EFFECTS OF LOCAL AND GENERAL EXHAUST VENTILATION ON CONTROL OF CONTAMINANTS

Ventilation 1 EFFECTS OF LOCAL AND GENERAL EXHAUST VENTILATION ON CONTROL OF CONTAMINANTS A. Kelsey, R. Batt Health and Safety Laboratory, Buxton, UK British Crown copyright (1) Abstract Many industrial processes are performed indoors and any release of contaminants from processes must be controlled to limit worker exposure. The main objective of the present study was to use computational fluid dynamics (CFD) to explore the capture of a release of a contaminant under a range of local exhaust ventilation (LEV) and general exhaust ventilation (GEV) conditions. The study was planned as an initial phase of work designed to improve the environmental conditions of workers. Three unflanged LEV extracts of.415 m,.83 m and.166 m diameter were modelled using CFD. Flow rates through the extracts were varied to examine the effect on capture behaviour. Simulations using additional imposed air change rates from to air changes per hour were used to examine the influence of GEV on the LEV capture behaviour. The contaminant was modelled as massless particles tracked through the flow field from release points at different distances from the LEV. The CFD simulations predicted qualitatively similar behaviour to observations and LEV centreline velocities compared well with empirical results. The simulation results were examined using capture efficiencies and times. The results showed how high GEV air change rates led to a reduction in capture efficiency. However, higher LEV flow rates could overcome the effects of the higher air change rates, but with potential loss of comfort for workers with increased noise levels and energy requirements. Particle tracks (ensemble-averaged) and their corresponding residence times indicated that while some particles were captured rapidly others were resident for extended periods before capture, during which the flow could be disturbed and exposure could occur. Keywords: LEV, GEV, CFD, capture, control 1 Introduction The most common type of local exhaust ventilation (LEV) hoods are capture, or exterior, hoods. The process, source and contaminant cloud are situated outside capture hoods. Though it is known that more effective control can be achieved when sources are fully or partially enclosed, capture hoods are widely used as they can be added to processes more easily than other types of LEV and they are perceived to cause less interference with the process. However, capture hoods are susceptible to draughts and their effective capture zone, the region where a capture hood can capture a contaminant, is typically smaller than users expect (HSE, 11). In this paper we examine influences on the capture performance of LEV by carrying out simulations using Computational Fluid Dynamics (CFD). Method Page 1 / 6

Ventilation 1.1 Geometry, flow and release conditions Local exhaust ventilation capture behaviour was studied in an enclosure with plan dimensions 4 m % 4 m and a height of.9 m. The layout of the enclosure is shown in Figure 1. LEV GEV Source Figure 1 Enclosure layout General exhaust ventilation (GEV) was imposed by specifying flow rates through a circular outlet, with radius.15 m, set in the ceiling of the enclosure. There were four other ventilation openings, consisting of two offset diagonal pairs of openings on opposite walls. The LEV flow rate was also imposed over a circular outlet. This outlet formed one end of an unflanged cylinder of length. m. The axis of the cylinder was parallel to the floor, at a height of 1.5 m, on a vertical plane between inlets. Influences on capture performance of LEV were examined using a series of CFD simulations. The LEV was varied directly by changing the LEV diameter and flow rate. The LEV flow rates used correspond to a face velocity of 1 ms -1 for each of the diameters. The effect of ventilation in the enclosure was studied by imposing additional GEV flows. Using sources at different distances from the LEV the variation in capture performance with separation was examined. The parameter values used are shown in Table 1, simulations were attempted for all the combinations of parameter values, though in some cases converged results were not achieved. Table 1 Parameters varied in the study LEV diameters (m).415,.83,.166 LEV flow rates (m 3 s -1 ).135,.541,.164 GEV air changes per hour (ACH, hour -1 ), 5, 1, GEV flow rates (m 3 s -1 ).,.65,.13,.6 Separation, source to LEV (m).166,.331,.664. Numerical modelling The CFD simulations used STAR-CCM+ version 4. (CD-adapco, 8). Most of the simulations were run as steady state. Where necessary transient simulations were performed to check that results did not differ significantly from the steady state solutions. The calculations were isothermal and the air was modelled with constant density, as compressible effects were expected to be negligible. The Realisable k-ε model was used with -layer, all y + wall functions, as recommended by default in STAR-CCM+. The convergence criteria were set at 1-4 for all flow variables. Contaminant was modelled as massless particles injected at specified source positions with zero initial velocity. The particles were transported by the flow fields, with the influence of turbulent velocity fluctuations represented using the approach of Gosman and Ioannides (1983). Page / 6

Ventilation 1 The simulations used polyhedral meshes, with prism cells at walls. The mesh was locally refined in the region of the LEV. Mesh dependence was examined by performing simulations at two levels of refinement. Increased refinement showed little influence on the predicted flows and therefore the coarser mesh was used in the simulation results presented..3 Analysis Quantitative information was not available for comparison with the simulation of capture behaviour. However, a qualitative assessment was performed using images from experiments similar to the simulations. In addition predictions of velocity along the axis of the LEV were compared with those calculated using empirical expressions. Two measures of capture performance were used to analyse the capture behaviour. These were the capture efficiency (the percentage of released particles captured by the LEV) and capture times (the residence times of particles between release and capture by the LEV). The effect of parameter variation on these quantities was examined using lattice plots, Sarkar (8), created using the R statistical computing language, R Development Core Team (1). 3 Results 3.1 Contaminant dispersion Images from experiments, performed at HSL, using smoke to visualise the interaction of quiet sources with LEV are shown in Figure (Pocock, 1). These show that as GEV increases smoke no longer travels by a direct route from source to LEV. Images from equivalent CFD simulations are also shown in Figure, showing similar behaviour. The simulation results are for an ensemble of particle tracks, unlike the experimental images which show a single realisation. a) Air Changes per Hour b) 5 Air Changes per Hour c) Air Changes per Hour d) 5 Air Changes per Hour Figure Effect of air change rate on capture behaviour, source is 8 LEV diameters from LEV face Page 3 / 6

Ventilation 1 3. Centreline velocities For an unflanged circular hood, the centreline velocity can be predicted using the empirical simplified Dallavalle expression (Braconnier, 1988). V ( z) 1 (1) = V 1.7z + 1 where V is the face velocity of the LEV hood and V ( z) is the centreline velocity at a distance z LEV diameters from the face. The expression is only valid in the range z 1. 5 LEV diameters from the face (Braconnier, 1988). Figure 3 shows the predictions of this formula and the results from CFD simulations for three LEV diameters, providing confidence in the prediction of the near field flow of the LEV hood. In Figure 4 increased LEV flow rate is shown to control the influence of higher GEV air change rates on centreline velocity. 4 6 8 1 1 Centreline velocity (V V ) 1.1.1.1 Diameter (m) :.415 Diameter (m) :.83 Diameter (m) :.166 4 6 8 1 1 Centreline distance in LEV diameters 4 6 8 1 1 Figure 3 Centreline velocities: solid line simplified Dallavalle expression, points CFD simulations at Air Changes per Hour..1..3.4.5 Centreline velocity (ms 1 ) 1 1.1.1 Diameter (m) :.166 Flow rate (m 3 s 1 ) :.135 Flow rate (m 3 s 1 ) :.451 Flow rate (m 3 s 1 ) :.164..1..3.4.5..1..3.4.5 Centreline distance (m) 5 1 Air change rate (ACH) Figure 4 Effect of changing GEV on LEV centreline velocity 3.3 Capture performance Capture efficiencies at the same LEV flow rate show similar behaviour for all LEV diameters, Figure 5. The capture efficiency decreased with increasing separation and GEV air change rate. Higher LEV flow rates reduced the decrease in capture efficiency with separation and air change rate. Capture times increased with separation of source and LEV. Also, in addition to modifying the capture rate, increasing GEV air change rates increased the spread of capture times, Figure 6. This is due to the increasing spread of routes taken from source to LEV, seen in Figure. Reduced control, Page 4 / 6

Ventilation 1 due to either separation or increased GEV flow rates, increased the time between release and capture and would therefore increase the chances of exposure....4.6.8 1 Flow rate (m 3 s 1 ) :.135 Flow rate (m 3 s 1 ) :.451 Flow rate (m 3 s 1 ) :.164 8 6 4 Diameter (m) :.166 Capture efficiency (%) 1 Diameter (m) :.83 Flow rate (m s ) : 135 Flow rate (m 3 s ) : 135 Diameter (m) : 83 Flow rate (m s ) : 451 Flow rate (m 3 s ) : 451 Diameter (m) : 83 Flow rate (m s ) : 164 Flow rate (m 3 s ) : 164 1 8 6 4 8 6 4 Diameter (m) :.415 Diameter (m) : 415 Diameter (m) : 415...4.6.8 Distance from face (m) 5 1 Air change rate (ACH)...4.6.8 Figure 5 Changes in the capture efficiency with source distance from the LEV face 4 Conclusions CFD simulations have been used to examine the influence of LEV hood diameter and flow rate, GEV air change rate, and separation of source and LEV on the capture behaviour of LEV capture hoods. Qualitative comparison of results from the CFD simulations with smoke visualisations of capture showed similar behaviour. The LEV centreline velocities predicted using CFD were in agreement with the simplified expression of Dallavalle (Braconnier, 1988). Increasing GEV air change rate was shown to modify the predicted centreline velocities but the effect of increased GEV air change rates could be controlled by higher LEV flow rates or repositioning the hood closer to the source. In reality the latter would usually be more practical. Influences on capture were examined using capture efficiencies and capture times. For the same LEV flow rate increased GEV air change rates reduced the performance of LEV, but increased LEV flow rates could overcome the effects of higher GEV air change rates. This publication and the work it describes were funded by the Health and Safety Executive (HSE). Its contents, including any opinions and/or conclusions expressed, are those of the authors alone and do not necessarily reflect HSE policy. Page 5 / 6

Ventilation 1 5 References Braconnier R (1988) Bibliographic review of velocity fields in the vicinity of local exhaust hood openings, American Industrial Hygiene Association Journal 49, 185-198. CD-adapco (8) User guide STAR-CCM+ Version 4..7, CD-adapco. Sarkar D. (8) Lattice Multivariate Data Visualization with R, Springer. Gosman A D and Ioannaides E (1983) Aspects of computer simulation of liquid-fuelled combustors, AIAA, 7, 48-49. HSE (11), Controlling airborne contaminants at work: A guide to local exhaust ventilation (LEV), nd Edition, HSE. Pocock D J (1). Personal communication. R Development Core Team (1) R: A language and environment for statistical computing, R Foundation for Statistical Computing, Vienna, Austria. 1 1 1 Separation (m) :.166 Separation (m) :.33 Separation (m) :.664 6 4 Diameter (m) :.166 Separation (m) : 166 Separation (m) : 33 Separation (m) : 664 Probability density Diameter (m) :.83 Diameter (m) : 83 Diameter (m) : 83 6 4 Separation (m) : 166 Separation (m) : 33 Separation (m) : 664 6 4 Diameter (m) :.415 Diameter (m) : 415 Diameter (m) : 415 1 1 1 1 1 1 Capture time (s) 5 1 Air change rate (ACH) Figure 6 Distribution of particle parcel capture times at LEV flow rate of.135 m 3 s -1 Page 6 / 6