Flow Measurement and Instrumentation最新文献

{"title":"Design and optimization of a universal, cost-effective gas sensing and photoconductivity measuring system","authors":"M.A. Sebak ,&nbsp;A.K. Aladim ,&nbsp;Fatma Gami ,&nbsp;Abdalrahman M. Rayan ,&nbsp;Mahrous R. Ahmed ,&nbsp;Mohamed Asran Hassan","doi":"10.1016/j.flowmeasinst.2026.103191","DOIUrl":"10.1016/j.flowmeasinst.2026.103191","url":null,"abstract":"<div><div>This study details the design, implementation, and validation of a modular, economical instrumentation system for the concurrent assessment of gas-sensing and photoconductivity features in semiconductor materials. The developed platform integrates off-the-shelf components, such as an Arduino-based microcontroller, a high-precision data-logging multimeter, and a time-regulated motorized ball valve, resulting in a cost reduction exceeding 90 % compared to commercial alternatives, while maintaining scientific accuracy. The system includes a bespoke Pyrex glass chamber and a PID-regulated heating element, guaranteeing consistent and replicable environmental conditions for both bulk pellets and thin-film samples. To assess the system's performance, sensing measurements were performed on a CuO pellet subjected to CO₂ gas. Experimental findings indicated a maximum sensitivity of 33.68 % at high temperatures, with response and recovery durations recorded at 692 s and 476 s, respectively. Arrhenius analysis produced activation energies of 0.561 eV in the high-temperature domain and 0.161 eV in the low-temperature domain, demonstrating remarkable quantitative concordance with existing literature. Moreover, the integrated optical module, employing a modulated video projector as a multi-wavelength light source, effectively exhibited the system's ability to conduct photo-induced conductivity measurements. This dual-functional architecture offers a dependable, high-throughput option for researchers investigating intricate light-assisted sensing mechanisms, substantially reducing the financial barrier to advanced materials characterization.</div></div>","PeriodicalId":50440,"journal":{"name":"Flow Measurement and Instrumentation","volume":"108 ","pages":"Article 103191"},"PeriodicalIF":2.7,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145924935","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Modeling of ultrasonic blood flow sensor and its extended application discussion","authors":"Dandan Zheng,&nbsp;Manman Wu,&nbsp;Ying Xu,&nbsp;Zimeng Zheng,&nbsp;Chenglin Yv,&nbsp;Xueyong Chen,&nbsp;Haojun Fan","doi":"10.1016/j.flowmeasinst.2026.103200","DOIUrl":"10.1016/j.flowmeasinst.2026.103200","url":null,"abstract":"<div><div>—Extracorporeal membrane oxygenation (ECMO) serves as a critical life-support technology for patients with cardiopulmonary failure. Ultrasonic blood flow sensors, as core ECMO components, enable accurate real-time flow monitoring. This study addresses the variation in measurement characteristics of clamp-on ultrasonic flowmeters in different media by proposing a flow-acoustic coupled numerical simulation method and establishing a data-driven correction coefficient prediction model. Experimental validation was subsequently performed using a developed high-precision ultrasonic blood flow sensor prototype. First, a coupled flow-acoustic simulation model was built through collaborative modeling of CFD and acoustic modules. Experimental validation in pure water and blood-mimicking fluid showed absolute errors between simulation and experimental results within ±3 % (flow rate ≥1 L/min), confirming the accuracy of the simulation method. Second, based on simulation data from multiple media at different temperatures, an artificial neural network (ANN)-based prediction model for instrument correction coefficients was developed. The mean absolute percentage errors (MAPE) for the training and test sets reached 0.543 % and 1.028 %, respectively, outperforming regression trees and support vector machines. The proposed method effectively eliminates medium-dependent measurement errors, enhances cross-population applicability, and promotes the generalization of ultrasonic sensors to industrial and pharmaceutical fields, thereby improving cross-medium adaptation capability of ultrasonic flow sensors. Finally, the prediction model was tested using the developed prototype ultrasonic blood flow sensor. Within the 1–10 L/min flow range, the measurement error was reduced from a maximum of 16 % to within ±10 % after correction, with repeatability better than 1 %, achieving performance comparable to commercial Transonic sensors.</div></div>","PeriodicalId":50440,"journal":{"name":"Flow Measurement and Instrumentation","volume":"108 ","pages":"Article 103200"},"PeriodicalIF":2.7,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145976527","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An intelligent data-driven flow regime recognition method for horizontal air-water two-phase flow","authors":"Meng Wang ,&nbsp;Ruowei Li ,&nbsp;Yanyan Shi ,&nbsp;Zhen Yang ,&nbsp;Manuchehr Soleimani","doi":"10.1016/j.flowmeasinst.2026.103195","DOIUrl":"10.1016/j.flowmeasinst.2026.103195","url":null,"abstract":"<div><div>Flow regime recognition is very important in the two-phase flow measurement. However, considering that two-phase flow is much more complicated than single-phase flow, flow regime cannot be accurately identified by mechanism model. In this work, a novel intelligent data-driven method based on image encoding and transformer is proposed to recognize typical flow regimes encountered in the horizontal air-water flow. Dynamic experiment is carried out using a ring-shaped conductance sensor to collect voltage signals of bubble flow, bubble-slug flow, slug flow, slug-stratified flow and stratified flow. To highlight the characteristic differences between different flow regimes, the measured signals are encoded into two-dimensional images. To classify the encoded images of the five flow regimes, transformer models are then established. With the encoded images as the input of the model, flow regime identification is implemented by training of the model. The results demonstrate that the characteristics of different flow regimes can be better reflected in the encoded image with Gramian angular field. Meanwhile, the recognition accuracy of Swin Transformer is advantageous to that of Vision Transformer in the classification of the encoded images of the five flow regimes. Comparing with other identification methods, the method which combines Gramian angular field with Swin Transformer shows the best performance in the recognition of the flow regimes. The total accuracy reaches as high as 99.1 % This study offers an alternative for accurate flow regime recognition in two-phase flow measurement.</div></div>","PeriodicalId":50440,"journal":{"name":"Flow Measurement and Instrumentation","volume":"108 ","pages":"Article 103195"},"PeriodicalIF":2.7,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145976530","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Gas plug characteristic parameters measurement based on high-speed imaging technology","authors":"Jiaqi Wen ,&nbsp;Haodi Jia ,&nbsp;Jiale Li ,&nbsp;Lei Gao ,&nbsp;Lide Fang ,&nbsp;Ning Zhao","doi":"10.1016/j.flowmeasinst.2026.103196","DOIUrl":"10.1016/j.flowmeasinst.2026.103196","url":null,"abstract":"&lt;div&gt;&lt;div&gt;Gas-liquid two-phase plug flow is widely prevalent in industrial scenarios such as aircraft cooling, chemical production, and falling film evaporation. Conducting precise measurement research on the characteristic parameters of Taylor bubbles within plug flow is a key research focus in the field of two-phase flow. In gas-liquid two-phase plug flow, parameters such as the slug length and the void fraction are crucial for studying the dynamic characteristics of the two phases. The study adopts a visual sensor measurement system utilizing refraction-corrected high-speed imaging technology to achieve dynamic measurement of the characteristic parameters of gas plugs. Based on computer vision technology and integrated with deep learning methodologies, an optimized edge detection operator is employed to accurately locate and extract the contour of the gas plug. Data acquisition and piecewise polynomial fitting modeling are conducted for the characteristic parameters such as the length of the gas plug head, the height of the gas plug head, and the average void fraction of the gas plug. The results demonstrate that for the prediction model of the length-to-height ratio of the gas plug head, at the superficial velocities condition of 0.028&lt; &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;msub&gt;&lt;mi&gt;u&lt;/mi&gt;&lt;mrow&gt;&lt;mi&gt;s&lt;/mi&gt;&lt;mi&gt;g&lt;/mi&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; &lt;0.283, the mean absolute percentage error (MAPE) of the prediction results is 12.53 %, with over 80 % of prediction results falling within a ±20 % relative error range. At the superficial velocities condition of 0.354&lt; &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;msub&gt;&lt;mi&gt;u&lt;/mi&gt;&lt;mrow&gt;&lt;mi&gt;s&lt;/mi&gt;&lt;mi&gt;g&lt;/mi&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; &lt;0.495, the MAPE of the prediction results is 14.24 %, with over 80 % of prediction results within a ±20 % relative error range. For the prediction model of the average void fraction of the gas plug, at the superficial velocities condition of 0.028&lt; &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;msub&gt;&lt;mi&gt;u&lt;/mi&gt;&lt;mrow&gt;&lt;mi&gt;s&lt;/mi&gt;&lt;mi&gt;g&lt;/mi&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; &lt;0.354, the MAPE of the prediction results is 3.96 %, with 99 % of the prediction results within the range of ±15 % relative error. At the superficial velocities condition of 0.424&lt; &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;msub&gt;&lt;mi&gt;u&lt;/mi&gt;&lt;mrow&gt;&lt;mi&gt;s&lt;/mi&gt;&lt;mi&gt;g&lt;/mi&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; &lt;0.495, the MAPE of the prediction results is 8.26 %, with 83 % of the prediction results within the range of ±15 % relative error. On this basis, the machine learning method is adopted to improve the prediction accuracy the model. A Support &lt;strong&gt;Vector Regression (SVR)&lt;/strong&gt; based prediction model for the length-to-height ratio of the gas plug head and the average void fraction of the gas plug is established. The model prediction results show that, the MAPE of the height-to-length ratio of gas plug head prediction model is 4.74 %, while the MAPE of the gas plug average void fraction prediction model is 0.84 %, further improving the generalization ability of the mo","PeriodicalId":50440,"journal":{"name":"Flow Measurement and Instrumentation","volume":"108 ","pages":"Article 103196"},"PeriodicalIF":2.7,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145976444","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Research on the measurement method of pulverized coal concentration based on microwave loss","authors":"Jiangbo Qian ,&nbsp;Shuo Wang ,&nbsp;Ao Liu ,&nbsp;Zhijian Zhang ,&nbsp;Guanghai Lu","doi":"10.1016/j.flowmeasinst.2026.103199","DOIUrl":"10.1016/j.flowmeasinst.2026.103199","url":null,"abstract":"<div><div>Real-time, accurate, and non-contact measurement of pulverized coal concentration (PCC) in primary air ducts is critical for the safe and efficient operation of coal-fired boilers. Existing microwave attenuation-based methods suffer from narrow bandwidth, cumbersome installation, or lack of dynamic experimental validation. To address these limitations, this study proposes a PCC measurement system based on microwave attenuation theory. First, a mathematical model for the equivalent complex permittivity of the air-pulverized coal two-phase mixture was established by applying integral modification to the Maxwell-Garnett heterogeneous dielectric equation. After incorporating the dielectric parameters of lignite into the model, the dependence of the complex permittivity of the air-lignite mixture on electric field frequency and coal concentration was determined. Based on microwave attenuation theory, a relationship between microwave attenuation and dielectric permittivity was derived, thus establishing the theoretical foundation for concentration measurement. Second, a dual-ridged horn antenna operating in the 2–10 GHz range was designed and optimized, featuring flexible installation and stable radiation performance. Finally, dynamic experiments were conducted under reflective and transmissive modes, simulating actual power plant scenarios with six PCC levels (0–1.0 kg/kg). The maximum relative deviations between the measured values and the true concentrations obtained by the reflection and transmission arrangements were 8.59 % and 7.31 %, respectively. Results confirm that the proposed system achieves wide bandwidth, flexible deployment, and reliable accuracy, verifying its feasibility for gas-solid two-phase flow concentration measurement.</div></div>","PeriodicalId":50440,"journal":{"name":"Flow Measurement and Instrumentation","volume":"108 ","pages":"Article 103199"},"PeriodicalIF":2.7,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145976602","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-interface fusion speed pulsation model and behavior analysis for high-torque hydraulic motors","authors":"Yu Fang ,&nbsp;Haifeng Yang ,&nbsp;Gongcheng Zhang ,&nbsp;Jianxiong Wu ,&nbsp;Yiman Duan ,&nbsp;Xuguang Li ,&nbsp;Junhui Zhang ,&nbsp;Bing Xu ,&nbsp;Chao Zhang","doi":"10.1016/j.flowmeasinst.2025.103148","DOIUrl":"10.1016/j.flowmeasinst.2025.103148","url":null,"abstract":"<div><div>Speed pulsation of the hydraulic motor not only reflects its inherent characteristics but also significantly affects its output stability and operation safety. However, there is still a significant challenge to reveal the speed pulsation mechanism owing to the complex friction and leakage characteristics of multi-interfaces in the hydraulic motor, especially under low speeds. To address this issue, this study proposes an innovative multi-interface fusion speed pulsation model that comprehensively incorporates the friction and leakage losses for high-torque hydraulic motors. Firstly, a coupled rigid-fluid dynamics framework is established to determine the speed pulsation behavior. Within this framework, multiple methods are employed to describe the friction and leakage losses in different interfaces according to their lubrication characteristics. Furthermore, a dedicated speed pulsation test rig is developed to experimentally validate the proposed model. Finally, the influence of multi-interface friction and leakage on the speed pulsation behavior is quantitatively revealed, providing new insights into the intrinsic pulsation mechanism and offering valuable guidance for the design optimization of high-torque hydraulic motors.</div></div>","PeriodicalId":50440,"journal":{"name":"Flow Measurement and Instrumentation","volume":"108 ","pages":"Article 103148"},"PeriodicalIF":2.7,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145684977","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Gas-liquid flow regime information in horizontal and vertical pipes using machine learning approach on time varying pressure drop signals","authors":"Madhu V ,&nbsp;Suresh M ,&nbsp;Umapathy M ,&nbsp;Ezudheen P ,&nbsp;Vishakh Prakash","doi":"10.1016/j.flowmeasinst.2025.103176","DOIUrl":"10.1016/j.flowmeasinst.2025.103176","url":null,"abstract":"<div><div>Flow regime information helps provide insight into complex interaction between phases in two-phase flow. The phase distribution and flow pattern significantly impact behaviour of pressure drop at any section of flow conduit. In the study, experimental investigations were undertaken at a 100 mm NB air-water test loop for fully developed horizontal flow for the regimes slug, plug and stratified flow and further the regimes slug, churn flow regimes in vertical flow. Measurements for pressure loss were obtained using high response differential pressure sensors coupled with high speed sampling measurements at the Data Acquisition System.</div><div>The acquired time-series differential pressure signals were assessed for patterns using machine learning approaches. Techniques like Random Forest, non-linear regression models did not yield resolvable results for a range of tests. Three machine learning approaches namely, Gated Recurrent Unit(GRU), Long Short Term Memory (LSTM) and Shapelet Transform Classifier (STC) were attempted that reveal reasonably good prediction for flow regimes.</div><div>As per analysis, the STC model yielded accuracy in the range 71–73 % for horizontal flow regimes viz., stratified flow, slug flow and plug flow, based on pressure loss in horizontal line and venturimeter. For the vertical flow regimes slug flow and churn flow, the LSTM and GRU indicate similar performance (72–73 % prediction accuracy). LSTM and GRU however yielded prediction accuracy of the order of 89–91 % for differential pressure at the venturimeter located about 10D downstream of the bend. The study indicates feasibility for wider use of machine learning approaches on time varying pressure drop information for assessment and prediction of flow regimes in two phase gas liquid flow.</div></div>","PeriodicalId":50440,"journal":{"name":"Flow Measurement and Instrumentation","volume":"108 ","pages":"Article 103176"},"PeriodicalIF":2.7,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145883594","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Experimental and numerical simulation of the gas-liquid micro-cyclone separator with slotted outflow pipe","authors":"Junfeng Zhang ,&nbsp;Xin Wang ,&nbsp;Shenglei Zhang ,&nbsp;Yang Lv","doi":"10.1016/j.flowmeasinst.2025.103153","DOIUrl":"10.1016/j.flowmeasinst.2025.103153","url":null,"abstract":"<div><div>Gas-liquid cyclone separators are widely used in various industrial processes. To reduce energy loss in traditional cyclone separators and enable their application in chemical absorption-based CO₂ capture systems for recovering liquid droplets from flue gas, this paper proposes a slot-type outflow pipe cyclone separator. Building upon existing research on slotted outflow pipes, this study optimizes the slot configuration and investigates how different slot angles affect separator performance. An innovative control group was established to determine whether the presence of a through-hole at the bottom of the outflow pipe impacts the separator's applicability. The RSM model simulates the gas-phase flow field, while the DPM model, within the multiphase framework, tracks droplet particles. Results show that the slotted outflow pipe structure not only reduces static pressure appropriately but also optimizes pressure distribution and turbulence intensity, enhances flow field stability, and suppresses the short-circuit flow phenomenon. When the bottom of the outflow pipe is unsealed and the groove angle is 90°, cutting particle size and Critical particle size reach optimal values, with reduction rates of approximately 20 % and 23 %, respectively, compared to the Lapple-type. The pressure drop reduction is about 3.3 %. At a constant liquid concentration of 8.1 g m⁻³, the intake volume reaches 22.56 m³ h⁻¹, yielding a separation efficiency of 80 %. At an intake volume of 34.21 m³ h⁻¹, the separation efficiency increases to 89.35 %, with a lower pressure drop than for the other two outflow pipe configurations.</div></div>","PeriodicalId":50440,"journal":{"name":"Flow Measurement and Instrumentation","volume":"108 ","pages":"Article 103153"},"PeriodicalIF":2.7,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737203","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Analysis of hydraulic characteristics of submerged gate discharge based on hybrid artificial intelligence models","authors":"Dingye Cao ,&nbsp;Guodong Li ,&nbsp;Shanshan Li","doi":"10.1016/j.flowmeasinst.2025.103155","DOIUrl":"10.1016/j.flowmeasinst.2025.103155","url":null,"abstract":"<div><div>With the expansion of water conservancy projects and the complexity of water resources scheduling problems, calculating submerged flow has become increasingly critical and challenging, and improving its computational efficiency and accuracy has become a key problem. This study analyzes the hydraulic characteristics of sluice gate submerged flow based on a long-distance water transmission project. Random Forest (RF) and its optimization algorithm Firefly Algorithm-optimized Random Forest (FA-RF) were used to build an intelligent prediction model, with performance evaluated through index and uncertainty analysis. The Sobol sensitivity analysis method is used to quantify the hydraulic parameters affecting submerged flow, identifying core parameters influencing the submerged discharge coefficient <em>C</em>_<em>d</em>. The results show that for the submerged flow, the greater downstream water depth causes stronger upstream flow disturbance. In the testing phase, the indicators of the FA-RF model are better than RF, and the mean absolute percentage error (MAPE) and root mean square error (RMSE) are reduced by 20.9 % and 24.3 %, respectively, indicating that the firefly algorithm optimizes the performance of the RF algorithm. The upstream Froude number <em>Fr</em> and the ratio of downstream water depth to gate opening <em>y</em>_<em>3</em><em>/e</em> were identified as core parameters affecting <em>C</em>_<em>d</em>. <em>C</em>_<em>d</em> increases with <em>Fr</em> and decreases with <em>y</em>_<em>3</em><em>/e</em>.</div></div>","PeriodicalId":50440,"journal":{"name":"Flow Measurement and Instrumentation","volume":"108 ","pages":"Article 103155"},"PeriodicalIF":2.7,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737283","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prediction of solid-liquid two-phase flow velocity in horizontal pipelines via ensemble learning","authors":"Lili Pang ,&nbsp;Hanchuan Dong ,&nbsp;Xiaotong Kong ,&nbsp;Pan Zhang ,&nbsp;Jianhua Feng ,&nbsp;Lide Fang","doi":"10.1016/j.flowmeasinst.2025.103156","DOIUrl":"10.1016/j.flowmeasinst.2025.103156","url":null,"abstract":"<div><div>Accurate estimation of the cross-sectional average velocity in liquid-solid two-phase flows is critical for understanding and managing multiphase transport processes. However, the complex behavior of suspensions, particularly when particle characteristics are unknown, makes accurate velocity estimation highly challenging. To overcome this limitation, this study presents a novel approach that integrates multi-frequency ultrasonic backscatter measurements with an ensemble learning framework. Unlike conventional Doppler techniques, the proposed method considers both particle-fluid interactions and varying hydraulic conditions, enabling accurate velocity prediction under partially filled regimes. A dimensionless ensemble learning model based on the Stokes number was developed, incorporating particle size, concentration, and flow parameters to improve the accuracy of cross-sectional mean liquid velocity estimation. Experimental validation and comparison with existing data verify that the proposed Gradient Boosting Tree model achieves a mean absolute percentage error below 5.21 %, with 98 % of predictions falling within a relative uncertainty of ±20 %. These results highlight that integrating multi-frequency ultrasonic backscatter technology with ensemble learning constitutes a novel and robust approach for the precise prediction of horizontal liquid-solid two-phase flow velocities.</div></div>","PeriodicalId":50440,"journal":{"name":"Flow Measurement and Instrumentation","volume":"108 ","pages":"Article 103156"},"PeriodicalIF":2.7,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145684980","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}