Figure - uploaded by Yanan Camaraza-Medina
Content may be subject to copyright.
Source publication
Dimensional analysis was used on this study with the aim of stablishing a model for prediction of the monoethanolamine heat exchangers output. The passive experimentation method was applied to gather 14 400 data points, since the exchangers are installed in an
online amine treating unit. After identification of those parameters having a relevant ef...
Contexts in source publication
Context 1
... procedure consisted of measurement and observation of input and output variables within the normal working regime of investigated heat exchangers, without any manipulation from the researchers, thus analyzing the heat transfer mechanisms as it actually happens [20]. Readings of mass flowrates, inlet and outlet temperatures were performed on each fluid, with the primary instrumentation (Table 1) linked to a Siemens PLC and the SCADA system. A 14 400 data points database was gathered by recording the variables every minute during ten successive days. ...Context 2
... procedure consisted of measurement and observation of input and output variables within the normal working regime of investigated heat exchangers, without any manipulation from the researchers, thus analyzing the heat transfer mechanisms as it actually happens [20]. Readings of mass flowrates, inlet and outlet temperatures were performed on each fluid, with the primary instrumentation (Table 1) linked to a Siemens PLC and the SCADA system. A 14 400 data points database was gathered by recording the variables every minute during ten successive days. ...Citations
... Despite several parameter estimation methods are applied to obtain heat transfer correlations from experimental data, as briefly summarized by Tam et al. [17,18], there are only two approaches (according to the authors best knowledge) that eludes anticipated selection of the Nusselt-equation functional form. The first one is symbolic regression, performed through Genetic Programming algorithms, which not only allows determination of the equation constants, but also its mathematical structure [19][20][21]. ...
... However, the resultant expressions are aleatory, have a limited use in practical applications, and are deprived of a theoretical support. The second approach is referred here in as the 'Nusselt-equation simulated evolution method', and was devised to link relevant analogies among momentum, heat and mass transfer [18]. It automatically converges into the lowesterrors correlation, through successive modifications of the Nusselt number fitting function. ...
... This method was recently devised by Sá nchez-Escalona et al. [18], as a novel approach to obtain the mean heat transfer coefficients for single-phase non-laminar fluid flow inside tubes. It essentially consists on a comprehensive Nusseltnumber equation, according to Eq. (13), that is able to evolve into three different functional forms derived from the analogies among momentum, heat and mass transfer as proposed by Reynolds-Colburn, Prandtl and von Karman. ...
Reliability of heat exchangers thermal analysis strongly depends on the equations selected to determine local convective heat transfer coefficients. Chosen analogy among momentum, heat and mass transfer also plays a remarkable role. Within this context, the aim of the study was to validate a novel approach to obtain mean forced convective film coefficients under single-phase non-laminar fluid flow conditions, inside tubes. It relied on a comprehensive Nusselt-number equation that is able to evolve into different functional forms according to Reynolds-Colburn, Prandtl and von Karman analogies.
Parameters estimation was carried out through Genetic Algorithms. Applied experimental database was numerically obtained by Taler by solving the energy conservation equation for fully developed turbulent flow in tubes with constant wall heat flux. Application of the method provided a new correlation, valid for 0.1 ≤ 𝑃𝑟 ≤ 103 and 3 × 103 ≤ 𝑅𝑒 ≤ 106 . Besides attaining a better fit to the experimental data as compared to benchmark expressions, it correlated very well with the results of reference models (Skupinski, Seban & Shimazaki, Gnielinski, Camaraza-Medina, Petukhov, and Sandall). The first assessment provided mean and maximum relative errors of 2.41% and 19.45%, respectively, while the second comparison resulted in deviations over the Nusselt number up to 20% in 92.59% of the data points. The implemented solution overcomes the drawbacks of non-linear and symbolic regression methods by allowing evolution of the regression function within a controlled mathematical environment. Future model improvements should investigate different fitting-intervals along with
higher turbulence regions.
... Ambos escenarios de modelación se consideran apropiados para simular el desempeño de los intercambiadores de calor MEA-MEA en condiciones cambiantes de operación, ya que los índices de error calculados no tienen una implicación significativa en el proceso tecnológico examinado (tabla 3). Este enfoque de "caja negra" empleado en la modelación y simulación del objeto de estudio no sólo elude el cálculo del coeficiente global de transferencia de calor, sino que ofrece una alternativa para evaluar el desempeño de los equipos teniendo en cuenta que las propiedades termo-físicas de las soluciones de amina dependen de varios parámetros (temperatura, presión, fracción másica y carga de CO2), no están totalmente caracterizadas y son difíciles de obtener [18]. En el período analizado (22 días) se considera que la red neuronal con entrenamiento único no perdió la capacidad de generalización, ya que además de reproducir con exactitud los datos para los cuales fue entrenada (99,89 % de correlación y 0,438 K 2 error cuadrático medio) se lograron predicciones confiables frente a los juegos de datos nuevos [2,15,17]. ...
Esta investigación propuso un modelo de red neuronal artificial con entrenamiento dinámico para predecir las temperaturas de salida de ambos fluidos en un sistema de intercambiadores de calor para monoetanolamina, realizando el entrenamiento, validación y pruebas con 31 680 juegos de datos obtenidos a través del método de
experimentación pasiva. Con el perceptrón multicapa 4-3-2 se lograron correlaciones superiores al 98,76 %, y se corroboró que el entrenamiento dinámico proporciona respuestas más precisas que el entrenamiento único. Aplicando el primer enfoque se calcularon errores absolutos promedio de 0,419 y 0,372 K (para las variables
temperatura de salida de la amina rica y de la amina pobre, respectivamente), comparados con 2,214 y 1,181 K para el segundo. Tales desviaciones no tienen una implicación significativa en el proceso tecnológico examinado, por lo que el modelo se considera apropiado para simular el desempeño de los intercambiadores de calor objeto de estudio.
... The study focused on the hydrodynamics of flow in the corrugated channels using the Particle Image Velocimetry (PIV) technique for Reynolds numbers of (4000) and (6000). More recently, Dimensional analysis was used to establish a model for prediction of the mono-ethanolamine heat exchangers output [18]. The Buckingham pi theorem and the method of repeating variables were implemented to predict the outlet temperature of known operating conditions for such heat exchangers. ...
Dimensional analysis was utilized on this research to establish a shortcut model for predicting hydrogen sulphide gas discharge temperature in jacketed shell and tube heat exchangers. Since the equipment belongs to an online industrial facility, the passive experimental method was applied. Selection of the heat transfer process parameters was followed by application of the Buckingham Pi-theorem and the repeating-variables technique. After formulation of the dimensionless groups, approximation of the explicit model equation was carried out through a least-squares multivariate linear regression. The model predictive ability performance was appraised by comparing predictions versus measured discharge temperatures, hence attaining a Pearson correlation of 97.5 %, a mean absolute error of 2.1 K, and 1.7 % maximum deviations. The explicit equation that was obtained is pertinent to studied heat exchangers, when 0.55 ≤ ṁ1 ≤ 0.60, 1.06 ≤ ṁ2 ≤ 1.09, and 0.22 ≤ ṁ3 ≤ 0.24 (fluids flowrate, kg/s). It can be used as an alternative calculation method for quick anticipation of the equipment performance, which overcomes computation of the overall heat transfer coefficients.
