https://bxscience.enschool.org/ourpages/auto/2014/9/5/42400025/Example%209.pdf

Applications of artificial neural networks for thermal analysis of heat exchangers – A review

Liquid and gas flow sensors are important components of the micro total analysis systems (μTAS) for modern analytical sciences. In this paper, we proposed a self-powered triboelectric microfluidic sensor (TMS) by utilizing the signals produced from the droplet/bubble via the capillary and the triboelectrification effects on the liquid/solid interface for real-time liquid and gas flow detection. By alternating capillary with different diameters, the sensor’s detecting range and sensitivity can be adjusted. Both the relationship between the droplet/bubble and capillary size, and the output signal of the sensor are systematically studied. By demonstrating the monitoring of the transfusion process for a patient and the gas flow produced from an injector, it shows that TMS has a great potential in building a self-powered micro total analysis system.

Self-Powered Triboelectric Micro Liquid/Gas Flow Sensor for Microfluidics.

https://research.tees.ac.uk/ws/files/9903105/Application_of_soft_computing_techniques_to_multiphase_flow_measurement.pdf

Application of soft computing techniques to multiphase flow measurement: A review

Well production forecasting based on ARIMA-LSTM model considering manual operations

Flow regime identification of swirling gas-liquid flow with image processing technique and neural networks

Measurement of gas and liquid flow rates in two-phase pipe flows by the application of machine learning techniques to differential pressure signals

Identification of flow regime in vertical upward air–water pipe flow using differential pressure signals and elastic maps

A novel gas and liquid flow rate measurement method is proposed for use in vertical upward and downward gas–liquid pipe flows. This method is based on the analysis of the time history of area-averaged void fraction that is measured using a conductivity wire-mesh sensor (WMS). WMS measurements were collected in vertical upward and downward air–water flows in a pipe with an internal diameter of 32.5 mm at nearly atmospheric pressure. The relative frequencies and the power spectral density of area-averaged void fraction were calculated and used as representative properties. Independent features, extracted from these properties using Principal Component Analysis and Independent Component Analysis, were used as inputs to artificial neural networks, which were trained to give the gas and liquid flow rates as outputs. The present method was shown to be accurate for all four encountered flow regimes and for a wide range of flow conditions. Besides providing accurate predictions for steady flows, the method was also tested successfully in three flows with transient liquid flow rates. The method was augmented by the use of the cross-correlation function of area-averaged void fraction determined from the output of a dual WMS unit as an additional representative property, which was found to improve the accuracy of flow rate prediction.

http://iopscience.iop.org/article/10.1088/0957-0233/26/1/015306/pdf

H. Shaban

Papers

Measurement of gas and liquid flow rates in two-phase pipe flows by the application of machine learning techniques to differential pressure signals

Identification of flow regime in vertical upward air–water pipe flow using differential pressure signals and elastic maps

The wire-mesh sensor as a two-phase flow meter

On the accuracy of gas flow rate measurements in gas–liquid pipe flows by cross-correlating dual wire-mesh sensor signals

Performance evaluation of conductivity wire-mesh sensors in vertical channels