Ann. occup. Hyg., Vol. 46, No. 7, pp. 629-635, 2002
© 2002 British Occupational Hygiene Society
Published by Oxford University Press
Determination and Interpretation of Total and Transversal Linear Efficiencies in Push–Pull Ventilation Systems for Open Surface Tanks
1 Universidad Politécnica de Cartagena, Departamento de Ingeniería Térmica y de Fluidos, 48 Paseo de Alfonso XIII, 30203 Cartagena, Murcia, Spain; 2 Universidad de Murcia, Departamento de Ingeniería Química, 30071 Espinardo, Murcia, Spain
Received 28 March 2002; in final form 12 June 2002
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
A real-scale pilot installation simulating an open surface treatment tank with a push–pull ventilation system has been designed. From experiments carried out, typical representations of the total and transversal linear efficiencies show that when total efficiency is related to push flow rate, taking as a parameter the pull flow rate, a parabolic profile is obtained with a maximum point or plateau that increases as the pull flow increases. When the transversal linear efficiency is analysed, three general zones where losses occur to the exterior can be detected: (i) when the push flow rate is low, any distortion in the wall jet, whether external (e.g. in the air flow inside the workshop) or internal (e.g. thermal effects), provokes an escape from contaminant; (ii) in the impact zone, where the push flow impacts on the tank surface, distortion increases as the push flow rate increases; (iii) when the push/pull flow rate ratio increases and preferential currents are produced inside the exhaust hood, these escape and cause substantial losses in efficiency.
Keywords: push–pull ventilation; tracer gas; capture efficiency; transversal linear efficiency; open surface tank ventilation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The push–pull ventilation systems used in surface treatment tanks have complex flow dynamics (Hughes, 1990; Flynn et al., 1995; Robinson and Ingham, 1996; Rota et al., 2001), as can be seen in Fig. 1, where the following mechanisms can be identified (Marzal et al., 2001): (i) the push flow forms a continuous semi-free curtain under which a depression is formed, which generates a feedback current; (ii) the curtain impacts on the tank, initially producing a strong distortion; (iii) the flow reorganizes into a wall jet, which, as it flows over the surface, continuously draws in external air, this flow forming vortices along the side walls; (iv) this current enters the bottom of the exhaust hood.
|
If the emission of contaminants from the tank surface is homogeneous, the concentration of the push flow increases gradually as it approaches the exhaust. Therefore, the greater the push flows escaping from the tank and the nearer such escapes are to the exhaust hood, the greater the loss of efficiency.
Visualization of flows using smoke is a good technique for analysing the behaviour of the above-mentioned flows [Woods and McKarns, 1995; Marzal et al., 2000; Maynard et al., 2000; American Conference of Governmental Industrial Hygienists (ACGIH, 2001)], although greater information is provided by tracer gases, which permit the capture efficiency to be quantified (Hampl et al., 1986; Niemelä et al., 1991; Bemer et al., 1998; Marzal et al., 2002a).
In this paper we describe the conclusions reached after studying the total and transversal linear efficiencies observed in a series of controlled experiments carried out in a pilot installation specifically designed for this purpose. The zones where escapes to the exterior are detected (losses of efficiency) and the emissions are quantified.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Figure 2 shows a diagram of the pilot installation used. The push flow is generated in 3a, which consists of a plenum that finishes in a longitudinally perforated tube through which uniform jets of air escape. These form a wall jet, which travels along the tank and enters the exhaust hood (2a), 30 cm high. The geometry and dimensions were previously optimized (Marzal et al., 2002a).
|
The tank, which measures 1.60 m wide by 1.80 m long, can slide below the exhaust hood, permitting three different lengths to be studied: 1.20, 1.53 and 1.80 m. A system of electrical resistances permit controlled and uniform heating of the surface up to 100°C.
As the push system we used a series of interchangeable tubes (1.6 m in length and 50 mm in diameter) closed at both ends and punctured along their length by holes (5, 10, 15 or 20 mm in diameter). The back was an open continuous slot (20 mm) to facilitate a uniform jet speed. The tubes were situated 0.10 m above the tank surface, resting on the side walls of the tank. Their circular cross-section meant that the angle at which the jets were emitted could be altered from 0° (parallel to the tank surface) to 45°. The flow rate, which is measured by a venturi, could be regulated from 0.0025 to 0.4 m3/s by varying the turning speed of the fan. The exhaust system had a rectangular entrance 1.60 m wide and a height which could be varied from 0.15 to 0.60 m. In this study a height of 0.30 m was used. A frequency variable fan varied the exhaust flow rate between 0.08 and 2.5 m3/s.
To determine capture efficiencies, the tracer gas used was sulphur hexafluoride, which was incorporated at a constant flow rate as follows. To evaluate total efficiency, the tracer gas was injected (at 55–75 cm3/min) and mixed with air (20 000 cm3/min) from the tank surface by means of a gas grid of 460 holes, each of which consisted of a 0.3 mm hypodermic needle to ensure uniformity of emission over the whole tank surface. The tracer mixed with the push flow and travelled towards the exhaust. A venturi (2b) in the exhaust duct measured the total flow captured and homogenized the air–tracer mixture. The concentration of sulphur hexafluoride (Ca) was measured by an infrared spectrometer (4a), which made a suitable number of measurements (six to ten), for which variation coefficients of less than 2% were obtained. The same tracer flow rate was then introduced into the exhaust duct without modifying any of the geometric or operational characteristics of the push–pull ventilation system. The concentration was again measured (CaT), this representing the (maximum) reference value. The total efficiency (ET) value is obtained by:
ET = (Ca/CaT) x 100
To determine the so-called transversal linear efficiency, we used a diffuser consisting of a tube (1.58 m in length and 8 and 10 mm in internal and external diameter) closed at both ends and connected in the centre to the tracer generating equipment (Fig. 3).
|
The tube was perforated along its length by 100 holes of 0.3 mm diameter, through which the tracer emerged uniformly; in this case the flow rate of tracer was kept constant at 75 cm3/min, mixed with air (500 cm3/min) and settled over the tank parallel to the push. The experiments started with the tube as near as possible to the push and its concentration was measured (Cal), as in the previous case. The diffuser was then gradually moved towards the exhaust and the concentrations were measured again. All the measurements were compared with the reference (maximum), CaT, obtained by introducing the tracer into the exhaust duct. The transversal linear efficiency, El, is obtained by:
El = (Cal/CaT) x 100
The relation between the above efficiencies is given by (Marzal, 1999):
where L is the total tank length and l is the distance from the push element to the diffuser tube.
Figure 4 shows a typical profile of transversal linear efficiency. From Fig. 4 and equation (2) we can deduce the following (Marzal, 1999).
|
1. The area of the rectangle 01CL0 represents the total mass flow of the tracer emitted from the tank, proportional to CaT.
2. The total losses of tracer (not captured by the exhaust hood) are represented by the area ADB1A.
3. According to the above, the amount of tracer captured by the ventilation system is represented by the area ADBCL0A, proportional to Ca. Consequently, the total efficiency is determined from:
ET = (area ADBCL0A/area 01CL0) x 100
4. The losses of tracer occur where variations in efficiency are observed. In this case, all the losses occur in the zone DB, where it is verified that dEl/dl 
0. In other words, the tracer emitted in the zone AD does not suffer losses before D. All the tracer that reaches zone BC and that is emitted in the zone is captured by the exhaust.
When the losses occur in two or more zones the representation of transversal linear efficiency is similar to that indicated in Fig. 5. Note that the losses in zone 1 are proportional to the area ABCGA and in zone 2 to area GCDE1G (point G is obtained by tracing a line parallel to the axis of the abscissa from C).
|
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Figure 6 shows a generic representation of several total efficiency profiles as a function of push flow rate, Qi, for a given tank, taking as a parameter the pull flow rate, Qa (Marzal et al., 2002b). In general, the profiles show a quasi-parabolic shape, with one point or plateau as maximum, the height and amplitude of which increases with the pull flow rate. To explain this morphology, we analyse the transversal linear efficiency for the conditions indicated by points x, y and z on one of the curves depicted in Fig. 6. These points are situated in one of the characteristic regions of the profile, obtaining typical representations indicated in Fig. 7.
|
|
At point x on the rising part of the total efficiency profile (Fig. 6), where the push flow rate is low, there are continuous losses of tracer from A to B (Fig. 7), coinciding with regions 1–3 and, partially, 4 of Fig. 1, where the wall jet is totally developed. This behaviour is due to the small momentum flux of the push flow, which produces great instability. Small outside distortions (movement of air in the workshop) or internal ones (higher temperature in the tank than outside, resulting in convective currents) lead to tracer gases escaping from the tank. These are not captured by the exhaust and so total efficiency is affected. These losses are slightly greater furthest from the push, as is reflected in the greater slope of the transversal linear efficiency curve near B (3–4 in Fig. 1). However, this does not necessarily mean that the total of air tracer flow lost increases as the wall jet advances, only that it might be related with the greater concentration of the wall jet as it nears the exhaust hood. After B, the transversal linear efficiency is 100%, i.e. a situation is reached in which the capture induced by the exhaust is sufficient to prevent any loss from the push flow.
Point y of Fig. 6 represents the value or optimal zone of the total efficiency profile. When the transversal linear efficiency is evaluated for this point, it is seen that tracer loss is only concentrated in the zone nearest the push (CD in Fig. 7 and 1–2 in Fig. 1) as a consequence of the impact of semi-free jets over the tank, no losses occuring after consolidation of the wall jet. Put another way, the momentum flux of the wall jet provokes a sufficient depression over the tank to overcome both external and internal distortions, so that any initial escape is returned to the zone of influence of the ventilation system. As an example of such a situation, we describe an experiment in which the transversal linear efficiency is evaluated under the following conditions: push element, resting on the tank, with 53 holes of 20 mm diameter; push flow rate 0.035 m3/s; tank length 1.53 m, temperature 50°C; pull flow rate 0.175 m3/s; exhaust hood height 30 cm; angle of push element 0°. The experimental results are illustrated in Fig. 8. Continuous escapes occur in the region between 0 and 600 mm, as a result of escapes through the initials jets and in the wall jet impact and reorganization zone.
|
Point z (Fig. 7) suggests two loss zones (EF and GH). The first is that mentioned for y, which increases slightly as the push flow rate increases, and the second (GH in Fig. 7 and 5 in Fig. 1) is located inside the exhaust hood itself. In this last case, we evaluated whether the losses were the result of ‘overflowing’, i.e. that perhaps the push flow rate at the entrance to the exhaust hood might be greater than the exhaust flow rate. However, our estimations of the push flow rate led us to discard this hypothesis and we think rather that with sufficient energy the flow overcomes the retention capacity of the exhaust. As an example of such a situation, we describe an experiment in which the transversal linear efficiency is evaluated under the following conditions: push element, resting on the tank with 103 holes of 10 mm diameter; push flow rate 0.043 m3/s; tank length 1.80 m, temperature 50°C; pull flow rate 0.185 m3/s; exhaust hood height 30 cm; angle of push element 0°. The experimental results are illustrated in Fig. 9. Continuous escapes occur in the region between 100 and 400 mm (curtain impact and wall jet reorganization zones), slight escapes through the initial jets (0–100 mm) and in the wall jet (400–1800 mm). Escapes are also detectable in the exhaust hood (1800 mm).
|
From the above it can be deduced that the loss of contaminants from the tank can be represented as a polynomic function in the form:
losses (%) = (100 – efficiency) (%) = A + B + C
where A, B and C are functions that depend on the momentum flux of the push and pull flow rates.
A represents the losses resulting from the impact on the tank surface, including those produced in the feedback current formed between the impact zone and the push element. These losses increase slightly as the push flow rate increases as long as the impact zone is near the push exit, since in this area the ‘accumulated’ concentration is still small. Given this, it would be of interest to design the push exit so that violent impacts on the tank surface could be avoided.
B represents the losses produced in the wall jet, which may be practically non-existent above a certain momentum flux of the wall jet.
Finally, C represents the losses from the exhaust hood as a result of preferential currents of the push curtain with a high momentum flux as it enters the hood, so that they exit again. These losses are greater as the push/pull flow rate ratio (Qi/Qa) increases, very probably depending on (Qi/Qa)2 given that the momentum flux of both flows are quadratic functions of their respective flows. To avoid the formation of these preferential currents baffles should be placed inside the hood to break the internal recirculation of air and to conduct the total flow towards the exhaust duct.
To summarize, the quantitative treatment of the efficiency (or losses) or models that attempt to reflect the real behaviour of these ventilation systems should not be based on a study of the behaviour of the wall jet, but should take into account the above-mentioned contributions and the combined influence of flows and the momentum fluxes of both the push and pull flows.
|
FOOTNOTES |
|---|
* Author to whom correspondence should be addressed. E-mail: fj.marzal@upct.es
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
ACGIH. (2001) Industrial ventilation—a manual of recommended practice, 24th edn. Cincinnati, OH: American Conference of Governmental Industrial Hygienists.
Bemer D, Muller JP, Dessagne JM. (1998) Comparison of capture efficiencies measured by tracer gas and aerosol tracer techniques. Indoor Air; 8: 47–60.
Flynn MR, Ahn K, Miller CT. (1995) Three-dimensional finite-element simulation of a turbulent push-pull ventilation system. Ann Occup Hyg; 39: 573–89.
Hampl V, Niemelä R, Shulman S, Bartley D. (1986) Use of tracer gas technique for industrial exhaust hood efficiency evaluation—where to sample? Am Ind Hyg Assoc J; 47: 281–7.[Web of Science][Medline]
Hughes RT. (1990) An overview of push-pull ventilation characteristics. Appl Occup Environ Hyg; 5: 156–61.
Marzal F. (1999) Metodologías de evaluación de sistemas de ventilación de baños de tratamiento superficial. Tesis doctoral, Departamento de Ingeniería Térmica y de Fluidos, Universidad Politécnica de Cartagena, Murcia, Spain.
Marzal F, González E, Miñana A, Baeza A. (2000) Open surface tanks ventilation: some design criteria. In Ventilation 2000, Proceedings of the 6th International Symposium on Ventilation for Contaminant Control, Helsinki, 2000. Vol. 2, pp. 41–44.
Marzal FJ, González E, Miñana A, Baeza A. (2001) Control de emisiones en baños rectangulares mediante ventilación por impulsión-aspiración: relación de flujos implicados. Montajes Instalaciones; 347: 89–92.
Marzal FJ, González E, Miñana A, Baeza A. (2002a) Influence of push element geometry on the capture efficicency of push-pull ventilation systems in surface treatments tanks. Ann Occup Hyg; 46: 383–93.
Marzal FJ, González E, Miñana A, Baeza A. (2002b) Analytical model for evaluating lateral capture efficiencies in surface treatment tanks. Am Ind Hyg Assoc J; 63.
Maynard A, Thompson J, Cain JR, Rajan B. (2000) Air movement visualisation in the workplace: current methods and new approaches. Am Ind Hyg Assoc J; 61: 51–5.
Niemelä R, Lefevre A, Muller JP, Aubertin G. (1991) Comparison of three tracer gases for determining ventilation effectiveness and capture efficiency. Ann Occup Hyg; 35: 405–17.
Robinson M, Ingham DB. (1996) Numerical modelling of the flow patterns induced by a push-pull ventilation system. Ann Occup Hyg; 40: 293–310.
Rota R, Nano G, Canossa L. (2001) Design guidelines for push–pull ventilation systems through computational fluid dynamics modelling. Am Ind Hyg Assoc J; 62: 141–8.
Woods JN, McKarns JS. (1995) Evaluation of capture efficiencies of large push-pull ventilation systems with both visual and tracer techniques. Am Ind Hyg Assoc J; 56: 1208–14.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L.-C. TSENG, R. F. HUANG, C.-C. CHEN, and C.-P. CHANG Aerodynamics and Performance Verifications of Test Methods for Laboratory Fume Cupboards Ann. Hyg., March 1, 2007; 51(2): 173 - 187. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. KULMALA, P. HYNYNEN, I. WELLING, and A. SAAMANEN Local Ventilation Solution for Large, Warm Emission Sources Ann. Hyg., January 1, 2007; 51(1): 35 - 43. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. F. HUANG, S. Y. LIN, S.-Y. JAN, R. H. HSIEH, Y.-K. CHEN, C.-W. CHEN, W.-Y. YEH, C.-P. CHANG, T.-S. SHIH, and C.-C. CHEN Aerodynamic Characteristics and Design Guidelines of Push-Pull Ventilation Systems Ann. Hyg., January 1, 2005; 49(1): 1 - 15. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||