Gas Volume Fraction Measurement of Oil–Gas–Water Three-Phase Flows in Vertical Pipe by Combining Ultrasonic Sensor and Deep Attention Network

Ren, Weikai; Jin, Ningde; OuYang, Lei; Zhai, Lusheng; Ren, Yi

doi:10.1109/tim.2020.3031186

Cited by 11 publications

(6 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Training a convolutional neural network to the target data without an LSTM layer to obtain pretrained convolutional filter weights would be a sub-optimal task due to the LSTM layer's ability to learn the important process feature trajectories. Therefore, the input waveforms would not be able to fit to the target data optimally without an LSTM layer and informative waveform features would not be learned (Bowler et al, 2020 and2021). Training convolutional and LSTM layers simultaneously would also be a difficult task especially with long time sequences and limited training data used in the present case studies.…”

Section: Convolutional Feature Extractionmentioning

confidence: 99%

“…For ultrasonic techniques, the speed of sound is commonly used as a feature as it is dependent on the density and compressibility of the material it passes through and is calculated by measuring the sound wave time of flight and distance travelled (Utomo et al, 2001, Utomo et al, 2002, Supardan et al, 2003, Sun et al, 2005. The changing amplitude between consecutively acquired waveforms can be used as a feature to identify process states and has been applied to determine flow regimes (Ren et al, 2021;Abbagoni and Yeung, 2016). Other process information can also be used to aid the prediction accuracy of the ML model, such as the temperature, material composition and concentration (Sun et al, 2005), or mass flow rate (Wallhäußer et al, 2014).…”

Section: Introductionmentioning

confidence: 99%

“…CNNs have been used previously with US signals (Virupakshappa et al, 2018, Meng et al, 2017, Munir et al, 2019, Munir et al, 2020, Bowler et al, 2020. However, previous work has also shown that Long Short-Term Memory (LSTMs) neural network layers are required to accurately monitor time-evolving processes (Bowler et al, 2020 and2021). LSTMs are able to retain process information from previous timesteps and are a type of recurrent neural network which uses gate units to reduce the likelihood of vanishing or exploding gradients.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Convolutional feature extraction for process monitoring using ultrasonic sensors

Bowler¹,

Pound²,

Watson³

2021

Computers & Chemical Engineering

View full text Add to dashboard Cite

Section: Convolutional Feature Extractionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Convolutional feature extraction for process monitoring using ultrasonic sensors

Bowler¹,

Pound²,

Watson³

2021

Computers & Chemical Engineering

View full text Add to dashboard Cite

“…So many methods and strategies to find flow velocity in the pipe, especially in one phase of the custody transfer. In the industrial process, due to its high accuracy and absence of moving parts, the transit time method is the most frequently used or widely used method [1][2][3][4][5][6][7]. However, the transit time usually takes advantage of sending ultrasonic waves from upstream to downstream and vice versa.…”

Section: Introductionmentioning

confidence: 99%

The development of simultaneous transducer ultrasonic with dual-transducer to measure flow velocity in the pipe

Wiranata,

Kurniadi

2023

Eureka: PE

View full text Add to dashboard Cite

There are still several obstacles to calculating fluid flow rate measurement on custody transfer, such as the distribution of one-phase fluid flow rate, less sTable pressure, and fluid flow rate changes, which are still the main problems in the process of a measuring system. To calculate the flow rate of one phase fluid, there is a method often used, namely transit time. In practice, the transit time works to send the ultrasonic pulse from upstream to downstream, which had to have a delay because the transducer must switch the function of a transducer to become transmitter or receiver. So, this paper proposed a new strategy of measurement multipath ultrasonic flowmeter (UFLW) with a simultaneous transit time method using a dual transmitter and receiver. The simultaneous method is a measuring technique that utilizes a pair of ultrasonic transducers as both trigger and receiver. The first transducer serves as the transmitter, while the second functions as the receiver, capturing the signal at the same time without changing their positions or roles. In order to implement the configuration setup, let’s try to use 3 paths configuration with 6 pairs of transducers, 3 on the upstream and 3 on the downstream. To estimate the flow velocity, let’s use long short-term memory (LSTM), which is one of the recurrent neural networks (RNN) architectures in the deep learning algorithm, and to evaluate the performance, let’s use the Bland-Altman plot and root mean squared error (RMSE) and validation loss of the LSTM model. The result shows RMSE 0.289 from the actual flow velocity, which means an LTSM with simultaneous multipath ultrasonic can reduce the error between prediction and actual measurement

show abstract

“…Hage et al 17 discuss a similar method based on a modified Wood’s equation for evaluating the velocity and phase change of an acoustic wave propagating along the annulus of the borehole. Adaptations of this principle are also evaluated numerically 18 and, experimentally 19 – 25 .…”

Section: Introductionmentioning

confidence: 99%

Effect of bubble size on ultrasound backscatter from bubble clouds in the context of gas kick detection in boreholes

Indimath

Fiorentini

Bøklepp

et al. 2023

Sci Rep

View full text Add to dashboard Cite

Early detection of gas influx in boreholes while drilling is of significant interest to drilling operators. Several studies suggest a good correlation between ultrasound backscatter/attenuation and gas volume fraction (GVF) in drilling muds, and thereby propose methods for quantification of GVF in boreholes. However, the aforementioned studies neglect the influence of bubble size, which can vary significantly over time. This paper proposes a model to combine existing theories for ultrasound backscatter from bubbles depending on their size, viz. Rayleigh scattering for smaller bubbles, and specular reflection for larger bubbles. The proposed model is demonstrated using simulations and experiments, where the ultrasound backscatter is evaluated from bubble clouds of varying bubbles sizes. It is shown that the size and number of bubbles strongly influence ultrasound backscatter intensity, and it is correlated to GVF only when the bubble size distribution is known. The information on bubble size is difficult to obtain in field conditions causing this correlation to break down. Consequently, it is difficult to reliably apply methods based on ultrasound backscatter, and by extension its attenuation, for the quantification of GVF during influx events in a borehole. These methods can however be applied as highly sensitive detectors of gas bubbles for GVF $$\ge$$ ≥ 1 vol$$\%$$ % .

show abstract

Gas Volume Fraction Measurement of Oil–Gas–Water Three-Phase Flows in Vertical Pipe by Combining Ultrasonic Sensor and Deep Attention Network

Cited by 11 publications

References 39 publications

Convolutional feature extraction for process monitoring using ultrasonic sensors

Convolutional feature extraction for process monitoring using ultrasonic sensors

The development of simultaneous transducer ultrasonic with dual-transducer to measure flow velocity in the pipe

Effect of bubble size on ultrasound backscatter from bubble clouds in the context of gas kick detection in boreholes

Contact Info

Product

Resources

About