In the later stage of natural gas well production, the bottom hole pressure decreases, the gas flow velocity decreases, the liquid carrying rate decreases, the wellbore liquid accumulation increases, the liquid accumulation hinders the natural gas migration, and the production decreases. The liquid discharge efficiency of traditional technologies such as gas lift, bubble discharge, natural gas circulation and high pressure pump is poor[1–3]. In order to solve the problem of serious fluid accumulation in wellbore, a new technology of nozzle atomization for drainage and gas recovery is adopted[4–6]. The technology uses the energy of natural gas to atomize the liquid accumulated in the bottom hole through the nozzle, and the droplets are carried out of the wellbore along with the natural gas. This method reduces energy consumption, improves energy utilization, reduces the operation and maintenance cost of gas wells, and can effectively solve the problem of wellbore fluid accumulation.
Many scholars have carried out related research on nozzle atomization technology, carried out corresponding atomization experiments[7–8], and achieved ideal results. Han Han et al.[9] designed a kind of internal mixing atomization and dedusting nozzle. The flow field in the nozzle was simulated by FLUENT. The results show that with the increase of water supply pressure, the flow velocity in the nozzle increases, the air velocity decreases, and the gas-liquid relative velocity decreases. They carried out spray dust reduction experiments through atomizing nozzles. The experimental results show that when the water supply pressure increases, the range, droplet volume fraction and droplet size all increase, and the dust reduction efficiency of total dust and respirable dust increases at first and then decreases. KOMAG Mining Technology Research Institute[10] has developed a kind of water spray nozzle, which can be effectively used to spray dust reduction at the transfer point of shearer, roadheader and conveyor. Feng[11] carried out atomization drainage tests on low-pressure and low-production gas wells using nozzles at the production site of West Sichuan gas field. The atomization effect of nozzles is good, the liquid carrying rate of wellbore is increased by 23.4%, and the liquid production is increased by 200m3/d. Ni et al.[12] put the supersonic nozzle atomization device at the depth of 1000-2000m to carry out field experiments, which shows that the outlet velocity is greatly higher than that of ordinary nozzles, and the liquid-carrying efficiency is increased by 41.3%.
In recent years, the experimental means have also been innovated due to the progress of science and technology, through the in-depth study of spray experiments, in order to obtain better atomization effect, so as to achieve better economic value. Benjanin et al. [13] studied the internal and external flow field of swirl nozzle. According to the experimental results, the model formulas of mass flow coefficient and droplet velocity were obtained. Seoksu et al. [14] studied the internal structure and static pressure of swirl nozzle, and the study showed that there was backflow vortex in the swirl atomization process, and the pressure drop inside the nozzle was larger under high injection pressure. Hyung yu Lee et al.[15] carried out experiments on pass nozzles, and optimized the pass nozzles by CFD simulation, optimizing four design variables: nozzle inlet length, outlet length, inlet diameter and radial position. The results show that the optimized nozzle reduces the total pressure loss and increases the mass flow, and the optimized pre-swirl system reduces the aerodynamic loss, increases the mass flow rate under a certain pressure, and satisfies the pressure margin of blade cooling. The supersonic atomization nozzle model was established by Los Alamos[16] laboratory in the United States, and the supersonic atomization combustion process of internal combustion engine was studied. The analysis of the complex motion of evaporation, fragmentation and turbulent diffusion of two-phase chemical fluid can effectively improve the efficiency of atomization combustion.
The atomization effect of the nozzle can be evaluated by studying the droplet size SMD and its distribution, which are the key parameters of the atomization performance. Haoqi Lilan et al.[17] studied the droplet size distribution in the flow field of external air atomizer through experiments. They divide the atomization field into several observation areas, and through the measurement of several local observation areas, the relationship of droplet size distribution in the whole atomization field is obtained, which provides a certain reference for the study of nozzle atomization field. it provides a basis for intuitively understanding the droplet size distribution of the nozzle atomization field. Yu Hai-long et al.[18] developed a new type of gasified coal water slurry nozzle and studied its atomization performance. They discussed in detail the effects of nozzle working load and gas flow rate on atomization particle distribution, Sauter mean diameter SMD and nozzle atomization angle. Hyun Kyu Suh et al.[19] studied the effect of cavitation flow on the atomization characteristics of diesel fuel in different size nozzles through the flow visualization system, and used the particle measurement system to determine the atomization characteristics such as SMD and droplet average velocity. The results show that the cavitation in the nozzle enhances the fuel atomization performance, and the longer the nozzle orifice length is, the more fuel atomization is. Y.Xia et al.[20] used laser particle size system to measure the droplet diameter of different spray fields. It was found that the droplet size in the spray center was the smallest and the SMD at the spray edge increased.
The influence of atomization parameters on atomization effect was studied, and the experimental studies on the influence factors of different operating parameters on droplet diameter were carried out, and the corresponding atomization model was established to improve the atomization efficiency. Nonnenmacher et al.[21] studied the hollow cone pressure swirl nozzle based on the theory of internal and external flow field of the nozzle, and established the simulation program model of flow coefficient and droplet diameter. According to the simulation program, the Sauter average diameter SMD of the nozzle can be predicted. Wang et al.[22] established the pre-film atomization model by using neural network algorithm, and concluded that with the increase of oil pressure difference, the pre-film device appeared anti-fog effect, and the atomization effect was poor. According to the similarity of droplet breakup, Liu et al.[23] constructed a finite random fragmentation model (FSBM) for the blast atomization process before film formation. The droplet size distribution is simulated by using this model, and the simulation results are consistent with the experimental results of the pre-film blast atomizer. This model can accurately determine the nonlinear relationship between the average droplet diameter SMD and the size distribution of the blast spray.
According to the previous research and analysis of nozzle atomization experiments, there are many studies on the factors affecting the droplet diameter of atomization parameters, and the current nozzle atomization research is limited to the simple relationship and law among atomization parameters, but there is no complex atomization model based on SMD of nozzle atomization parameters, no accurate prediction of SMD, and the qualitative analysis of the influencing factors of SMD is time-consuming and laborious. Therefore, in this paper, a new algorithm-genetic algorithm optimized BP neural network is used to establish an atomization model for the new nozzle atomization parameter SMD, and compared with the traditional BP neural learning network to establish a reliable and accurate atomization model. It can quickly predict the atomization target parameter SMD, improve the prediction efficiency and accuracy, and has certain significance for setting the working condition parameters of downhole nozzles and improving the drainage efficiency.