Search

The supply of species gases and removal of liquid water in a PEM fuel cell are performed through gas flow channels. Therefore, optimising the flow channel configurations is critical for maximising performance of fuel cells. In this paper, the ANSYS PEM Fuel Cell Module is used to simulate a PEM fuel cell with an active area of 100 cm2 and four different flow channel configurations: single-channel serpentine, two-channel serpentine, three-channel serpentine, and parallel channel with headers under two set of different operating conditions. Under both sets of operating conditions, the simulation results indicated that 3-channel serpentine configuration has the best performance due to uniform distribution of species gases over the catalyst layer. It was concluded from the simulation results that the pressure-drop along the 3-channel serpentine configuration is less compared to other serpentine channel configurations, which will require less power for the blowing in of the species gases. The results obtained from simulation under more optimal operating conditions (set 2) used by Hwang’s URFC, predicted increase in the performance of all the four channel configurations compared to the set 1 operating conditions due to higher cell temperature and stichometry of inlet gases. Under set 2 operating conditions, the values of pressure drop across all the channel configurations have increased substantially due to higher flow rates of inlet gases.

In the present work, a modified Pressure Swirl Atomizer (PSA) was introduced to operate as an Assisted Pressure Swirl Atomizer (AAPSA). Different mass flow rates of assisted air were experimentally tested. The studied parameters are the orifice length to diameter ratio (L/Do), swirl chamber length/diameter ratio (Ls/Ds) and swirling passage size (width × depth). The spray characteristics were studied using water as the atomized liquid. The spray was photographed using a high-speed camera to determine the spray shape, the cone angle and the spray breakup length. Moreover, photos taken by the high-speed camera were processed using IMAGE J software to obtain the droplet size diameters. The spray concentrations were measured by the patternator tubes technique. The results indicated that the spray cone angle increased by increasing the value of the L/Do ratio or by decreasing the assisted air mass flow rate. The spray concentration maximum value is shifted inward with the assist air injection. The breakup length decreased by about 65% for the PSA and 77% for the AAPSA when the value of the Ls/Ds ratio is decreased from 1.163 to 0.664. With the injection of assist air at a rate of 2 g/s, the Sauter Mean Diameter (SMD) is reduced by about 64% compared to a PSA having the same geometry.

Small tube irrigation is a new type of micro-irrigation that the emitters were replaced by small tubes. The small tubes of 4 mm diameter were connected to the laterals through the flow stabilizer or directly to the laterals. The discharge formulations and total heads of the system have been driven from the basic formulas of friction losses and other minor losses of the laterals. Based on hydraulic analysis of laterals, the cost function of lateral and pump, the operation cost and the objective function of lateral diameter optimization technique was established. In the study, a typical irrigation system of fruit tree was taken as an example, in which the length of lateral pipe was 100 m, and assuming that the small tube flow range is 10–40L/h, lateral diameter ranges 12–25 mm, and the field slope ranges −20%∼20%. The results showed that the minimum total cost was obtained with a lateral diameter of 13.6 mm at 10 L/h flow rate, for the flow rate of 20L/h, the minimum cost is obtained at the diameter of 13.6 mm for down-slopes and 17.2 for flat land and up-slopes, and for the flow rate of 30L/h and 40L/h the minimum cost is obtained at 17.2 mm diameter for any slopes.

This study focuses on the heat transfer characteristics of the evaporation terminal, the cool distribute unit (CDU) and refrigerant flow distribution of a water cooled multi-spilit heat pipe system (MSHPS) used in data center. The finite time thermodynamic analysis, the exergy method and the software SIMULINK was employed to build the simulation model of the combined system. The results show that the IT servers should concentrate on arranging at the location below 1.3 m. The CDU has a heat transfer of about 74 J in a period of 6 s. And the optimum flow rate of the CDU is 0.82 kg/s. The flow distribution characteristic of a CDU which connect 2 heat pipe evaporator terminals of 6 kW was calculated, and the working fluid is R22. Then the free cooling time,part time free cooling and energy saving potential in major cities of China were analysised. The energy saving potential is from 61% to 25%. The results are of great significance for the operational control and practical application of a MSHPS and other pipe-net systems.

How to achieve uniform mixing of highly viscous fluids with low energy consumption is a major industry demand and one of the hot spots of mixing research. A typical multistage rotor-stator mixer (MRSM) equipped with a distributor was investigated to disclose the effects on the mixing performance and power consumption for highly viscous fluids via numerical simulation, considering the influence factors associated with different geometric parameters of both MRSM and the distributor. The mixing index and power consumption are used to evaluate the performance of the mixers. The dimensionless correlations for the mixing index and the power consumption are established considering the factors including the flow rate, rotor speed, the number of mixing units. Adopting the optimized mixer with the distributor (X1-T1), the mixing index increases to 0.85 (obviously higher than 0.46 for the mixer T1 without a distributor), meanwhile the corresponding power consumption is about 1/5 of that of T1 achieving the same mixing effect. It illustrates that the distributor can significantly improve the mixing of highly viscous fluids in the MRSM without the cost of large power consumption. These results would provide a guidance on the design and optimization of multistage rotor-stator mixers in industrial applications.

The excessive exploitation of rare earth elements (REEs) has caused major losses of non-renewable resources and damage to the ecosystem. The processes of mining and smelting produce massive amounts of wastewater with low concentrations of REEs. Consequently, the enrichment and recovery of low-concentration REEs from wastewater has significant economic and environmental value. For this purpose, operation under large phase ratios (the flow rate ratio between the aqueous phase and extractant) is more desirable and economically viable. However, the traditional REE extraction process suffers from the uneven dispersion of the extractant and the difficulty of phase separation, which leads to long extraction times and large consumption of extractants. Hence, there is an urgent need to develop a green and efficient technique to extract low concentrations of REEs from wastewater. In this work, a droplet-based microfluidic technique was used to continuously extract and recover low-concentration REEs at large phase ratios. Snowman-shaped magnetic Janus nanoparticles were added to the continuous phase as emulsifiers to facilitate uniform extractant dispersion and rapid phase separation. Several key factors affecting the extraction efficiency, including pH, residence time, and the amount of added Janus nanoparticles, were systematically investigated. Compared to batch extraction, droplet-based microfluidic extraction with the addition of Janus nanoparticles showed the advantages of a large specific surface area and fast phase separation during extraction. Meanwhile, the Janus nanoparticles exhibited good emulsification performance after three extraction cycles. In summary, the Janus nanoparticle-stabilized droplet generated by microfluidic methods provides a feasible path for the efficient enrichment and recovery of low-concentration REEs.

There are various models which describe the hydro-dynamical aspects of the fluid flow and solute transport in small diameter permeable tubes without incorporating the slip effects, which play an important role to predict the flow behavior with higher accuracy. This study aims to investigate the slip effects on the slow, steady-state and axisymmetric flow of an incompressible second order Rivin-Ericksen fluid in a small diameter cylindrical tube with constant wall permeability. The momentum equations have been derived in cylindrical coordinates and are converted in the form of a single compatibility equation which governs the stream function of the flow. Navier slip condition has been used to take slip phenomenon into account near the wall of the tube. Compatibility equation has been solved analytically subject to the associated boundary conditions by using Langlois recursive approach method. The detailed expressions have been obtained for velocity components, pressure, volume flow rate, pressure drop along longitudinal axis, wall shear stress, fractional reabsorption and leakage flux. Behavior of all the flow variables against the slip parameter has been examined graphically in detail. Overall, the solutions obtained her e show a great agreement with the work in literature. One of the key findings in this study is that the volume flow rate Q is found to be independent of the slip parameter ϕ .

Relight of jet engines at high altitude is difficult due to the relatively low pressure and temperature of inlet air. The penetration of initial flame kernel affects the ignition probability in the turbine engine combustor greatly. In order to achieve successful ignition at high altitude, a deeper penetration of initial flame kernel should be generated. In this study, a Gliding Arc Plasma Jet Igniter (GAPJI) is designed to induce initial flame kernel with deeper penetration to achieve successful ignition at high altitude. The ignition performance of the GAPJI was demonstrated in a model combustor. It was found that GAPJI can generate plasma with deeper penetration up to 30.5 mm than spark igniter with 22.1 mm. The discharge power of GAPJI was positively correlated with flow rate of the carrier gas, approaching 200 W in average. Ignition experiments show that GAPJI has the advantage of extending the lean ignition limit. With GAPJI, the lean ignition limit of the combustor is 0.02 at 0 km, which is 55.6% less than that with spark igniter (0.045). The evolution of flame morphology was observed to explore the development of the flame kernel. It is shown that the advantage of a high penetration and continuous releasing energy can accelerate the ignition process and enhance combustion.

Because of the rapid advancement of microdevices technologies, electroosmotic flow (EOF) has gotten a lot of coverage in recent years. EOF has a range of benefits over other micropumps, including the ability to create continuous pulse-free flows and elimination of moving parts. Our aim is to investigate the unsteady electroosmotic flow (EOF) of an electrolyte solution of generalized fractional second grade hybrid nanofluid (Cu-TiO 2 / Water) confined between vertical two coaxial tubes. The energy equation will be altered by the fractional Cattaneo model. The finite Hankel technique and Laplace transformation will be employed to acquire the solution of the mathematical model. Since obtaining the inversion of Laplace of the transformed functions in the closed-form is a challenging task, the numerical inversion algorithm will be used. The results showed that the hybrid nanofluid became more viscous with increasing the volume concentration which reduces the electroosmotic (EO) velocity and the flow rate. Moreover, the electroosmotic flow is boosted by the free convection force.

Streams and rivers, particularly smaller ones, often do not maintain steady flow rates for long enough to reach equilibrium conditions for sediment transport and bed topography. In particular, streams in small watersheds may be subject to rapidly changing hydrographs, and relict bedforms from previous high flows can cause further disequilibrium that complicates the prediction of sediment transport rates. In order to advance the understanding of how bedforms respond to rapid changes in flow rate, a series of flume experiments were performed where the flow was reduced rapidly from equilibrium conditions. Sediment transport rates and bed elevation data across the flume and over a 15-meter-long test section were collected during the experiments to allow detailed examination of evolving bedform dynamics. It was found that relict bedforms stopped moving completely after flow reductions, and the mode of sediment transport was shifted to small bedforms that arose rapidly over dune stoss sections throughout the test section. The changes in sediment transport with time as the sand bed adjusted to the new flow rate was found to agree with predictions based on the relations proposed in Wren et al. (2020). Wavelet analysis is used to visualize changes in length and amplitude scales during the bed transition process.

The discharged volatile organic compounds (VOCs) have aroused more and more attention because of increasing serious air pollution. Accurate prediction of the emission is very important for the industry. Therefore, two kinds of model were developed to predict the VOCs removal efficiency in the RPB, including empirical model and artificial neural network (ANN) models. The empirical model was mainly including gas and liquid mass transfer, effective contact area and liquid holdup. The ANN models were including Cascade-forward back propagation neural network, Feed-forward distributed time delay neural network, Feed-forward back propagation neural network and Elman-forward back propagation neural network. The input parameters were dimensionless numbers, such as high gravity factor, liquid Reynolds number, gas Reynolds number and dimensionless Henry coefficient. The output parameter was removal efficiency. The outlet concentrations of different VOCs predicted by the ANN model were much better than those of the empirical model. And the disadvantages and the advantages were also analyzed. The effects of high gravity factor, gas flow rate and liquid flow rate on the removal efficiency in the RPB were simulated by ANN model with high precision. And then, operation conditions could be timely adjusted to meet the emission standards.

Multi seeds oil expeller was designed fabricated and evaluated using screw press principle. It was designed and fabricated based on the data obtained on the properties of the oilseeds. The performance of the machine was carried out using a 3x3x3 factorial experimental design with the following type of seeds (palm kernel, soya bean and groundnut seed), worm shaft speed of 47.6, 87and 92 rpm, and moisture content of 5, 10, and 15 % w.b as factors. Data collected include oil yield, volumetric flow rate of oil, throughput capacity and efficiency of the machine. The results showed that increase in moisture content as stated has direct effect on oil yield and efficiency of the expeller, but has an inverse effect on volumetric flow rate for the three oil seeds. Increase in worm shaft speed as stated increased oil recovery for palm kernel from 19.1 to 23.0%, soya beans from 0.0 to 16.0% and groundnut from 7.3 to 31.7%. Increase in worm shaft speed as stated increased the oil yield, throughput capacity as well as the efficiency of the machine for all the oil seeds. The efficiency increased from 39.4 to 52.1%; 0.0 to 82.1% and 53.1 to 72.0% for palm kernel; soya bean and groundnut, respectively. These results were found to be significant (p

Drop microfluidics has driven innovations for high throughput, low input analysis techniques such as single-cell RNA-seq. However, the instability of single emulsion (SE) drops occasionally causes significant merging during drop processing, limiting most applications to single-step reactions in drops. Here, we show that double emulsion (DE) drops address this critical limitation and completely prevent content mixing, which is essential for single entity analysis. DEs show excellent stability during thermal cycling. More importantly, DEs undergo rupture into the continuous phase instead of merging, preventing content mixing and eliminating unstable drops from the downstream analysis. Due to the lack of drop merging, the monodispersity of drops is maintained throughout a workflow, enabling the deterministic manipulation of drops downstream. We also developed a simple, one-layer fabrication method for DE drop makers. This design is powerful as it allows robust production of single-core DEs at a wide range of flow rates and better control over the shell thickness, both of which have been significant limitations of conventional two-layer devices. This approach makes the fabrication of DE devices much more accessible, facilitating its broader adoption. Finally, we show that DE droplets effectively maintain the compartmentalization of single virus genomes during PCR-based amplification and barcoding, while SEs mixed contents due to merging. With their resistance to content mixing, DE drops have key advantages for multistep reactions in drops, which is limited in SEs due to merging and content mixing.

Microfluidics is characterized by the manipulation of fluids in submillimeter channels and has great application potential in encapsulation. To further extend the application of microfluidics in food industries, a 3D printed microfluidic device is used to encapsulate vitamin A and improve its stability. Two natural macromolecules, sodium alginate and gelatin, are added to water as the continuous phase to generate monodisperse emulsion. Under different flow rate ratios, the diameter of droplets decreases with the increase of continuous flow rate. However, at the same flow rate ratio, varying the dispersed and continuous flow rates does not significantly change the diameter and size distribution of emulsion collected. The prepared O/W (oil/water) single emulsion can form microgel particles and avoid degradation of vitamin A by simulated gastric acid; the encapsulated vitamin A will not be released until particles reach simulated intestinal tract. In the simulated digestion in vitro, no vitamin A is released for 2 h in the acidic environment; under an alkaline or neutral environment such as those in intestinal fluids, vitamin A can be released from the microgel particles within 2.5 h. Using the presented approach, emulsions encapsulating vitamin A have been prepared and can potentially be applied to encapsulate other oil-soluble substances in the food industry.

This paper investigates the force features of inclined pipes under the action of waves and flows through both test and numerical simulation. Firstly, the wave and flow loads on inclined pipes were examined through tests in circulating water flumes. Next, the unsteady Reynolds-averaged Navier-Stokes (RANS) equations were adopted to numerically simulate the test conditions. Multiple factors were taken into consideration: immersion depth, inclination, flow rate, wave height, and cycle. The results show that, the wave and flow loads on the pipes both increase with wave height and flow rate. When the inclination β = 0°, the flow force is most rapid, conversely, it is most slow when the inclination β = 30°. Furthermore, when the pipe was immersed by 75%, the change of inclination directly affected the ratio of vertical force to horizontal force. For the pipes with different inclinations, the wave load increases basically at the same rate with the rising wave height. With the widening of the inclination, the wave and flow loads on the pipes become less sensitive to immersion depth, and the maximum loads remain near the free water surface.

The study of multiphase flows has earned much attention from researchers due to its numerous applications in mechanics and industry such as sewage management, synthetic fabrication, electricity generation, and medication processes. In the present work, we examined the Poiseuille flow of a couple stress fluid between parallel plates in presence of velocity and temperature slip boundary conditions. The effects of thermal conductivity and thermal radiations, variable viscosity, and non-uniform magnetic field on velocity and temperature fields in the existence of hafnium nanoparticles are also investigated. The temperature-dependent viscosity model, namely, Vogel’s model is utilized. The non-linear system of equations is solved numerically by using the shooting technique due to its fast convergence and prosperity to initial approximation. It is examined that the velocity slip parameter speeds up the flow rate of fluid but decelerated the temperature field while the temperature slip parameter upsurges the temperature of the fluid. It is also investigated that the concentration of nanoparticles enhanced both the velocity and temperature of the fluid. Further, the comparison of Newtonian and non-Newtonian (Couple stress fluid) is expressed with the help of graphs. To check the validation of results, a comparison with previous literature is also presented.

The bubble formation dynamics and size manipulation in the slurry of polystyrene microspheres in the microfluidic T-junction were visually investigated by a high-speed camera. Based on the evolution law of the bubble neck width with time, the formation process of bubbles is divided into three stages: filling, squeezing and pinch-off. The particle concentration has an obvious effect on the squeezing stage, while less impact on the filling and pinch-off stages. In the squeezing stage, the evolution of the dimensionless minimum neck width of bubbles with time could be described by a power-law relationship. The increase of the particle concentration or continuous phase flow rate could lead to the increase of body flow of the continuous phase flow rate and the enhancement of the squeezing force acted on the bubble neck, correspondingly, the power-law index α in the squeezing stage enlarges. Moreover, the bubble size increases with the increase of the gas phase flow rate and the decrease of the particle concentration and continuous phase flow rate. However, the effect of the particle concentration on the bubble size weakens with the increase of the continuous phase flow rate. In addition, a new prediction correlation of the bubble size for the slurry system in a T-shape microchannel was proposed with good prediction accuracy.

Electrochemical drilling (ECD) provides an alternative technique for drilling multiple small holes in difficult-to-machine materials in numerous industrial applications such as for aeroengines. The value and fluctuation of electrolyte flowrate can seriously affect the machining stability and hole quality in ECD. In particular, when drilling multiple holes, the distribution and fluctuations of the electrolyte flowrate in each channel could influence the uniformity of the electrolyte flowrate among multiple tube electrodes, thereby affecting the machining stability and the maximum feed rate. Thus, an eight-channel flow control system was developed to measure and regulate in real time the electrolyte flowrate supplied into each individual tube electrode. This paper proposes ECD of small holes with real-time flowrate control to improve the uniformity of electrolyte flowrate in each tube electrode and reduce the fluctuations in the electrolyte flowrate. In single-hole drilling, when the electrolyte flowrate was regulated in real time, the hole quality and machining stability were considerably better than without regulation. This is because the electrolyte flowrate remains basically constant, which stabilizes the flow field. Moreover, by considering the hole profile, it was found that an electrolyte flowrate of 200.0 mL/min can be acceptable for ECD of small holes. When the eight-channel flow control system was used in multiple-hole drilling, the uniformity of the electrolyte flowrate in each tube electrode was obviously improved, which led to a more stable process. Additionally, the maximum feed rate can attain 2.40 mm/min in multiple-hole drilling. Based on these findings, a matrix (5 × 32) of multiple small holes was successfully fabricated with a satisfactory diameter consistency, as the machining stability and machining efficiency had been enhanced.

In the titanium dioxide industry, there is a lack of a low-cost and high-efficiency treatment method for chloride containing tail gas. In this paper, the removal of HCl from the titanium dioxide industry by gas cyclone-liquid jet separator was studied, while Ca(OH)2, Na2CO3, NaOH solution, and water were used as absorbents. This paper investigated the influence of gas cyclone-liquid jet separator’s various process parameters on the removal rate of hydrogen chloride gas. The mechanism of mass transfer in the process of removing hydrogen chloride was discussed, and the effect and feasibility of HCl gas removal in the gas cyclone-liquid jet absorption separator were studied. The results showd that the removal efficiency of hydrogen chloride maintained above 95%, up to 99.9%, and the total mass transfer coefficient reached 0.28 mol·m-3·s-1·kPa-1. Under the same conditions, the absorption effect and total mass transfer coefficient of weak basic absorption liquid can be greatly improved by increasing the flow rate of absorption liquid, but the absorption effect and total mass transfer coefficient of strong alkaline absorption liquid can’t be improved obviously. The larger the inlet gas volume, the higher the gas concentration, the lower the absorption efficiency and the lower the total volumetric mass transfer coefficient.

Thermoelectric generator can directly convert thermal energy into electrical energy, and applying it to recover waste heat from automobile exhaust is conducive to energy saving. In this paper, a thermoelectric generation experimental system is built, the power output performance and channel power consumption of the flat-plate thermoelectric power generation system under different exhaust gas flow rates are experimentally studied. The results show that when the exhaust flow rate is 40 Nm3/h and the exhaust inlet temperature is 400 °C, the maximum output power of the system is 26.8 W, and the maximum net output power is about 23.3 W. In addition, the exhaust flow rate is changed from 10 Nm3/h to 40 Nm3/h, and the proportion of power consumption in the hot air channel is changed from 1.91% to 12.85%, which has increased by more than six times.

The aims of this study is to design and optimize the functioning of a full continuous combined process based on electrocoagulaion-adsorption on crude Tunisian clay to treat a real textile effluent. The clay characterization shows that the used clay is a rich-smectite clay. The response surface methodology (RSM) technique based on Box-Behnken design (BBD) was used to optimize the process. At optimum conditions which are initial pH solution of 8.24, effluent flow rate of 0.5 L·min−1, voltage of 70 V, and added suspension of clay flow rate of 100 ml·min−1 the achieved color, chemical oxygen demand (COD) and total suspended solid (TSS) removal efficiencies were respectively 96.87%, 89.77% and 84.46% with 0.75USD·m−3 as total cost. The additional laboratory experiments at optimum conditions agree with the predicted results, which confirm the accuracy and the capability of RSM to predict results in the defined space. Finally the designed process could be a good eco-friendly alternative to treat and reuse wastewater in industrial process with reasonable cost.

In this study, we investigated the performance of a synthetic resin for the adsorption of Li from pre-desilicated solution which is the waste liquid produced by extracting aluminum from fly ash. The adsorption kinetics and isotherms of the resin were obtained and analyzed. The saturated adsorption sites of the resin were in agreement with the quasi-second-order kinetic model. Then, the pore diffusion model (PDM) was applied to represent the lithium adsorption kinetics which confirming that the external mass is the limiting step. Moreover, we evaluated the adsorption properties of this resin in fixed-bed mode. We established a feasible extraction process for Li from strong alkaline solutions with low Li concentrations. The process parameters, such as the flow rate, initial adsorption solution concentration, water washing process, desorption agent concentration, and flow rate were studied. The desorption rate of the Li+ ions was directly proportional with the concentration of the desorption agent. The time required to accumulate Li decreased as the hydrochloric acid concentration and flow rate increased. Time of the peak appeared increased from 0.5 bed volume (BV) to 2.5 bed volume (BV) as the concentration was increased from 1 to 3 mol·L−1, and the peak increased from 231 to 394 mg·L−1. The resin presented good selectivity for Li+ ions and could effectively separate impurity ions from the pre-desilication solution.

Many aspects relating to thermal, economic, and reliability challenges are involved while designing a Photovoltaic/Thermal (PV/T) system under high performance. When compared to standard PV systems, PV/T systems are more effective in locations with high irradiance. To eliminate the excess heat from a PVT system, an optimal cooling system is necessary, resulting in improved the temperature of PV module which directly effect on the overall performance. The current study is focused on a new economic and essay construction thin and thick (3 mm and 15 mm) cooling cross-fined channel box that have covered modeled. In terms of cell temperature and output cooling water temperature, the effects of various irradiation levels and cooling water flow rates are examined. Software that is based on finite elements, the problem was solved numerically in a three-dimensional model using ANSYS (19.2). It has been discovered that when the cooling fluid flow rate increases, the average surface temperature distribution of PV/T system and outlet cooling water temperature are reduced for thin and thick box heat exchanger and ideal with a cooling fluid flow rate of 3 L/min. The average surface temperature and water-cooling temperature are 30.7 °C and 38.9 °C for thin box heat exchanger, respectively, whereas for thick box heat exchanger are 30.8 °C and 32.4 °C, respectively, under solar irradiance 1000 W/m2 and optimal flow rate.

In this study, the effects of wax deposition on submarine multiphase pipelines are investigated. In order to understand the mechanism of wax deposition in submarine multiphase pipelines, a wax deposition model in a submarine pipeline was established using the OLGA wax deposition module, and key factors influencing this phenomenon were analyzed. Additionally, a multifactor impact analysis of wax deposition was performed via the orthogonal test method. The results indicated the relative influence of five factors on oil-gas-water three-phase wax deposition, namely, oil flow rate, water content, inlet temperature, gas-oil ratio, and outlet pressure. Then, the main influencing factors of wax deposition in a multiphase pipeline flow were extracted, and the wax deposition prediction model was established. The wax deposition rate under different operating conditions was simulated using the OLGA wax deposition simulation module. The SPSS software was used to perform nonlinear regression analysis under different working conditions. In this manner, the constant terms in the wax deposition prediction model for a submarine multiphase pipeline were obtained. The prediction model could be used to programmatically predict wax deposition along a multiphase pipeline by programming the initial waxing conditions. By using this method, the wax deposition prediction model of an M − N submarine multiphase pipeline was obtained under different working conditions. After comparing the predicted and the OLGA simulation results, it was determined that the wax deposition model established could be used to accurately calculate the wax deposition in a submarine multiphase pipeline within an allowable error range. If the physical properties and operating parameters of the conveying medium in a multiphase pipeline are given, the developed oil-gas-water three-phase wax deposition model can be used to predict the distribution of wax deposition in submarine multiphase pipelines without laboratory experiments. The results of this study can help production units understand the waxing situation of pipelines in time in order to scientifically formulate a pigging plan and avoid pipeline blockages and shutdowns.

s Feed-water valves (FWVs) are used to regulate the flow rate of water entering steam generators, which are very important devices in nuclear power plants. Due to the working environment of relatively high pressure and temperature, there is strength failure problem of valve body in some cases. Based on the thermo-fluid-solid coupling model, the valve body stress of the feed-water valve in the opening process is investigated. The flow field characteristics inside the valve and temperature change of the valve body with time are studied. The stress analysis of the valve body is carried out considering mechanical stress and thermal stress comprehensively. The results show that the area with relatively high-velocity area moves gradually from the bottom of the cross section to the top of the cross section with the increase of the opening degree. The whole valve body reaches the same temperature of 250 °C at the time of 1894 s. The maximum stress of the valve body meets the design requirements by stress assessment. This work can be referred for the design of FWVs and other similar valves.

A valve-based piezoelectric pump operated by three bending transducers is proposed in this work. The major features of the proposed pump are self-priming and high pressure. The structure and operation principle of the proposed pump are presented and analyzed in detail. The output displacement of the transducer can be obtained by Timoshenko beam theory; on this basis, the flowrate and pressure models of the pump are established based on gap flow theory and extended Bernoulli equation. A prototype of the pump is fabricated and the experimental system is established. The results indicate that the maximum output flowrate is 58.6 mL/min under the frequency of 160 Hz; and the maximum output pressure is 28.7 kPa at the frequency of 180 Hz; and the maximum output power is 6 mW at the backpressure of 12.5 kPa. Besides, the self-priming of the pump is 10.7 kPa; thus, there is no need to fill the pump during working process, which can bring a convenience to the operation of the pump.

Commercial scale jet-loop and trickle bed reactor simulations have been performed for the reductive aminolysis of glucose with dimethylamine (DMA) in combination with an experimental assessment in a (large) lab-scale fed-batch and trickle bed reactor. Based on the intrinsic kinetics of the sequence of homogeneous and heterogeneously catalyzed reaction steps in the reductive aminolysis reaction network, coupled with a different flow pattern and catalyst-to-liquid ratio, significant differences in the yield of N,N,N’,N’-tetramethylethylenediamine (TMEDA) and N,N-dimethylaminoethanol (DMAE) can be expected for both reactor types. Yet, TMEDA is obtained in the highest yields irrespective of the reactor type. In the jet-loop reactor, which is operated overall as a batch reactor, an optimized TMEDA yield up to 57 % is simulated, while a non-negligible yield of DMAE, up to 12 %, is also achieved. The optimized operating conditions are such that they correspond to the high DMA to glucose ratio used during lab-scale fed-batch experimentations, also explaining why the jet-loop product spectrum is almost identical to the latter. This is ensured by the efficient heat-exchange in the jet-loop reactor, which is, in contrast, one of the main bottlenecks for exploitation of the reductive aminolysis in a trickle bed reactor. The second challenge to address in the trickle bed reactor is the limited extent of hydrogen transfer from gas to liquid, which can only be partially overcome by increasing the feed flow rates. While isothermal operation indicates TMEDA yields of about 60 %, non-isothermally simulated TMEDA yields only amount to 40 %. The latter yield loss was also validated experimentally in the trickle bed reactor and confirmed that avoiding degradation reactions remains challenging as it is critical in the reductive aminolysis of glucose.

Although many studies have been done on finding operating conditions of hydrogen-fed fuel cells before, it remains one of the most critical points in determining its parameters in the process. So this paper aims to investigate experimentally the reactant gases flow rate and cell voltage which have a significant impact on the current density of a 3-cell Proton Exchange Membrane fuel cell stack having a 150 cm2 active layer. In this case, to determine the optimum values, Design of Experiment and Response Surface Methodology was applied to the experimental system at low 1.5 V, medium 1.8 V, and high 2.1 V. Then, they were compared with each other. In this context, keeping the hydrogen flow rate low and obtaining high current density is one of the main targets; at low voltage values, it was concluded that the flow rate should be increased due to the reaction rate increases with temperature. In general, the effect of humidification and cell temperature on performance was seen more prominently at 1.8 V. The highest current density values that were 313.66 mA/cm2, 336.75 mA/cm2, and 323.48 mA/cm2, respectively, were reached at flow rates of 1 L/min,1.3 L/min,1.6 L/min.

The present work pertains to experimental studies on the unsteady phenomena in the saturated flow boiling of R407C stemming from the oscillatory refrigerant flow rate with the triangle shape in the horizontal slender annular channel. Attention is mainly paid to examine the effects of different parameters, including mean mass flux, amplitude and period of the oscillating mass flux, saturation temperature of R407C, input thermal flux, and gap size on the boiling diagrams and heat transfer coefficients. Firstly, the steady-state saturated flow boiling is studied as baseline cases. It was concluded that beyond the onset of nucleate boiling region, the mass flux of refrigerant has little influences on the boiling diagrams, which indicates that the fully developed nucleate boiling is the major mechanism. The saturated flow boiling heat transfer coefficient is enhanced as the gap size is reduced. Then, the experiment focuses on the temporal changes in the wall temperature Tw and the heat transfer coefficient hr stemming from oscillatory flow conditions. When the mass flux is oscillated periodically with the triangle shape, the Tw and hr are also oscillated temporally and enhanced with increasing in either the amplitude or frequency of oscillating mass flux. At higher input thermal flux, Tw decreases with the decreasing mass flux, which has opposite tendency compared to the single-phase convection with lower thermal flux. For intermediate exerted thermal flux, there is a negligible oscillation in the Tw and hr for most cases; the amplitudes of oscillations are lower than 0.2 °C. Besides, the significant oscillations in the wall temperature and heat transfer coefficient are noted the case with large period of the oscillating mass flux.

Floodwaters can damage railway tracks by scouring the ballast, and this can, in turn, lead to dire consequences, including train accidents. Therefore, the accurate prediction of water flow through railway ballast is of significant importance to assess and mediate the risk of ballast damage. In particular, ballast permeability is an important parameter because it controls the flow rate through the ballast. Here we develop a model that includes the random permeability effect using the Karhunen-Loeve expansion. We also allow for the free boundary. The computation is carried out using the finite element method, and it is shown that the same elements can compute the Karhunen-Loeve expansion as are used in the numerical solution of the fluid flow. All calculations are carefully validated against either analytic solutions or results from the literature. It is shown that randomness in the permeability increases the outward flow compared to the constant permeability case. It is also shown that allowing for the free boundary changes significantly the flow characteristics.

In this paper, we present a parallelized microfluidic device with improved ladder–tree distributors for the scaled-up production of monodisperse microspheres. Owing to the improved distribution form comprising a ladder distributor in the center and tree distributors around the circumference, the uniformity of the generated droplets for our device was improved by 32.9% and 86.1% compared to those achieved in devices with conventional tree and ladder distributors, respectively, in the final production stage with 32 droplet generators, indicating the potential of our device for further scale-up while maintaining the monodispersity of the generated droplets. The effect of the 2D scale-up of our device was investigated by varying the continuous-phase flow rate and viscosity. Using our device, γ-Al2O3 microspheres with a diameter of 598.1 μm and a CV of 2.58%, as catalyst support for microreactors, were continuously synthesized at high production rate. The nickel-based catalyst prepared from γ-Al2O3 microspheres with large specific surface area and pore volume had an average NiO size of 14 nm, which was 60.9% of that supported on industrial support. The CO methanation conversion of the Ni@Al2O3 microspheres was 1.7 times larger than Ni@Al2O3 industrial support at 140 °C.

A two-phase natural circulation test using SMART integral test loop (SMART–ITL) was conducted to explore thermo-hydraulic phenomena of two-phase natural circulation in the SMART reactor. Specifically, the test examined the natural circulation in the primary loop under a stepwise coolant inventory loss while keeping the core power constant at 5% of the scaled full power. Based on the test results, three flow regimes were observed: single-phase natural circulation (SPNC), two-phase natural circulation (TPNC), and boiler–condenser natural circulation (BCNC). The flow rate remained steady in the SPNC, slightly increased in the TPNC, and dropped abruptly and maintained in the BCNC. Using a natural circulation flow map, the natural circulation characteristic in the SMART–ITL was compared with those in pressurized water reactor simulators. In the SMART–ITL, a BCNC regime appeared instead of siphon condensation and reflux condensation regimes because of the use of once-through steam generators.

Abrasive waterjet (AWJ) has a significant impact on the dimensional accuracy and machining damage of Cf/SiC composites. Thus, this paper aims to investigate the hole characteristics and formation mechanisms of surface damages (e.g. crack damage, tearing damage, delamination damage and burr damage). The results show that the diameters of hole entrance and hole exit varies with the traverse speed, jet pressure, standoff distance, and abrasive flow rate within certain parameters range. The over-machined phenomenon has the greatest impact on the hole dimensional accuracy, which can be effectively enhanced by improving the jet trajectory. The formation mechanisms of various damages are also obtained. These results will guide realizing high-precision and low-damage hole machining for ceramic matrix composite material.

As an innovative thermal energy technology, the spray-type packed bed has advantages of high efficiency and low cost. A liquid distributor is the key component for the spray-type packed bed for scattering heat-transfer liquid drops evenly. In this study, the distribution performance and pressure drop of the perforated plate distributors of different orifice diameters were studied experimentally. The experimental results indicate that orifice diameter has a greater effect on the distribution performance compared to flow rate. With an increase in flow rate, the flow pattern through the distributor changes from the uncovered drop to the covered drop and then to the jet flow. The covered drop pattern shows the best performance with a good distribution and a small pressure drop simultaneously, which is the design and optimization principle of the distributor for a spray-type packed bed.

In aero-engines, mortise-tenon joint structures are often used to connect the blades to the turbine disk. The disadvantages associated with conventional manufacturing techniques mean that a low-cost, high-efficiency, and high-quality nickel-based mortise–tenon joint structure is an urgent requirement in the field of aviation engineering. Electrochemical cutting is a potential machining method for manufacturing these parts, as there is no tool degradation in the cutting process and high-quality surfaces can be obtained. To realize the electrochemical cutting of a mortise-tenon joint structure, a method using a tube electrode with helically distributed jet-flow holes on the side-wall is proposed. During feeding, the tube electrode rotates along its central axis. Flow field simulations show that the rotational speed of the tube electrode determines the direct spraying time of the high-speed electrolyte ejected from the jet-flow holes to the machining area, while the electrolyte pressure determines the flow rate of the electrolyte and the velocity of the electrolyte ejected from the jet-flow holes. The machining results using the proposed method are verified experimentally, and the machining parameters are optimized. Finally, mortise and tenon samples are successfully machined using 20 mm thick Inconel 718 alloy with a feeding rate of 5 μm/s.

Waves are an alternative energy source that can be used for electricity generation. Wave Energy Converter (WEC) system in perforated breakwater is potentially applicable WEC system for coastal area. The magnitude of wave energy generated is determined by the volume of sea water inside the perforated breakwater. This volumetric flow rate is calculated using the flow velocity at perforated holes on the structure slope. Therefore, this research aims to study the velocity magnitude by analyzing the interrelation among wave steepness, wave run-up and relative velocity. The method used consists of applying numeric 3D flow model in the perforated structure of the breakwater with the variation of wave height, wave period and structure slope. The result shows that, the steeper the structure, the bigger is the relative run up (Ru/H). The higher the relative run up, the higher are the relative run-up velocities (V/Vru). As the velocity increase, the volumetric flow rate inside perforated breakwater will be higher, which leads to higher wave energy. Hence, it can be concluded that the higher the velocities (V/Vru), the higher is the wave energy generated.

In order to improve the surface properties of Al alloy, TiAlN ceramic coatings composed of alternating TiAl layer and TiAlN layer were prepared by FCVA technology. The influence of N2 flow rate on the microstructure, mechanical and tribological performance of multilayer TiAlN coatings was systematically studied. The results showed that the coating exhibited a multi-phase structure with mainly fcc TiAlN phase and a small amount of TiN and AlN phases. As the N2 flow rate increased from 50 to 80 sccm, the preferred orientation varied from TiAlN (220) plane to (111) plane due to lattice distortion. The hardness and elastic modulus of multilayer TiAlN coatings first increased and then decreased with the increase of N2 flow rate. However, the average friction coefficient and wear rate of multilayer TiAlN coatings exhibited an opposite trend to the hardness, H/E and H3/E2 values. The adhesion gradually decreased from 27.3 to 14.9 N due to the continuous increase of the compressive stress in the coating with the increase of N2 flow rate. When the N2 flow rate was 70 sccm, the coating showed the lowest average friction coefficient and wear rate, which was attributed to the highest hardness and the best resistance to cracking and plastic deformation.

Monodispersed microbubbles and microdroplets are widely used as reaction carriers in microfluidics. In this study, the generation processes of bubbles and droplets in a step-emulsification microfluidic device are compared to show the similarities and differences in the emulsification process. By changing the placement of the microdevice, the effects of buoyancy and gravity on the generation of bubbles and droplets are introduced, and the feedback mechanism of the bubble layer and the effect of droplet accumulation on the emulsification process are clarified. Finally, based on the analysis of the difference of the pinch-off of the dispersed phase between the bubble and the droplet in this configuration, the Plateau-Rayleigh instability processes for the formation of bubble and droplet are revealed by using a high-speed camera system, and the reasons for the difference of the operating ranges of the gas flow rate and liquid flow rate in the dripping flow regime are explained.

In order to improve the design of PSA system for fuel cell hydrogen production, a non-isothermal model of eight-bed PSA hydrogen process with five-component (H2/N2/CH4/CO/CO2=74.59/0.01/4.2/2.5/18.7 % (vol)) four-stage pressure equalization was developed in this article. The model adopts a composite adsorption bed of activated carbon and zeolite 5A. In this article, pressure variation, temperature field and separation performance are stimulated, and also effect of providing purge (PP) differential pressure and the ratio of activated carbon to zeolite 5A on separation performance in the process of producing industrial hydrogen (CO content in hydrogen is 10 µml·ml-1) and fuel cell hydrogen (CO content is 0.2 µml·ml-1) are compared. The results show that Run 3, when the CO content in hydrogen is 10 µml·ml-1, the hydrogen recovery is 89.8%, and the average flow rate of feed gas is 0.529 mol·s-1; When the CO content in hydrogen is 0.2 µml·ml-1, the hydrogen recovery is 85.2%, and the average flow rate of feed gas is 0.43 mol·s-1. With the increase of PP differential pressure, hydrogen recovery first increases and then decreases, reaching the maximum when PP differential pressure is 0.263 MPa; With the decrease of the ratio of activated carbon to zeolite 5A, the hydrogen recovery increases gradually. When the CO content in hydrogen is 0.2 µml·ml-1 the hydrogen recovery increases more obviously, from 83.96% to 86.37%, until the ratio of activated carbon to zeolite 5A decreases to 1. At the end of PP step, no large amount of CO2 in gas or solid phase enters the zeolite 5A adsorption bed, while when the CO content in hydrogen is 10 µml·ml-1, and the ratio of carbon to zeolite 5A is less than 1.4, more CO2 will enter the zeolite 5A bed.

In this work, a six-bed pressure swing adsorption (PSA) process was investigated to produce medical oxygen from air, which uses the combination of six-way rotating distribution valve and PSA and has the main advantage of effectively saving space compared to the traditional two-bed or four-bed PSA process and can obtain greater productivity. The mathematical model of adsorption beds was developed based on the separation mechanism and the interaction among different equipment. Moreover, a pilot-scale device has been constructed to verify the accuracy of mathematical model by experiment. The oxygen product conformed to the medical standard (> 93vol%) with a recovery of over 57%. Some related parameters were also discussed in detail, such as step time, ratio of length to the diameter, flow rate of product.

In order to remove N2 from low quality natural gas, a mathematical model has been established by Aspen adsorption, using the CH4-selective sorbent silicalite-1 pellets. The dynamic adsorption isotherm was first simulated by breakthrough simulation of a CH4/N2 mixture at different adsorption pressures and feed flow rates based on breakthrough experiments. The resulting simulated CH4 dynamic adsorption amounts were very close to the experimental data at three different adsorption pressures (100, 200, and 300 kPa). Moreover, a single-bed, three-step pressure swing adsorption (PSA) experiment was performed, and the results were in good agreement with the simulated data, further corroborating the accuracy of the gas dynamic adsorption isotherm obtained by the simulation method. Finally, based on the simulated dynamic adsorption isotherm of CH4 and N2, a four-bed, eight-step PSA process has been designed, which enriched 75% (vol) CH4 and 80% (vol) CH4 to 95% (vol) and 99% (vol), and provided 99% (vol) recovery.

A population balance model based on the multiple-size-group (MUSIG) approach was developed to investigate the multi-size bubbly flow in a slab continuous casting mold. Firstly, effects of the bubble size group number, the initial bubble diameter, and the calibration factors for breakage and coalescence in MUSIG model on the multi-size bubbly flow were discussed in detail. Then, effects of both the water and gas flow rates on the bubble size distribution, the gas volume fraction, and the turbulence intensity of liquid phase were investigated. A relatively high water flow rate intensified the turbulence in the upper recirculation region of the mold and also eased the turbulence near the submerged entry nozzle (SEN) by promoting the breakage and horizontal transportation of bubbles; thus, an appropriate water flow rate is essential for better stability with a lower turbulence intensity near the SEN to prevent exposed slag eye formation and slag entrainment. With increase in the gas flow rate, the mean bubble diameter and gas volume fraction increased significantly, the turbulence near the SEN was also significantly increased, but this effect was much weaker in the region away from the SEN.

In the wake of novel technologies such as artificial intelligence (AI), machine learning (ML), internet of things (IoT) to cater to a wide spectrum of industrial applications, eventually the demand for ultra-fast processors/data centers has been on rise. In further, these novel technologies inherently result a drastic increase of power density in data centers, where air-cooling is the popularly utilized method to handle the thermal loads. The present research study involves in numerically investigating the effectiveness of single-phase immersion cooling process, an alternate solution to air-cooling method, on governing the heat load of the densely packed servers. A dielectric fluid, EC-110 is chosen for the simulation on a suitably developed CAD model of a server, of rated design power of 150 W. Four different processing unit powers of value 65 W, 90 W, 115 W and 150 W are supplied as input load to the cooling fluid in the simulation work. The condition of inlet temperature of the dielectric fluid is varied from 25℃ to 40℃ and the flow rates are varied from 0.5 lpm to 2.5 lpm. The resulting chip junction temperature in the server is estimated for different flow rates and inlet temperature conditions of the fluid, under immersion cooling technique. It proves that this cooling technique has the potential to ensure safe operation temperature of the server at low flow rates of the fluid with high inlet temperature, at high power densities.

Heterogeneous nitrogen-doped carbon-incarcerated copper nanoparticle catalysts have been developed. The catalysts showed high activity for oxidation of alcohols to the corresponding aldehydes, including aliphatic substrates, in the presence of N-oxyl radical co-catalyst. The catalysts were applied to a flow reaction and a long lifetime (up to >5 days) was achieved by tuning the flow rate of oxygen gas. By applying the catalysts, two-step, sequential, continuous-flow synthesis of unsaturated carbonyl compounds was demonstrated, and the product was further converted into a bioactive heterocycle without isolation of intermediates.

Platelets (PLTs) are responsible for stopping the bleeding. They are small cell fragments produced from megakaryocytes (MKs) in the bone marrow. Low platelet count is a significant health problem for a patient. PLTs can usually be stored for up to 5 days prior to transfusion. Instantaneous production of PLTs from isolated and stored MKs is crucial for the patient’s health. Thanks to microfluidic platforms, PLTs can be produced instantaneously from MKs. Herein, we have computationally studied fluid dynamics in the microchannels with slit structures and different inlet geometries. Analysis of the flow dynamics was performed by the commercial analysis software. The effects of flow rates and the angle between the inlet channels on the MKs trapping were investigated. The optimization of the angle between inlet channels and flow rates of main and pressure flows was done with Response Surface Methodology (RSM) by counting the trapped MKs. The optimum conditions lead to the percentage of trapped MKs were 100 with a relative deviation of

Nanofluid offers remarkable thermal and optical properties favourable for direct solar absorption. The nanofluids prepared by the conventional two-step synthesis method have low colloidal stability, while that synthesized through the one-step method is costly. Hence the nanofluid synthesized using an economical one-step method has great significance. In the present study, a highly stable C-dot/water nanofluid was synthesized using an economical one-pot synthesis method. The optical characterisation, corrosion analysis and cost estimation of the nanofluid were conducted. The influence of C-dot/water nanofluid on the performance of direct absorption solar collector was analyzed. The direct absorption parabolic solar collector employed with C-dot/water nanofluid yielded a maximum thermal efficiency of 73.41% at Reynolds number of 2952, while that for water was 15.79%. Thermodynamic analysis of the system and cost estimation of the nanofluid was performed to establish its commercial suitability in various solar thermal devices. The maximum exergy destruction was found to be 924.3 W and was more or less constant at all flow rates. The main highlight of the new C-dot/water nanofluid is its significantly high colloidal stability and was found to be stable for more than six months. The corrosion rate of the new C-dot/water nanofluid was obtained as 0.094 mm/year, while that for the base fluid was 0.372 mm/year. With superior optical performance, corrosion resistance, and low production cost, the C-dot nanofluid has the potential to be a prospective working fluid in various direct absorption solar thermal systems.

Mixed ionic-electronic conducting (MIEC) ceramic membranes present an attractive alternative oxygen separation technology to cryogenic distillation and pressure swing adsorption. The model of Tan and Li has been extensively adopted in the oxygen permeation modeling through MIEC hollow fiber membranes. To evaluate the performance of the La0.8Ca0.2Fe0.95O3-δ-Ag hollow fiber membrane module here, the oxygen permeation is simulated numerically by feeding the air into the shell while passing sweep gas or creating a vacuum in the permeate side under different configuration. The simulation results show that countercurrent flow configuration is preferable in the sweep gas mode given its capability to separate oxygen completely under moderate flow rate of feed air. For vacuum mode, both cocurrent and countercurrent flow configurations result in a similar oxygen permeation performance, which lies between the sweep gas mode with countercurrent and cocurrent flow configurations. When abundant oxygen presence is created in the feed side by elevating its flow rate, the sweep gas mode loses its oxygen separation efficiency and is thus outperformed by the vacuum mode.

A new optimization strategy based on the mixed quantitative struc- ture–retention relationship (QSRR) model is proposed for improving the RP- HPLC separation of aripiprazole and its impurities (IMP A-E). Firstly, experi- mental parameters (EPs), namely mobile phase composition and flow rate, were varied according to Box–Behnken design and thereafter, an artificial neural network (ANN) as a QSRR model was built correlating EPs and sel- ected molecular descriptors (ovality, torsion energy and non-1,4-van der Waals energy) with the log-transformed retention times of the analytes. Values of the root mean square error (RMSE) were used for an estimation of the quality of the ANNs (0.0227, 0.0191 and 0.0230 for the training, verification and test set, respectively). The separations of critical peak pairs on chromatogram (IMP A- B and IMP D-C) were optimized using ANNs for which the EPs served as inputs and the log-transformed separation criteria s as the outputs. They were validated by application of leave-one-out cross-validation (RMSE values 0.065 and 0.056, respectively). The obtained ANNs were used for plotting response surfaces upon which the analyses chromatographic conditions resulting in optimal analytes retention behaviour and the optimal values of the separation criteria s were defined. The optimal conditions were 54 % of methanol at the beginning and 79 % of methanol at the end of gradient elution programme with a mobile phase flow rate of 460 μL min-1 Нова оптимизациона стратегија заснована на грађењу мешовитих модела за кван- тификовање односа структуре и ретенционог понашања (QSRR) предложена је за уна- пређење RP-HPLC раздвајања арипипразола и његових нечистоћа (IMP А-Е). Експери- ментални параметри (EP), састав мобилне фазе и брзина протока, варирани су најпре у складу са Box–Behnken дизајном, а затим је награђена вештачка неуронска мрежа као QSRR модел који повезује ЕP и одабране молекуларне дескрипторе (овалност, торзиона енергија и не-1,4-ван дер Валсова енергија) са логаритамски трансформисаним ретен- ционим временом аналита. Вредности средње квадратне грешке (RMSE) коришћене су за процену квалитета мреже (0,0227, 0,0191 и 0,0230 за тренинг, верификацију и тест сет, редом). Раздвајање критичних парова пикова на хроматограму (IMP А-B и IMP D-C) оптимизовано је коришћењем мрежа за које су ЕP послужили као улази, а логаритамски трансформисани критеријуми сепарације s као излази. Ове мреже су валидиране при- меном унакрсне валидације изостанка (RMSE вредности, редом, 0,065 и 0,056). На основу награђених мрежа, конструисани су дијаграми површина одговора чијом ана- лизом су дефинисани услови при којима се постиже оптимална ретенција аналита, односно вредности критеријума сепарације s, а који су подразумевали 54 % метанола на почетку и 79 % на крају програма градијентног елуирања са брзином протока мобилне фазе од 460 mL min -1