This study investigates the instabilities of Poiseuille-Rayleigh-B & eacute;nard convection of Jeffreys fluids in an inclined fluid-porous system subjected to a gravity-driven plane Poiseuille flow. In light of the linear stability analysis and the spectral collocation method, we find that some neutral curves of oscillatory instabilities are bimodal and the dominant mode shifts between the long-wave and short-wave branches with the variation of the depth ratio. For some depth ratios, the transition of the dominant mode will shift from short-wave branch to the long-wave branch with the increase of angles. In addition, the neutral curves appear trimodal at the depth ratio d = 0.2 and the inclination angle 4 degrees <= B <= 6 degrees. For the local Reynolds number 1.96 <= Ref <= 4.1, the dominant instability is the short-wave branch when 0 degrees <= B <= 5 degrees. When Ref and B are not in these intervals, it is always dominated by the long-wave branch. The strain retardation time and the stress relaxation time are found to have opposite influences on the instabilities. Moreover, the greater angles delay onset of the oscillatory instability and the short-wave modes vanish near B = 60 degrees and 70 degrees for both viscoelastic times. In a slightly inclined system, the Froude number has the same effect on long-wave and short-wave branches. On the contrary, the Froude number has an opposite effect on two branches with a greater inclination angle: stability increases for long waves but decreases for short waves as the Froude number increases.

A review concerning the development and performance of water-based fixed fire fighting systems (WFFFS) in road tunnels is presented. Conventional WFFFS types, including automatic sprinkler system, deluge system, and water mist system, have gained recognition as valuable supplements to ventilation systems in combating tunnel fires. Full-scale projects through international collaborations have been instrumental in formulating guidelines for facilitating the installation and optimal functioning of WFFFS in road tunnels. Extensive small-scale experiments, along with numerical investigations documented in literature, greatly contribute to complete exploring the performance of WFFFS. Owing to these efforts, advantages of WFFFS, such as reducing fire size, cooling hot smoke, and limiting smoke spread are well affirmed, and the shed mechanisms behind these fire fighting effects now have been well understood. However, WFFFS also present challenges from aspects such as disrupting smoke stratification and augmenting the production of toxic combustion products. From the review, further development of WFFFS in road tunnels relies on establishing methods of specifying design fires that accounts for suppression effects of WFFFS, as well as gaining deeper insights into the cooperative and interactive mechanisms between WFFFS and mechanical ventilation systems.

Shear significantly influences turbulence in the energy-containing range of shear-dominated flows, and the longitudinal structure functions do not have a universal form as they do in homogeneous isotropic turbulence. Despite this, the relative scaling of structure functions exhibits universal sub-Gaussian behaviour in shear-dominated flows, in particular for turbulent boundary layers, channels and Taylor–Couette flows. Our investigation of a turbulent vertical buoyancy layer at $Pr = 0.71$ using direct numerical simulation shows this universality even in moderate-Reynolds-number buoyancy-driven but shear-dominated boundary layers. It is demonstrated that the universality is related to the energy density of the eddies, which attains a hierarchical equilibrium in the energy-containing range of shear-dominated turbulence. We conjecture that the universal sub-Gaussian behaviour of the energy density of the energy-containing range, which was considered to be non-trivial in prior studies, is related to the universal anomalous scaling exponents of the inertial subrange turbulence. Based on this conjecture, we propose a hypothesis that relates large-scale eddies and the intermittent dissipation field in shear-dominated turbulence, highlighting a relationship between large and small scales. A phenomenological model is also developed to predict the scaling, which is verified using data from a turbulent boundary layer, half-channel and vertical buoyancy layer at friction Reynolds numbers spanning four orders of magnitude. Excellent agreement is observed.

This paper investigates the turbulent structure of stratified open-channel flow subjected to a radiative volumetric heat source modelled by the Beer-Lambert law, for Prandtl numbers (Pr) varying from 0.07 to Pr=7. Direct Numerical Simulation (DNS) was employed to model the open-channel flow. To overcome the increased computational resources required to resolve the thermal fields when Pr>1, a multi-resolution method using quadratic interpolation was employed to resolve the temperature and momentum fields on different spatial and temporal resolutions. This scheme was implemented in an in-house computational fluid dynamics (CFD) code. To further reduce the computation cost, the DNS of Pr=2.2 and 7 fluids were initialised using the outputs of minimal channel simulations. The simulations were conducted for Pr=0.07, 0.22, 0.71, 2.2, and 7 under neutral (lambda=0), near-neutral (lambda=0.1), and stable (lambda=0.5) thermal stratification. The results demonstrate that Pr significantly affects the flow structure and turbulence characteristics of stratified flows, particularly near the free surface. This includes higher velocity, temperature gradient, and buoyancy effects for Pr=7 compared to lower Pr values. For stratified Pr=7 flow, examination of the Reynolds stresses and turbulent heat flux reveals significant damping of turbulence near the surface, with flow displaying near-laminar behaviour.

Wake effects in the Anholt offshore wind farm have been investigated using both operational data and a Large Eddy Simulation (LES) model of a group of five turbines within the wind farm. Analysis of operational data showed that the variations of main shaft speeds of the downstream turbines were almost six times those of the upstream turbine at near-rated operation. The aim of the LES was to study the impact of atmospheric stability on the wind turbine array performance and compare this with the field data. An LES precursor method was used to model the near-neutral and unstably stratified atmospheric boundary layers that represent typical conditions in winter and summer, respectively, and the turbines in wind farm model were simulated using an actuator line method. It was found that LES with the actuator line method and generic turbine design data can generate a reasonable mean power generation trend for the Anholt wind farm under near-neutral and unstable conditions. The maximum difference in the mean power output between the LES and averaged operational data was approximately 20%.

_ Sail aerodynamics has been a researched topic since the start of competitive racing, such as the America's Cup and Whitbread Round the World Race. Recently, with restrictions on the use of wind tunnels and towing tanks in the America's Cup, there has been an increase in computational fluid dynamics (CFD) to study the aerodynamics and performance of yacht sails. These studies typically model steady-state conditions, but in contrast, this paper demonstrates the implementation of dynamic wind conditions in Reynolds Averaged Navier Stokes CFD and the effects it has on the aerodynamics of a yacht sailing upwind. A user-defined function was used to give a velocity profile and vary the wind direction with time to simulate a gust condition with a period of 10 s. In addition to the gusting model, steady models were run for the apparent wind angles (AWA) experienced in the gust to see if the dynamic wind conditions could be modelled by a series of steady state snapshots. The results show minor differences in the streamlines of two models and the pressure distributions show no significant difference. There is only a slight difference in sail forces at high AWAs. Although further research would be required this information could be valuable to anyone using CFD to test sail designs. Keywords CFD; transient; steady; upwind; UDF; RANS

Direct numerical simulations of stratified open channel flows subject to a varying surface heat flux are performed. The influence of the diurnal heating time on the spatial and temporal variation of mixing in the flow and the characteristics of the mean flow state are examined. The control parameters are the bulk stability parameter λ_B , defined through the ratio of the channel height δ and a bulk Obukhov length scale ℒ_B , and the diurnal time scale t̂ , defined as the ratio of the heating time to an eddy turnover time. The Prandtl number Pr and Reynolds number Re_τ have values of 1 and 400. Simulations are performed over t̂ = 1 to 24 and λ_B = 0.6 to 26. Two key flow features are used to classify the flow regimes observed, namely the laminar layer depth (LLD) and stratified layer depth (SLD) where the LLD is defined as the depth from the free surface when the buoyancy Reynolds number Re_B≈ 7 and the SLD is the depth from the free surface when the turbulent Froude number Fr ≈ 1 . This study attempts to characterise how these length scales vary across the diel cycle. The LLD is a viscous length scale and a regime map of a viscous parameter, the bulk Obhukov Reynolds number Re_ℒ , and t̂ is presented to classify the LLD behaviour. A regime map of λ _B and t̂ is presented to classify the behaviour of the SLD. Three classifications for each layer depth behaviour within a diel cycle form the basis of the regime maps for this paper: a neutral flow where the LLD or SLD does not exist (denoted by NL and NS), a stratified flow where the LLD or SLD are diurnally varying (denoted as DL and DS) and a persistent layer of the LLD or SLD (denoted as PL and PS). The transition between the NL to DL is t̂∝ Re_ℒ^4.5 , DL to PL is t̂∝ Re_ℒ^- 0.5 , NS to DS is t̂∝λ _B^0 and DS to PS is t̂∝λ _B^1 . The regime maps may be used as a predictive tool to determine when suppressed mixing regimes occur in rivers. At each flow depth, the flow sweeps though a range of mixing states across the diel cycle. The local mixing efficiency are briefly assessed and found to scale well with the instantaneous Fr number according to the regimes proposed by Garanaik and Venayagamoorthy (J. Fluid Mech., vol. 867, 2019, pp. 323-333). This paper reports on direct numerical simulations of stratified open channel flows subject to a varying surface heat flux. The results have found that:

In a windy environment, the flame tilt angle and base drag length are two parameters used to describe the flame geometry of liquid fuel fires. This work describes a numerical investigation using the FDS code to model three typical fire sizes in a longitudinally ventilated tunnel and a windy open space. The CFD modelling methodology was validated against experimental data from the literature. According to the findings, when the ventilation is weak, a backlayering flow occurs in the tunnel and causes a two-layer flow pattern upstream, which reduces the flow space and accelerates the ventilation. Consequently, the flame base drag length tends to remain constant, and the flame tilt angle becomes very difficult to predict. Unlike open fires, when the upstream is free of backlayering, the tunnel area and the heat release rate have been proven to be two additional factors that influence the flame geometry. The confined space causes more concentrated momentum of wind and speeds up the fuel consumption upstream of the dragging flame base. The former results in a more significant flame tilt, and the latter shortens the stay of fuel vapor on the floor. By introducing a dimensionless heat release rate and dimensionless tunnel cross-sectional area, a series of new prediction models covering all the governing parameters were derived, which have successfully estimated the longitudinal distance from the fire center to the maximum ceiling temperature in the tunnel.

Side-wall sprinkler systems used for fire suppression in road tunnels have been found to redistribute the longi-tudinal air flow within the tunnel, with the most notable feature being an acceleration of the airflow near the ceiling. The present work is driven by the motivation of utilizing the acceleration effect to block the upstream smoke propagation in cases when the ventilation rate is sub-critical. A nozzle prototype that has practical ap-plications in road tunnels was selected for a side-wall water spray system in a 300 m road tunnel, and a total of 60 cases were modelled using the CFD code FDS (version of 6.7.3) to investigate its smoke blocking effect. Experiments and correlations in the literature have validated the accuracy of the FDS method. The confinement velocity reduction rate, which is determined as the reduction of the ventilation rate required to stop the smoke front point at a given position, was found to be positively affected by both the water spray rate and the venti-lation rate and was successfully correlated using dimensionless momentums of these two factors. By comparing the confinement velocity reduction rates with the increases in the ceiling velocity in the absence of tunnel fire, utilization efficiencies of the spray-induced accelerations were further obtained.

This study investigates the coherence of turbulent fluctuations in a turbulent vertical natural convection boundary layer immersed in a stably stratified medium (turbulent buoyancy layer). A turbulent buoyancy layer of a fluid having a Prandtl number of $0.71$ at a Reynolds number of $800$ is numerically simulated using direct numerical simulation. The two-point correlations reveal that the streamwise velocity fluctuations are coherent over large streamwise distances, with the length scale of the streamwise coherence being greater than the boundary layer thickness. This is due to large-scale motions (LSMs), similar to the LSMs observed in canonical wall-bounded turbulence despite the stark differences in flow dynamics. Both high-speed (positive) and low-speed (negative) streamwise velocity fluctuations form LSMs, with their streamwise length scales increasing with increasing wall-normal distance. High-speed LSMs are composed of upwash flow with high temperatures, while low-speed LSMs are composed of downwash flow with low temperatures. Both high-speed and low-speed LSMs meander appreciably in the streamwise direction, with the degree of meandering being correlated with the sign of the spanwise velocity fluctuations. The LSMs exhibit coherence across significant wall-normal distances and contribute significantly to the turbulence production in the outer layer. Examining the one-dimensional energy spectra of the turbulent buoyancy layer shows that the LSMs are the dominant energy-containing motions, implying that the length scale of the energy-containing range is of the order of boundary layer thickness. Notably, wall-normal velocity, spanwise velocity and buoyancy fluctuations do not form LSMs with streamwise length scales comparable to streamwise velocity fluctuations.

This paper presents a simplified model of tunnel smoke stratification, which can be used to study the effect of sprinkler water spray on stratified stability. The stratified flow downstream of a tunnel fire is represented by a two-layer thermal flow, with the hot upper layer representing the smoke and the cold lower layer representing the ventilation air. The dragging effects of sprays from sprinkler nozzles with various K-factors were numerically studied using FDS code (version 6.7.3). Experiment data reported in the literature successfully validated the accuracy of the CFD method. Spray-induced smoke movements, including the smoke logging distance and the horizontal movement, were determined by releasing a series of massless tracers from the smoke layer inlet. The findings of the work indicate that the dragging effect of a water spray is droplet-size-dependent. The smaller the droplets, the larger the dragging force for a given volumetric spray rate. By applying dimensional analysis and introducing the volume mean diameter into the governing parameters, some correlation models that account for the droplet size were proposed to predict the spray-induced smoke movements and smoke layer heat loss. The new correlation models help complete previous work by the authors and are of practical use for tunnel fire safety design where water sprays are included alongside a longitudinal ventilation system.

This paper reviews and evaluates the application of computational fluid dynamics (CFD) to predict the impact of tropical cyclone wind speeds. The literature on this topic is sparse; however, in the past decade an increasing number of studies have been carried out. Historical tropical cyclone events have motivated several studies, and researchers have adopted a methodology of coupling numerical weather prediction models (NWPs) with CFD models. This computational approach is relatively new and allows researchers to predict the impacts of tropical cyclone wind speeds more accurately. However, upon review it is clear the approach taken by researchers in this field do not adhere to the best practice guidelines established by wind engineers. The majority of the researchers have not validated their numerical results and even fewer have ensured the CFD model has a homogeneous atmospheric boundary layer (ABL), both of which are important to the accuracy of the results. Based on the findings of this review, CFD can be coupled with NWP models to model a higher resolution tropical cyclone event. However, following the guidelines of wind engineers can improve the quality of such investigations.

Expansion devices are a key component that affects the performance of the Organic Rankine Cycles (ORCs), and its improvement has been identified as one of the most important parts of future studies in ORCs. A theoretical model of a modified revolving vane expander was developed in this study. The model was to investigate the inherent physical processes and the effect of losses on the mechanism. The developed model was validated with the experimental results. The experiments were carried out using an improved experimental rig with direct coupling of the dynamometer and expander to eliminate any externally exerted misalignments. The results of the theoretical validation showed that the torque is generally slightly underpredicted, but the trends are very closely predicted. Moreover, a little underprediction was observed in the theoretical study of mass flow rate but still within the uncertainty of the experiments. The experimental results showed the prototype could generate up to almost 0.8 N.m and rotated up to 1700 rpm at a suction pressure of 3 bar(g). Moreover, the prototype demonstrated up to 55% volumetric efficiency and 11% isentmpic efficiency at 3 bar(g) of the suction pressure.

Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Reinforcement Learning for Vapor Compression Cycle Control 38 Pages Posted: 8 Jul 2022 See all articles by Tech Logg DingTech Logg Dingaffiliation not provided to SSRNStuart NorrisUniversity of AucklandAlison SubiantoroUniversity of Auckland Abstract Reinforcement Learning (RL) was applied to control the compressor speed and valve opening percentage of a vapor compression cycle (VCC). The controller must track the chamber temperature and superheat setpoints while the VCC was subjected to varying cooling loads and ambient temperatures. This study developed a dynamic air conditioning simulation model to train the RL agent where the VCC cycle was used to cool an air chamber. The validation against experimental data showed that the model matched the true dynamics with an average MAPE of 2.54%. Four RL controllers were trained, including the full-state (FS) or economic (ES) observations and either the Feed-forward Neural Network (FFNN) or Recurrent Neural Network (RNN) control policy representation using the Soft Actor-Critic Algorithm. The controllers were tested and compared to a PID controller using their rise times, settling times, overshoots, steady-state errors, and power consumption. The study found that the overall performance increased from the PID, ESFFN, FSFFNN, ESRNN to FSRNN. All the RL controllers had better chamber temperature tracking performances and lower power consumption than the PID controller. However, they had higher superheat steady-state errors despite having shorter settling times and overshoots. The RL agents were significantly more robust to sensor noise than the PID controller, and the FS controllers were more stable than ES controllers. Besides that, The ES controllers were more robust than the FS controllers, but the RL agents were less robust than the PID controller. Lastly, the FFNN and RNN did not significantly influence the learned control policies. Keywords: artificial intelligence, Reinforcement learning, Vapor Compression Cycle, Refrigeration, Air conditioning, control Suggested Citation: Suggested Citation Ding, Tech Logg and Norris, Stuart and Subiantoro, Alison, Reinforcement Learning for Vapor Compression Cycle Control. Available at SSRN: https://ssrn.com/abstract=4157083 Tech Logg Ding (Contact Author) affiliation not provided to SSRN ( email ) No Address Available Stuart Norris University of Auckland ( email ) Private Bag 92019Auckland Mail CentreAuckland, 1142New Zealand Alison Subiantoro University of Auckland ( email ) Private Bag 92019Auckland Mail CentreAuckland, 1142New Zealand Download This Paper Open PDF in Browser Do you have a job opening that you would like to promote on SSRN? Place Job Opening Paper statistics Downloads 0 Abstract Views 1 PlumX Metrics Related eJournals Information Systems eJournal Follow Information Systems eJournal Subscribe to this fee journal for more curated articles on this topic FOLLOWERS 223 PAPERS 4,640 Feedback Feedback to SSRN Feedback (required) Email (required) Submit If you need immediate assistance, call 877-SSRNHelp (877 777 6435) in the United States, or +1 212 448 2500 outside of the United States, 8:30AM to 6:00PM U.S. Eastern, Monday - Friday. Submit a Paper Section 508 Text Only Pages SSRN Quick Links SSRN Solutions Research Paper Series Conference Papers Partners in Publishing Jobs & Announcements Newsletter Sign Up SSRN Rankings Top Papers Top Authors Top Organizations About SSRN SSRN Objectives Network Directors Presidential Letter Announcements Contact us FAQs Copyright Terms and Conditions Privacy Policy We use cookies to help provide and enhance our service and tailor content. To learn more, visit Cookie Settings. This page was processed by aws-apollo4 in 0.177 seconds

In the presence of a water spray system, the fire-induced pressure losses in a road tunnel can change, but this effect is currently not considered in the design of longitudinally ventilated tunnels. In this study, theoretical analyses are combined with FDS modelling to investigate the pressure losses due to flow acceleration, stack effect and friction in tunnels that use a water mist or a deluge system. Analytical models for pressure loss calculations have been derived, which indicate the evaporation rate of the water sprays, the spray-induced flow resistances and the decay coefficient of the mean sectional temperature rise are the three key parameters required for the model. According to dimensional analyses, the spray-induced resistance is a function of the momentum ratio between the water spray and the air flow, and the evaporation rate is mainly determined by the fire size and the total mass flow rate of water sprays. Corresponding empirical correlations are proposed. Comparisons between analytical results and FDS simulations concluded the proposed correlations are reliable and can give accurate predictions.

A water mist system can reduce the tunnel critical velocity since it cools the smoke and reduces the buoyancy force. To study this, a theoretical model was derived using the theory of the critical Froude number and a global energy balance concept. The evaporation rate of the mists was found to be the dominating factor affecting the ratio between the critical velocities with and without the suppression. To test this, the CFD code FDS 6.7.3 was used to model a 60 m long tunnel with a water mist operating downstream to cool the smoke. Nine fire sizes in the range of 10 483 kW to 42 988 kW were modelled. With water mist operation, the theoretical model can accurately predict the reduction in critical velocity for small fire sizes. By the addition of a calibration factor to account for the change in the critical Froude number, the developed model can also provide reliable predictions for large fires whose critical velocities are independent of the heat release rate. The findings prove that the critical Froude number theory can be applied even in the presence of a water mist system and reveal a fact that only when the mist cools the rising hot plume and backlayering flow can the critical velocity be reduced.

The use of an expander is a promising technique to improve the performance of vapor compression refrigeration systems, but its use in a system with HFC refrigerants has not been extensively explored. This paper studies the performance of a four-intersecting-vane expander prototype and the effect of different expander location ar-rangements to the overall system performance. A customized R134a vapor compression test rig was constructed for the experiments. The rig's coefficient of performance was determined when using either a throttling valve or an expander for the same operating conditions. The relationships between expander placement arrangements, expander rotational speed, evaporating load and condensing load on the performance of the expander and the thermodynamic behaviors of the system were also investigated. Transient behaviours of the system are discussed too. The experimental results show that the expander could improve the performance of the system by 6.4% in comparison with a system with a throttling valve. The maximum expansion power efficiency was 34.9% at 500 rpm. Of the four expander arrangements, that in which the expander is located right after the condenser exhibited the highest performance. The results show the potential benefits of using a four-intersecting-vane rotary expander for energy recovery of a vapor compression refrigeration system, and the importance of the correct placement of the expander in the system.

In this paper, a four-intersecting-vane-rotary expander prototype is presented, and the performance was measured experimentally in dynamic and static conditions. The mechanism is different from conventional rotary vane machines because it has a non-circular stator with radius of between 35.7 and 40.7 mm. The prototype was designed to be eventually implemented in a refrigeration system. However, for testing compressed air was used as the working fluid. The expander was tested at rotational speeds of up to 1750 rpm, suction pressures up to 5 bar(g) and a discharge pressure of 0 bar(g). The effect of different operating parameters and lubricating oil grades on the expander was experimentally studied under dynamic conditions. Then, the internal leakage characteristics of the prototype was analyzed under static conditions. The maximum volumetric and isentropic efficiencies measured were 31.2 % and 45.6 %, respectively. Generally, higher speed and a more viscous oil improved the volumetric efficiency. The expander's isentropic efficiency was significantly affected by the suction pressure, viscosity of oil and rotational speed. The static leakage test showed that the prototype's housing was properly sealed. The average contributions of internal leakages through the radial clearance, vane tips, rotor slot and end face gaps were 37.2 %, 33.3 %, 16.2 % and 14.1 %, respectively. It was observed that the internal leakage paths were generally independent of each other. (c) 2021 Elsevier Ltd. All rights reserved.

Velocity Prediction Programs (VPPs) are commonly used to help predict and compare the performance of different sail designs. A VPP requires an aerodynamic input force matrix which can be computationally expensive to calculate, limiting its application in industrial sail design projects. The use of multi-fidelity kriging surrogate models has previously been presented by the authors to reduce this cost, with high-fidelity data for a new sail being modelled and the low-fidelity data provided by data from existing, but different, sail designs. The difference in fidelity is not due to the simulation method used to obtain the data, but instead how similar the sail’s geometry is to the new sail design. An important consideration for the construction of these models is the choice of low-fidelity data points, which provide information about the trend of the model curve between the high-fidelity data. A method is required to select the best existing sail design to use for the low-fidelity data when constructing a multi-fidelity model. The suitability of an existing sail design as a low fidelity model could be evaluated based on the similarity of its geometric parameters with the new sail. It is shown here that for upwind jib sails, the similarity of the broadseam between the two sails best indicates the ability of a design to be used as low-fidelity data for a lift coefficient surrogate model. The lift coefficient surrogate model error predicted by the regression is shown to be close to 1% of the lift coefficient surrogate error for most points. Larger discrepancies are observed for a drag coefficient surrogate error regression.