A Study of Water-mist Recirculation System From Scale Kitchen Hood Ventilation

Water-mist spray was installed in several heavy-duty commercial kitchen hood ventilation (KHV) systemsfor grease emission control and safety purpose. Unfortunately, the water-mist process increased the water consumption, which is a part of the operational cost to the user. The aim of this research is to analyzethe water-mist spray characteristics with different fluid properties. The fluid properties are dependent on recycled water-mist wastewater quality. The 8 cycles of wastewater sample was collected from the experimental work at each cycle of 30 minutes cooking loads and the fluid properties was determinedusing viscometer and analytical balance. Thus, the experimental workwas conductedusing scale water-mist kitchen hood ventilation system.Simulation works are performed using Computational Fluid Dynamics (CFD) to determine the water-mist spray characteristics. The inlet watermist pressure studied in this research is 1.67 bar. As the number of cycle increased the percentage of fat, oils and grease (FOG) increased. The streamlinevelocity atnozzle inlet and outlet was compared between experimental and simulation. The highest outlet velocity was recorded from experimental and simulation is from fresh water (at 0 cycle) 10.937 m/s and 10.610 m/s respectively. The water-mist spray activation at 8 cycles shows the lowest outlet velocity is 10.767 m/s from experimental and 10.434 m/s from simulation. The percentage of absolute error of outlet velocity between experimental and simulation at 0 cycle (freshwater) is 2.99 % which is the lowest among the cycles. The highest absolute error is 3.093 % was recorded at nozzle outlet at cycle 8 (WW8). Keywords-water-mist, recirculation, Kitchen hood, CFD, Kitchen Hood Ventilation (KHV)


LITERATURE REVIEW
Majority of cooking operation especially in commercial kitchens generate emission and particulate matter that significantly effect to human health and reducing the air quality. Cooking-particulates are emitted from small scales food outlets in the form of smoke and fumes from barbecuing, roasting, and wood-fired cooking [3][4] [7].
Some wood-fired pizza ovens use gas for cooking with a small wood fire to add flavour. The gas fire probably reduces odours but not particles in the wood-smoke. The particles consist of partially burnt fats emitted from the cooking meat and wood-smoke. These are known to contain materials detrimental to health such as polyaromatic hydrocarbons (PAHs) and fine particles (PM 2.5 ) [7] [8].
Studies on air emissions generated from barbecuing chickens have revealed many unsaturated hydrocarbons, a form of VOCs that is highly reactive in the photochemical process. The primary sources of these emissions were the produced from the cooking processes. For averaging 100 kg per day of cooked meat for 300 days per year, the estimated VOC emissions produced from a small barbecue chicken shopis roughly equivalent to running a typical car for 20,000 km per year.
Grease particles and aerosols can also be generated from certain cooking process. From ASHRAE 1375 final report, the results of the emission tests indicated that the wok had the largest total grease emission in the plume, 247 lb grease/1000 lb food cooked, nearly a factor of two over the appliance with the second highest emission, the solid fuel broiler at 142 lb grease/1000 lb food cooked. The conveyor broiler had 50.2 lb grease/1000 lb food cooked whereas the other appliances were less than 14. The total grease emissions in the exhaust duct were less than the values in the plume primarily because the plume mass emissions are often dominated by particles larger than 10 μm in size. These large particles rarely make it into the exhaust duct to be sampled by the instrumentation located there. Thus the total grease mass emissions in the exhaust duct ranged from 72.5 lb grease/1000 lb food product for the solid fuel broiler to 2.64 lb grease /1000 lb food product for the conveyor pizza oven [9].
Particles larger than 10 μm often dominated the total mass effluent in the plume whereas grease vapor dominated the effluent in the exhaust. Much of the grease mass emission was found to be in the vapor phase or associated with particles smaller than 1μm in size [7] [9]. From the previous investigation, the condensable vapor grease produced from cooking operations are measured from U.S Environmental Protection Agency (EPA) method 5 [10] and method 202 [11] to determine the condensable particulate emission from inside and outside the plenum location. Now days, many of the kitchen hood are equipped with mechanical grease filter such as cyclonic KSA and baffles plate to filter the grease emission from cooking process. During cooking operation, the efficiency of mechanical grease filter or extractor only efficient to capture particulate size larger than 10μm [1], [7], [9]. In real situation, the particulate size range from 0.5μm and up to 20μm [9] depends on the cooking styles. None of the mechanical filters is effective in capturing particles less than 2.5μm [1]. The cyclonic filter traps at least twice the amount of grease compared to the best baffle filter in the particle range from 2.5 to 10μm at the exhaust airflow 274 cfm/ft (424 L/[s·m]) and three times more grease at 144 cfm/ft (223 L/[s·m]) and 196 cfm/ft (303 L/[s·m]) exhaust airflow. The previous study reported the filtration efficiency for all filters increases with the airflow as a result of higher pressure drop across the filter [1].
To increase the emission filtration efficiency, the kitchen hood is equipped with the combination of the mechanical grease filter and water-mist spray. It has been revealed the cold-mist sprays are able to capture the small grease particulate size by increasing the diameter of grease particles from the temperature drop and grease vapor solidified during the cold-mist process [12], [13]. Until now, there is no publication and report documented to reveal and prove the result even though it is possible in theory.

III.
METHODOLOGY To study the grease filtration efficiency by the water-mist recirculation system, the experimental parameter was based on Halton U.S and Halton Europe kitchen hood operating conditions. For the installation of the kitchen hood ventilation system was based on ASTM 2519-05 and UL1046. The grease machine was selected as cooking load in this study for consistency and the standard operating condition comply with UL 1046 standard. Table 1 shows the experimental configuration selected to perform this study.

A. Experimental Setup
The experimental investigation was performed from grease loading by utilizing the grease machine and scaled (1:2) kitchen hood canopies as in Fig. 1. The grease machine was built based on a design detailed in UL 1046 (Grease filters for exhaust ducts) purpose for loading primary filters and water mist with grease for data consistency and continuous measurement on grease removal efficiency testing.  Fig. 2. The inlet pressure of working fluid (fresh water& recycled water WW2-WW8) was set to 1.67 bar at water-mist spray flow rate 0.29 l/min per nozzle. The grease loading system includes a chamber (vapor box), burner, dossing pump, oil, and water. An oil and water flow rate was set to 10±0.5 ml/min and 24.5±0.5 ml/min using 2 dossing pump. Oils and water weight are recorded using a digital scale before and after the test to assure the correct amounts of oil and water loaded during the test. A heat source was produced from 3 burner placed at the bottom of the vapor box and was maintained heated at range of 385±14 °C. The temperature of the heat source was determined by 2 chromal-alumel (k-type) thermocouple placed in the vapor box.
The exhaust fan and burner was activated at first 12 minutes to stabilize the flow rate and heat source temperature. The oils and water was loaded right after the temperature stable at range 377°C to 399°C. The test for temperature of the grease loaded was set to 30 minutes sampling time. 30 minutes after sampling, the oils, water, and burner was stop immediately supply to the system. The wastewater from water-mist waste samples was collected inside the strainer to determine the fluid properties. This samples then was delivered to the laboratory to measure the mass and viscosity of the fluids. Viscometer was used to measure the viscosity of the wastewater as in Fig. 3. Mass of the wastewater was measured using analytical balance as in Fig. 4. This two measurement will provide us the data to determine the density of the wastewater samples. Before performing a simulation work, some of the fluid properties details need to be calculated. Table 2 shows the results of wastewater from water-mist properties. The density and viscosity slightly increase cycle by cycle which shows the others wastewater properties are also increase. The molar mass of wastewater increased as the density and viscosity increased.

B. Simulation Setup
In this research, the nozzle geometry was base on 1/8 KJSB 0.5 with flat fan spray type. The nozzle geometry was designed using SolidWorks. Then, the nozzle design was exported to ANSYS software for simulation analysis. The descriptions and the boundary condition used in this simulation analysis as in Table 3. RESULT AND DISCUSSION This research aims to simulate the spray characteristics of the water-mist recirculation system base on water quality from experimental data. There are 4 different types cycle of wastewater, which areWW2, WW4, WW6 and WW8 where WW2 is for water-mist at cycle 2 of recirculation and WW8 is for water-mist at cycle 8 of recirculation.

C. Actual Water-mist Spray Velocity
The water-mist spray outlet velocity was obtains using Bernoulli's equation. Bernoulli's Eqn. 1 and Eqn. 2was applied to determine the water-mist spray velocity at nozzle outlet, v 2 . The spray velocity for each watermist spray activation cycle may differ due to fluid density.   Table 4shows the actual spray velocity results obtained from experiment and calculation above. There are two positions, which areat nozzle inlet at 0.068 m height and at nozzle outlet at 0.052 m height. The distance between nozzle inlet and outlet is 16 mm. The inlet water-mist pressure is set at 1.67 baratfluid velocity 0.39 m/s. The spray velocity at nozzle inlet and outlet decreased as the number of cycle increased.

D. Water-mist Spray Velocity From Simulation
In this simulation, graph pattern for all water and wastewaters are same but slightly different in small margin. Thesprayvelocity at nozzle inletwas 0.391 m/s and growing steadily until reach the bottleneck of nozzle. The spray velocity increased drastically as the water flowing through orifice, 0.5 mm and slowing down as the water spread into the domain or in actual situation is spread into the plenum of kitchen hood. For WW2, the flow pattern shows the same but the overall velocity shows decreasing as the number of cycle of wastewater increase. From the table above, the result for WW2 at distance 2 mm (inlet) the velocity is 0.39m/s, which is 0.01 m/s lower than fresh water. This goes to all other wastewaters, the velocity decrease as the number of cycle of wastewater increases.
The wastewater contains 99% of water and 1% ofFOG. The FOG contains in the wastewater increased as the number of cycle increased. Thus, the viscosity and density of the fluids increased and more viscous as the number of cycle increased.
The highest velocity was recorded at velocity of 10.61 m/s, at distance 18 mm forfresh water. The spray velocity decreasing as the flow reached at distance 18 mm where the spray enters and disperse into the air. The lowest spray velocity was recorded at velocityof 0.291 m/s at distance 20 mm for WW8 as in Fig. 5 Fig. 6 shows the fluids flow steadily from distance 2mm to 13 mm. The pressure started to drop at point 16 mm and drastically drop afterward. The pressure was drop at point 16 mm due to passage that passes through bottleneck of nozzle. The inlet pressure for fresh water is 1.67 bar and it is similar for other cycles. The recorded pressure values almost same but there are slightly different at point 13mm between cycles of WW4 and WW6. The water-mist spray pressure increased between WW4 and WW6 and at start to drop at point 18 mm. This is because the WW6 is denser and more viscous than WW4. Furthermore, to get better understanding, the data at point 16mm shown increases of pressure flow between WW4, WW6 and WW8. As the number of cycle of wastewater increased the pressure generated from the nozzle outlet decreased. Table 5below shows the overall absolute error, E ABS of water-mist spray velocity at nozzle inlet and outlet between experimental and simulation.  Now, lets assume the maximum velocity drop allowed is 10 % from current operating velocity and the other parameters are as in Error! Reference source not found.. To estimate the operating cost, the recirculation system will be compared with current operating condition, which is recommended by manufacturer.

F. Absolute Error, E ABS of Spray Velocity Between Actual and Simulation
It is clearly shows us that, under these conditions, the water-mist hood represent an operating cost for the user. The recirculation water-mist system will reduce the water consumption by 918.74 euro per year and 1319.04 euro per year for Water-mist Hood design 1 and 2.

V.
CONCLUSION The water-mist recirculation system was tested experimentally by comparing the spray velocity between actual and simulation. A simulation work has been done to determine the effects of nozzle spray characteristic by wasted water-mist water spray for each cycle. Both experimental and simulation analysis are important to predict the life cycle of wasted water-mist water, and water consumption along the water-mist activation.
Based on the study using scale KHV system, the water-mist recirculation system has great potential and improvement to the current system operation. By adopting the recirculation system to current water-mist kitchen hood, the operational cost was reduced. From the analysis, the user could save up to 1319.04 euro per year per hood length. It is 95.83 %of water cost reduction for existing water-mist kitchen hood design 1 and 2.