DEVELOPMENT OF AN EFFICIENT T-TYPE STRAINER WITH ITS PERFORMANCE EVALUATION

This paper was recommended for publication in revised form by Assigned Editor Mustafa Kilic 1 Department of Mechanical Engineering, SPCE, University of Mumbai, Maharashtra, India *E-mail address: gmahi1193@gmail.com Orcid id: 0000-0003-4100-7252, 0000-0002-0564-1603 Manuscript Received 09 August 2018, Accepted 21 September 2018 DEVELOPMENT OF AN EFFICIENT T-TYPE STRAINER WITH ITS PERFORMANCE EVALUATION


INTRODUCTION
Strainer is a filtering device which helps to remove particulates from fluid or gas flow to protect sensitive downstream equipments. It collects dirt and debris without hampering flowing system's performance significantly. Usually, they are placed before pump, compressor, turbines, automatic valves and other inline equipment to protect them. Absence of strainers in a process line could be one of the main reasons for the requirement of very frequent maintenance. Strainers of different designs are used for diverse applications -flat type, cone and basket type, Y-type, T-type and vertical basket type. Figure 1 shows quarter vertical cut section of a T-type strainer. The basic components of a strainer are -container, wire mesh, punch plate and orifice. Container is designed to fit with piping system. The elements of strainer rest on sitting in the container. Wire mesh is actual filtering element which is made up of metal wires. Mesh metal wire diameter and mesh count varies as per application. Pun ch plate supports thin wire mesh which is relatively thicker than wire mesh and placed just outside or inside of it. Wire mesh and punch plate are held at their respective places by an annular ring-shaped structure called as orifice. This arrangement causes decrease in diameter of inlet pipe at strainer entry and thus produces effect of orifice.

Figure 1. Cut view of T-type strainer
Journal of Thermal Engineering, Technical Note, Vol. 6, No. 6, Special Issue 12, pp. 420-433, December, 2020 421 Debris particle flowing with fluids is either trapped in the mesh cavity or blocked. This results in reduced available opening for flow. Strainer sizing and selection are critical to insure cleaning cycle time, particle size retention and pressure drop requirements. Predicting possible pressure drop across a strainer is an important part of overall process system design. This information is generally provided by manufacturer in the form of performance data sheet or semi-empirical method present in various literatures. On many instances it is observed that the actual pressure drop through strainer comes out to be significantly different than what predicted by semi -empirical method.
Characterizing an industrial strainer regarding its hydraulic performance is a challenging task. Experimental testing of the device is reliable but time-consuming methodology. Few researchers tried theoretical and numerical approach to overcome with partial success. In a work on temporary conical strainer, Carlomagno et al. [1] developed a theoretical-experimental model to characterize the pressure drop and mechanical resistance. They used 6 inch cone geometry, with 40% porosity and a base angle of 81.35° with mass flow is 0.2 kg/s. They investigated the impact of cone angle, porosity and thickness and their effects on the performance of strainer. The clogging characteristic was also investigated. Through a simple experimental and analysis, Kang et al. [2] investigated flow through wire screen mesh and derived a proportional factor between the velocity of fluid and head loss which can estim ate coefficient of discharge, Cd, for wire mesh. Using PHOENICS computer code, Erdal and Andersson [3] carried out a 2D-numerical investigation for flow through a single-holed plate. They concluded that turbulent kinetic energy and pressure drop are more grid-sensitive than the mean axial velocity in the down-stream of the orifice. Flow through an oil filters has been numerically investigated by Iliev et al. [4]. They used Navier-Stokes-Brinkman system successfully to simulate flow through plain and porous media both. Malavasi et al. [5] performed an experimental work to investigate the requirement of the pressure losses through sharp-edged perforated plates with key parameters of geometrical and flow variables. They studied the dependency of pressure loss coefficient to the Reynolds number, the equivalent diameter ratio, relative thickness and the number and deposition of the hole. Filho et al. [6] carried out a numerical work to investigate pressure loss across perforated plates used in pressurized water reactors. They concluded that the standard K-ϵ turbulence model is a good compromise between computational time and accuracy. Tek am et al. [7] presented a comparison between numerical and experimental methods to predict the efficiency of an automotive oil strainer. Gaurav et al. [8] carried out a numerical investigation of a T-type strainer to assess the impact of strainer body size keeping strainer element of same size. They were successful in representing fluid flow distribution and associated pressure drop across strainer.
Literature review reveals that a lot of work has done by researchers on of different k ind of filtration methods. Industrial strainers have been less attended by researchers for characterizing their hydraulic performance -predicting pressure drop. Limited numerical work on limited strainer types are reported in the literature. A methodology of investigation to improve the performance of an existing strainer design is found to be not sufficiently established. There is no comprehensive theory for the prediction of hydraulic performances of strainers. Overall, literature in area of computational modeling of strainers seems to be scarcely published.
The present study deals with a 12 inch T-type industrial strainer shown in figure 2. Inlet and outlet diameter are 303.74 mm. It has 40 mesh (36 SWG) of thickness 0.001393 m with backing of perforate d sheet having 6 mm hole arranged with 8 mm triangular pitch. The direction of fluid flow is shown in figure by arrow. The objective of the present work is to develop a numerical methodology to capture the underlying physics accurately and be able to predict pressure drop. It is also intended to carry out a geometrical parametric study for the development of an efficient strainer. The scope of work is limited to the flow rate in the range of 260 m 3 /hr to 470 m 3 /hr.

MATHEMATICAL MODELLING
The flow through the strainer is assumed to be isothermal steady incompressible Newtonian fluid flow. The body force, viscous forces and pressure gradient are significant contributor to the momentum equation. Viscous and kinetic resistance to flow is dominant. There is no blockage of mesh.

Governing Equations:
Continuity equation, where, the transient term is intentionally shown to update velocity field evolving during iterative solution.
Turbulence Model: k-ω model solves 2 transport equations, one for the turbulent kinetic energy (k), and one for the turbulent frequency (ω).

k-Equation:
( ) All terms represent usual meaning.
where, C2 is the pressure jump coefficient which is calculated by following mathematical model [9].
The equivalent pressure drop coefficient (keq) is given as, Now, the pressure drop coefficient of punch plate (Kp) is given by [5], Porosity of the screen ( 2 ) 2 = The tabular coefficient (τ) depending on tp/dh is found out from hydraulic resistance handbook [10] and friction coefficient (λ) found from [11]. Pressure drop coefficient of wire mesh (Kw), Pressure discharge coefficient (Cw), calculated from Perry's chemical engineers handbook [9].

NUMERICAL IMPLEMENTATION
To investigate the problem numerically, the geometry of computational domain is created and meshed through ICEM-CFD as per drawing obtained from M/S Bombay Chemical Equipments, Mumbai. It is shown in figure 3.
Meshing is done such that all essential features of strainer can be captured easily. The fluid flow through strainer is considered to be steady. The effect of gravity has been neglected. The cells count in meshing is around 1 million.
As the flow is highly turbulent in nature, a number of turbulence models are tested, where k -ω SST model is found to be more appropriate for present case. Pressure based SIMPLE algorithm is used to couple pressure and velocity. Second order upwind scheme is used in all cases of upwinding. PRESTO scheme is used as pressure discretization scheme. Under relaxation factors are set as default. Inlet v elocity is varied from 1.01 m/s to 1.8 m/s for different flow rates. Outlet condition is taken as pressure outlet and considered as zero gauge pressure. All solid surfaces are considered to have no slip condition. Strainer elements are given jump condition based on semi -empirical correlation [9].

RESULTS AND DISCUSSION Flow Through T-type Strainer:
The flow resistance is expected as fluid passes through strainer element causing a pressure drop across strainer element. The flow inside bottom pocket of strainer should be supportive to trap foreign particles and debris which are harmful for downstream components.   Journal of Thermal Engineering, Technical Note, Vol. 6, No. 6, Special Issue 12, pp. 420-433, December, 2020

425
The contours of velocity from figure 5 shows that the trap section with stagnation ambience which is favourable for debris to reside before next cleaning operation is performed for the strainer.
In order to validate the model adopted for investigation of strainer, the performance testing of the strainer was done by M/s Bombay Chemical Equipments, Mumbai at testing facilities located at Central Water and Power Research Station, Khadakwasla, Pune. The experiments were conducted on same strai ner which were at 100 % clean condition. The pressure drop across strainer is recorded. Normal water without any debris have used as fluid media. The pressure drop was measured on inlet and outlet nozzle across the strainer assembly using gauges. Using semi-empirical formulae available in Perry's handbook [9], calculation for pressure drop was computed. Head loss across strainer, ∆ℎ = ∆ℎ + ∆ℎ (10) where, ∆ℎ =   Excessive filtration activity in specific portion of the strainer leads to quick wear and tear which needs frequent replacement of strainer element. On many systems, a robust strainer is desired due to several reasons. In order to avoid this problem and to improve the life of strainer following section present a numerica l investigation of strainer where few possible geometrical changes has recommended.

Flow Through Modified Strainer:
Punch plates are generally used in current strainers as support to mesh element for their longer life and to prevent deformation and erosion. The flow patterns predicted through previous investigation clearly shows that only a specific section of meshing element always has loads of flow which tends to reduce its life and major portion of
Inlet outlet not inline 03.70 Figure 9(f) shows a good flow distribution and least pressure drop among all cases. No punch plate is introduced here. Pressure drop is minimum amongst all. This arrangement appears to be better strainer design configuration for minimum pressure drop and better utilization of mesh element. Only constraint is that incoming and outgoing flow lines would not be inline. With the case of total punch plate partition i.e. configuration 8(e) gives better flow distribution and comparatively lower pressure drop among all similar cases. So, this arrangement will be useful in cases where flow distribution is required in strainer and if higher pressure drop is manageable.

Effect of Punch Hole Size of Total Punch Plate:
The configuration 8 (e) gives lesser pressure drop among all cases with good flow distribution. Punch hole is expected to influence flow distribution and pressure drop significantly. The case has been investigated for the punch holes of size 10, 15 and, 20 mm. Figure 10 shows the variation of P with punch hole size of punch plate.
The variation of pressure drop is non-linear and decreases with increase of punch hole size. Increase of punch hole from 10 mm to 15 mm leads to bigger P compared to change in P from 15 mm to 20 mm.

Changing strainer body size while keeping straining element size same
The investigation report presented in previous sub-sections was for strainer of body size 12 inch. Bigger body size of strainer is a crucial from pressure drop point of view and a significant improvement in performance is expected. This section deals with impact of body size on pressure drop across strainer. The strainer body size considered for investigation are 13 inch, 14 inch, 15 inch and 16 inch at a flow rate of 264.48 m 3 /hr. Figure 11 shows the variation of outer body from 13 inch to 16 inch by keeping the strainer element same as of original case.  An increased gap between strainer mesh and outer body tends to deflect the peripheral annulas fluid towards core and changing the overall momentum of upstream fluid. It makes the flow distribution better and helps in lower the pressure drop. The distribution of pressure and velocity for strainer with 14 inch outer body size is shown in figure  12. Figure 13 shows a significant non-linear variation of pressure drop with increasing body size of strainer. A variation from 12 inch body size to 16 inch is found to drop about 50% which is shown in figure 13. This significant change in pressure drop may attributed to increasing gap between strainer mesh and outer body which acting as damper to the momentum of approaching stream. It causes lesser pressure drop across strainer. The pressure and velocity distribution on a mid plane along flow direction is represented in figure 15. A drop of pressure across punch plate and development of a stagnant zone (little increase in pressure) due to flow obstruction created by meshing element can be observed. Further drop of pressure across meshing element can be observed. The pressure drop found to be 2.37 kPa at a flow rate of 264.48 m 3 /hr, Velocity is uniformly distributed and well balanced from longevity and erosion point of view.

CONCLUSION
Present work deals with a numerical investigation of pressure drop across T-type Strainer. A numerical methodology has proposed to estimate pressure jump coefficient which is crucial for numerical investigation. The work has been validated with a test rig data of strainer, conducted by Bombay Chemical Equipments, Mumbai. The effect of Re on pressure drop is also investigated which increases with Re. Six geometrical modifications for punch plate have proposed to be placed ahead of straining element. Total punch plate configuration is found to be most efficient amongst all. Variation of punch hole size is also investigated. It is found that bigger diameter is more appropriate without losing strength of the punch plate. Further investigation is carried out for bigger bod y size of strainer, a drastic drop in pressure drop is observed. An illustration of combining all good features for lesser pressure is found to be a good predictive methodology. Present investigation is useful for the strainer designer and fabricator to se lect a proper strainer for a specific application.

Af
Open area of Strainer At Curved surface area of strainer element α Permeability of the medium αw Opening area ratio of mesh screen β Equivalent diameter ratio β 2 The porosity of the screen C2 Pressure