid for both the critical and subcritical flows through different kinds of choke valves. The applications show the model is reliable for predicting the downstream temperatures of condensate gases with upstream pressures up to 85.54 MPa and temperatures up to 93.23 °C. The average absolute errors between the measured and calculated temperatures are expected for less than 2 °C by using the model.

The leakage process simulation of a Natural Gas Liquid (NGL) storage tank requires the simultaneous solution of the NGL’s pressure, temperature and phase state in the tank and across the leak hole. The methods available in the literature rarely consider the liquid/vapor phase transition of the NGL during such a process. This paper provides a comprehensive pressure-temperature-phase state method to solve this problem. With this method, the phase state of the NGL is predicted by a thermodynamic model based on the volume translated Peng-Robinson equation of state (VTPR EOS). The tank’s pressure and temperature are simulated according to the pressure-volume-temperature and isenthalpic expansion principles of the NGL. The pressure, temperature, leakage mass flow rate across the leak hole are calculated from an improved Homogeneous Non-Equilibrium Diener-Schmidt (HNE-DS) model and the isentropic expansion principle. In particular, the improved HNE-DS model removes the ideal gas assumption used in the original HNE-DS model by using a new compressibility factor developed from the VTPR EOS to replace the original one derived from the Clausius-Clayperon equation. Finally, a robust procedure of simultaneously solving the tank model and the leak hole model is proposed and the method is validated by experimental data. A variety of leakage cases demonstrates that this method is effective in simulating the dynamic leakage process of NGL tanks under critical and subcritical releasing conditions associated with vapor/liquid phase change.

When tubing is in a high-temperature and high-pressure environment, it will be affected by the impact of non-constant fluid and other dynamic loads, which will easily cause the tubing to vibrate or even resonate, affecting the integrity of the wellbore and safe production. In the structural modal analysis of the tubing, the coupling effect of the fluid and the tubing needs to be considered at the same time. In this paper, a single tubing is taken as an example to simulate and analyze the modal changes of the tubing under dry mode and wet mode respectively, and the effects of fluid solid coupling effect, inlet pressure, and ambient temperature on the modal of the tubing are discussed. After considering the fluid–structure interaction effect, the natural frequency of tubing decreases, but the displacement is slightly larger. The greater the pressure in the tubing, the greater the equivalent stress on the tubing body, so the natural frequency is lower. Furthermore, after considering the fluid–solid coupling effect, the pressure in the tubing is the true pulsating pressure of the fluid. The prestress applied to the tubing wall changes with time, and the pressures at different parts are different. At this time, the tubing is changed at different frequencies. Vibration is prone to occur, that is, the natural frequency is smaller than the dry mode. The higher the temperature, the lower the rigidity of the tubing and the faster the strength attenuation, so the natural frequency is lower, and tubing is more prone to vibration. Both the stress intensity and the elastic strain increase with the increase of temperature, so the displacement of the tubing also increases.

In order to solve the problem of demulsification difficulties in Liaohe Oilfield, 24 kinds of demulsifiers were screened by using the interface generation energy (IFE) module in the molecular dynamics simulation software Materials Studio to determine the ability of demulsifier molecules to reduce the total energy of the oil–water interface after entering the oil–water interface. Neural network analysis (NNA) and genetic function approximation (GFA) were used as technical means to predict the demulsification effect of the Liaohe crude oil demulsifier. The simulation results show that the SDJ9927 demulsifier with ethylene oxide (EO) and propylene oxide (PO) values of 21 (EO) and 44 (PO) reduced the total energy and interfacial tension of the oil–water interface to the greatest extent, and the interfacial formation energy reached −640.48 Kcal/mol. NNA predicted that the water removal amount of the SDJ9927 demulsifier was 7.21 mL, with an overall error of less than 1.83. GFA predicted that the water removal amount of the SDJ9927 demulsifier was 7.41mL, with an overall error of less than 0.9. The predicted results are consistent with the experimental screening results. SDJ9927 had the highest water removal rate and the best demulsification effect. NNA and GFA had high correlation coefficients, and their R²s were 0.802 and 0.861, respectively. The higher R² was, the more accurate the prediction accuracy was. Finally, the demulsification mechanism of the interfacial film breaking due to the collision of fluorinated polyether demulsifiers was studied. It was found that the carbon–fluorine chain had high surface activity and high stability, which could protect the carbon–carbon bond in the demulsifier molecules to ensure that there was no re-emulsion due to the stirring external force.