Electrical capacitance tomography is an industrial process tomography technology, mainly used to measure and plot two-phase flow and multi-phase flow. The technology is based on the properties of various dielectric constants between the phases of the measured substance. This work focuses on the improvement of particle swarm optimization and intelligent algorithms, especially for the parameter control of particle swarm optimization and intelligent algorithms. The efficiency of group optimization algorithms and intelligent algorithms that solve optimization problems. This paper mainly aims at the image reconstruction process in electrical capacitance tomography system, and proposes an image reconstruction algorithm based on intelligent algorithm and particle swarm algorithm. Combined with the experimental environment of the self-made electrical capacitance tomography system, the actual imaging effect of the algorithm was compared and analyzed with the traditional imaging algorithm, and the verification of the algorithm improvement effect was theoretically completed. According to the finite element analysis method, the internal area of the sensor is subdivided on the entire network, and three flow modes are modeled, which provides conditions for the construction of the following experimental environment. Based on the theory, the principle of the classic Landweber imaging algorithm is discussed in detail. Using the experimental environment with built-in electrical capacitance tomography system, experiments were conducted to visualize the gas–liquid two-phase flow. The traditional Landweber algorithm and imaging algorithm proposed in this paper are used to reconstruct the image using the obtained volume data. Through comparative analysis of the resulting images, the results show that the imaging algorithm proposed in this paper improves the accuracy of flow pattern recognition and image accuracy, which proves the improvement. The feasibility of algorithm and particle swarm algorithm.

电容层析成像是一种工业过程层析成像技术，主要用于测量和绘制两相流和多相流。该技术基于被测物质各相之间的各种介电常数的特性。本工作侧重于粒子群优化和智能算法的改进，特别是粒子群优化和智能算法的参数控制。组优化算法和解决优化问题的智能算法的效率。本文主要针对电容层析成像系统中的图像重建过程，提出一种基于智能算法和粒子群算法的图像重建算法。结合自制电容层析成像系统的实验环境，将该算法的实际成像效果与传统成像算法进行对比分析，从理论上完成了算法改进效果的验证。根据有限元分析方法，在整个网络上对传感器内部区域进行细分，对三种流动模式进行建模，为后续实验环境的搭建提供了条件。在此基础上，详细讨论了经典Landweber成像算法的原理。利用内置电容层析成像系统的实验环境，进行了气液两相流可视化实验。采用传统的Landweber算法和本文提出的成像算法，利用获得的体数据重建图像。通过对所得图像的对比分析，结果表明本文提出的成像算法提高了流型识别的准确性和图像的准确性，证明了改进。算法和粒子群算法的可行性。