demand-controlled ventilation for commercial kitchens

7/30/2019 Demand-Controlled Ventilation for Commercial Kitchens

1/13

3 6 A S H R A E J o u r n a l a s h r a e . o r g N o v e m b e r 2 0 1 2

Food service facilities have high energy consumption with equipment

and commercial kitchen ventilation (CKV) being the primary energy

consumers in a restaurant. Exhaust hood airow drives HVAC energy

consumption for CKV, so the rst step in reducing this exhaust airow is

designing high efciency hoods with low capture and containment (C&C)

airow rates. The next step is using demand control ventilation (DCV) to

further reduce exhaust airow when cooking is not taking place under

the hood, but when appliances are hot and ready for food preparation.

Airow reduction is not the sole ob-

jective o a DCV system; it also must

ensure exhaust airow and the corre-

sponding supply airows are increased

to C&C levels as soon as cooking starts

(to avoid spillage o convective heat and

cooking euent into the kitchen space).

The current NFPA-96 Standard1 and In-

ternational Mechanical Code

2

require

that a hood operate at ull design air-

ows whenever ull-load cooking activ-

ity occurs underneath an exhaust hood.

DCV has evolved rom simple two-

speed an control systems to proportion-

al control with variable requency drives

(VFDs) based on exhaust temperature.

This improvement allowed or varying

airows throughout the day. Then, an

optical sensor was added to the temper-

ature-based control to detect cooking

activity taking place under the exhaust

hood to urther enhance perormance.

The latest system introduced to the

market added measurement o exhaust

airow and automated balancing o mul-

tiple exhaust hoods connected to a single

an (or a dedicated an) and modulation

o replacement air or the space. Future

systems need to be designed to consider

the entire kitchen status to maximize en-ergy savings.

Laboratory testing was conducted or

common appliances in the commercial

kitchen to evaluate system perormance

when equipped with various DCV al-

gorithms: operating at a fxed exhaust

About the Authors

Derek Schrock is U.S. research director and Jimmy

Sandusky is research engineer at Halton Company

in Scottsville, Ky. Andrey Livchak, Ph.D. is direc-

tor of global research and development at Halton

Group Americas in Bowling Green, Ky.

By Derek Schrock, Member ASHRAE; Jimmy Sandusky, Associate Member ASHRAE; Andrey Livchak, Ph.D., Member ASHRAE

Demand-Controlled Ventilation

For Commercial KitchensT

etraImages/Corbis

This article was published in ASHRAE Journal, November 2012. Copyright 2012 ASHRAE. Posted at www.ashrae.org. This article may not be copied and/or

distributed electronically or in paper orm without permission o ASHRAE. For more inormation about ASHRAE Journal, visit www.ashrae.org.


2/13

N o v e m b e r 2 0 1 2 A S H R A E J o u r n a l 3 7

setpoint temperature, operating on a temperature curve to in-

crease exhaust airow proportional to the temperature dier-

ence between exhaust and space temperature and operating on

a temperature curve in combination with a cooking activity

sensor (CAS) to drive the system to design when cooking is

detected. Additionally, an evaluation was done to determineenergy savings or a DCV system with balancing dampers in-

stalled on a our exhaust hood, island confguration.

Test Setup

The objective o the frst round o tests was to compare per-

ormance o DCV systems that use temperature sensors only

to those that incorporate cooking activity and temperature sen-

sors. Currently, only two manuacturers oer the latter. One

design uses optical opacity sensors to detect the presence o

cooking euent in a hood cavity. Another design uses inrared

(IR) temperature sensors to monitor the surace temperature

o cooking appliances. Data rom these IR sensors along with

space temperature and hood exhaust temperature sensors are

analyzed to interpret the status o cooking appliances (idle,

cooking or o) and adjust hood exhaust airow accordingly.

A 72 in. (1.8 m) long wall canopy exhaust hood was confg-

ured to simulate various DCV control algorithms available on

the market: exhaust temperature-based system that operates at

a fxed setpoint, exhaust temperature-based system that oper-

ates on a curve and exhaust temperature coupled with a cook-

ing activity sensor system (includes IR sensors).

Fixed setpoint exhaust temperatures o 90F, 100F and

130F (32C, 38C and 54C) were evaluated. For these con-

fgurations, the minimum exhaust airow rate was 80% o de-

sign airow rate (a common value or temperature only basedsystems due to the limited ability to detect when cooking

starts and ramp-up o exhaust airow). Exhaust airow was

varied by a VFD in an attempt to maintain the tested tempera-

ture setpoint.

For exhaust temperature systems that operated on a curve,

the minimum exhaust airow rate was again 80% o design.

When using the curve, exhaust airow was incrementally in-

the timer expiration, i no new cooking activity is detected, the

system returns to the curve control algorithm.

The exhaust hood was installed at 80 in. (2 m) above fn-

ished oor with a temperature sensor mounted in the exhaust

collar. The inrared sensors were positioned in the ront, in-

terior ace o the canopy to sense the cooking surace. The

temperature sensor was installed so that it was centered in the

hood collar.

The test protocol included a range o appliances that all

demonstrated similar trends, but due to space limitations only

data or appliances most commonly seen in kitchens is pre-

sented: a charbroiler, griddle and open-vat ryer.

Airows in Table 1 represent hood C&C airow, and when-

ever the hood operates below this value when cooking occurs,

the hood is spilling. Details o appliance uel source and prod-

uct cooked are shown as well.

During testing, each combination was evaluated at the idle

and cooking states. Exhaust airow rate and temperature wereplotted versus time. The onset o the cooking process was not-

ed to determine system response time.

Results and Discussions

Charbroiler

Figure 1 summarizes testing conducted with the charbroiler,

which exhibited the highest exhaust temperatures o all tested

Table 1: Cooking appliance and associated food product.

ApplianceFuel

SourceLoading

Design/Capture

And Containment

Airfow

600F

Charbroiler

Natural

Gas

Frozen

Hamburger Patties1,800 cm

400F Griddle,

Thermostatically

Controlled

Natural

Gas

Frozen

Hamburger Patties1,000 cm

350F Open-Vat

Fryer, Single Vat

Natural

Gas

Frozen

French Fries1,000 cm

2,000

1,900

1,800

1,700

1,600

1,500

1,400

150

145

140

135130

125

120

115

110

105

100

Airfow(cm)

Temperature(F

)

10:04:00a

10:07:22a

10:10:34a

10:13:16a

10:16:41a

10:19:25a

10:22:04a

10:24:42a

10:27:17a

10:29:50a

10:32:24a

10:34:59a

10:37:35a

10:40:11a

10:42:50a

10:46:25a

10:48:58a

10:51:33a

10:54:11a

10:56:51a

10:59:29a

11:02:06a

11:04:47a

11:07:25a

11:10:03a

11:12:43a

11:15:20a

11:17:58a

11:20:37a

11:23:16a

11:25:55a

11:28:35a

11:31:17a

11:34:01a

Temperature+ CAS

TemperatureCurve

Patties On

ConstantTemperature,

SP = 90F

ConstantTemperature,SP = 100F


Airfow Duct TemperatureDesign Airfow

Figure 1: Charbroiler testing.

creased as the temperature dierence

between exhaust and kitchen space in-

creased. This algorithm ensures C&C

o convective heat rom appliances in-

stalled under the hood.Minimum exhaust airow or the

system with cooking activity sensors

installed was limited to 40% o the de-

sign rate to ensure that the exhaust ans

were operated in their recommended

range. This system used the curve

temperature control as described previ-

ously when appliances are in idle mode

and transitioned to design exhaust air-

ow or an adjustable period (which was

set to seven minutes or this test) upon

detection o cooking activity. Following


3/13

3 8 A S H R A E J o u r n a l N o v e m b e r 2 0 1 2

appliances. With the exhaust temperature and cooking activity

sensor (CAS) algorithm, there was a small dierence (approxi-

mately 5%) in idle and cooking exhaust airow rates due to the

elevated exhaust temperature. The cooking activity sensor was

able to detect placement o patties on the cooking surace and

orce the system to design exhaust airow, albeit a small change.Following expiration o the cooking timer at approximately

10:13 a.m., the airow rate dropped, but was orced back to

design value as the exhaust temperature rose. With the tested

algorithm, when the exhaust temperature exceeded 120F

(49C), the system went to design exhaust airow (regardless

o the cooking activity sensor values signal) as this was the

upper limit o the temperature curve.

Comparable results were seen with the temperature only

system operating on a curve and constant temperature systems

with a setpoint o 90F and 100F (32C and 38C). Due to an

elevated exhaust temperature, the exhaust airow remained at

or near the design airow level or the entirety o the test due

to the temperature dierence threshold o the operating curve

being exceeded. When operated on the temperature curve,

there was a small decrease in exhaust airow at approximately

10:30 a.m. due to a decrease in exhaust temperature.

When the setpoint or the constant temperature system was

changed to 130F (54C) the exhaust airow dropped to 80%

o design and remained there or the duration o the test as

exhaust temperature did not exceed the setpoint. There was

one exception where the exhaust temperature peaked at 134F

(57C). Being 4F (2C) above the threshold, this value was

within the deadband o the system. Had the temperature con-

tinued to rise, the exhaust airow would have increased pro-portionally in an attempt to maintain a constant exhaust tem-

perature.

It should be noted that with the charbroiler, the exhaust tem-

perature rose steadily throughout the test as buildup on the

cooking surace occurred. I the test had lasted longer, the ex-

haust airow would have likely increased to design airows

or all tested systems.

Griddle

The results o the griddle testing are summarized in Fig-

ure 2. The exhaust temperature with the CAS algorithm oper-

ated at approximately 70% o design airow at idle. The ex-

haust airow was greater than the minimum setpoint o 40%

to maintain C&C o heat generated by the appliance. When

hamburger patties were placed on the surace, a decrease in

surace temperature was noted as a cooking signal, and airow

was driven to design almost immediately, and ell back to idle

ater the cooking timer expired.

Advertisement formerly in this space. Advertisement formerly in this space.


4/13

Advertisement formerly in this space.


5/13


This test showed that the constant temperature setpoints o

90F and 130F (32C and 54C) were not suitable or the ap-

plication. With a 90F (32C) setpoint the system remained at

design airow at all times due to the setpoint being lower than

the exhaust temperature, even at idle conditions. A site confg-

ured to operate in this manner was no dierent than having a

standard canopy hood without DCV. The opposite was true o

the 130F (54C) setpoint. The system remained at idle airow

as the exhaust temperature never exceeded the setpoint, result-

ing in both heat and cooking euent spilling to the space.

Open-Vat Fryer

Figure 3 shows data taken or the open-vat ryer. When

testing with the CAS, the idle airow rate uctuated between

For an exhaust temperature-based sys-

tem operating on a curve, exhaust airow

increased to the minimum idle rate o

80% o design. Ater patties were placed

on the griddle, the airow increased grad-

ually with the exhaust temperature andthen ell as the temperature decreased

near the end o the cooking cycle. Note

that airow was not at design or the en-

tirety o the cooking process.

When operating with a constant tem-

perature setpoint o 100F (38C), the

system was able to reach and maintain

design airow ollowing placing patties

on the cooking surace. However, the

initial response was not as ast as ob-

tained with the CAS.

Figure 2: Griddle testing.

1,200

1,100

1,000

900

800

700

600

140

130

120

110

100

90

80

70

60

Airfow(cm)

Temperature(F)

1:02:00p

1:06:55p

1:11:52p

1:17:04p

1:22:18p

1:27:34p

1:32:55p

1:38:21p

1:43:53p

1:49:22p

1:54:50p

2:00:13p

2:05:36p

2:10:49p

2:15:49p

2:20:44p

2:25:33p

2:30:12p

2:34:45p

2:39:16p

2:43:50p

2:48:18p

2:52:48p

2:57:19p

3:01:50p

3:06:20p

Temperature+ CAS

TemperatureCurve

Patties On


SP = 90F






6/13



7/13


40% and 50% o design as the appli-

ance fred to maintain oil temperature

in the vat. Upon dropping the baskets

o ries into the oil, the drop in temper-

ature detected by the sensor indicated

cooking was occurring and drove theexhaust an to design airow and then

returned to idle ater the cooking timer

expired.

With the algorithm operating on a

temperature curve, the system idled at

the minimum prescribed 80% o design

airow and increased proportionally as

the exhaust temperature rose. The tem-

perature dierence between exhaust and

space did not exceed the upper limit to

go to design airow.

With the constant temperature setpoint confgurations, all

systems idle at 80% o design. When the setpoint was 90F

(32C), the system did reach design airow ater the initia-

tion o the cooking process, but a signifcant time lag between

onset o cooking and meeting design airow was noted. For

100F and 130F (38C and 54C) confgurations, the system

remained idle at all times due to exhaust temperature not ex-

ceeding the setpoint. This was indicative o the requirement

to be very amiliar with the given cooking process when con-

fguring temperature-based DCV systems. For example, the

open-vat ryer potentially could perorm better with a lower

temperature setpoint value, but common practice dictates

90F (32C) and above setpoints to achieve any measurable

airow reductions.

Figure 3: Open-vat fryer testing.

1,200

1,100

1,000

900

800

700600

500

400

300

200

120

110

100

90

80

70

60

Airfow

(cm)

Tempera

ture(F)

2:58:00p

3:01:23p

3:04:37p

3:07:57p

3:11:21p

3:14:48p

3:18:11p

3:21:37p

3:25:01p

3:28:25p

3:31:45p

3:35:04p

3:38:19p

3:41:36p

3:44:52p

3:48:01p

3:51:12p

3:54:25p

3:57:39p

4:00:51p

4:04:06p

4:07:22p

4:10:39p

4:13:55p

4:17:16p

4:20:36p

4:23:58p

4:27:22p

4:30:46p

4:34:12p

4:37:36p

Baskets Dropped

Temperature

Curve


SP = 90F



Temperature

+ CAS




8/13



9/13


Response Time

In addition to evaluating

the exhaust airow achieved

with dierent types o DCV

systems, the response times

o the systems were comparedas shown in Table 2. An entry

o N/A indicates that the

system either was not able to

meet or remained at design

airow during the test. Note

or tested appliances. When using only the temperature sensor,

system response is more dependent on the correct selection set-

points and is always delayed. I improperly confgured, temper-

ature-based systems can operate at either idle or design airow

at all times, resulting in loss o C&C or no energy savings.

Each appliance shown has dierent exhaust temperatures

that represent the transition rom idle to cooking states. Rare-

ly, appliances are confgured so that each has a dedicated ex-

haust hood; the mixed lineup under a long hood is typical.

Appliance lineups will vary rom site to site, making a generic

temperature curve or setpoint nearly impossible to obtain.

Some confgurations can perorm well without the CAS.

For example, convection or conveyor ovens produce little to

the response time o the systems with the CAS was substan-

tially aster than the temperature-based systems. This is be-

cause the sensor actively monitors what is occurring at the ap-

pliance level, and can indicate when cooking occurs as soon as

the product is placed in/on the appliance and drive the exhaust

an to design airow or the entirety o the process; whereas,

a temperature-based system can only react to a by-product o

the cooking process: a change in temperature that takes longer

to observe.

Discussion

The inclusion o the cooking activity sensor ensures the sys-

tem goes to design airow at the onset o the cooking process

Table 2: Response time comparison.

Time From Start o Cooking (Seconds) When Design Airfow Reached

Appliance Temperature

+ Cooking

Activity Sensor

Temperature

Only Curve

Constant

Temperature,

SP=90F

Constant

Temperature,

SP=100F

Constant

Temperature,

SP=130F

Charbroiler 23 N/A N/A N/A N/A

Griddle 35 174 N/A 181 N/A

Open-Vat Fryer 23 N/A 297 N/A N/A



10/13



11/13

4 6 A S H R A E J o u r n a l a s h r a e . o r g N o v e m b e r 2 0 1 2

no smoke during the cooking process, which would be di-

fcult or an optical opacity or inrared sensor to interpret. In

some installations, a combination o temperature-based and

temperature/CAS-based systems may be necessary or the

system to perorm at an optimal level.

Space temperature will vary during cooling and heating pe-riods, aecting exhaust temperature as well. As a result the ex-

haust temperature setpoint must be adjusted rom one season

to another or DCV systems with exhaust temperature sen-

sors only. A space temperature sensor installed in the area to

constantly evaluate the dierence between exhaust and room

temperature maximizes energy savings potential o the DCV

system.

A an speed profle or each confguration was generated

to compare energy savings, rather than using average percent

reduction alone. These profles were used in conjunction with

an outdoor air load calculator to determine savings associated

with makeup air cooling and heating, as well as exhaust and

supply an energy. Table 3 compares the annual energy sav-

ings associated with both confgurations.

Although both confgurations save energy, the installation o

balancing dampers maximizes these savings by allowing thehoods to operate independently. Without dampers, when one

hood is in the cooking state, all are orced to design airow

regardless o state. The value o the balancing damper lies in

the ability to lower the airows or hoods that are not cooking

in single exhaust an, multiple exhaust hood confgurations.

In this particular case with our hoods connected to a single

exhaust an, the DCV system with balancing dampers saves

nearly twice as much energy when compared to a similar sys-

tem without balancing dampers. This is a conservative esti-

mate because additional savings when all hoods are in idle

mode are not accounted or. Indeed, when the DCV system

is in idle mode (appliances are hot, but no cooking occurs),

Case Study: DCV With and Without Balanc-

ing Dampers

For installations where the exhaust hoods each

have a dedicated exhaust an, balancing dampers

are not needed since the airows can be modu-

lated by changing the an speed. However, when

multiple exhaust hoods are connected to a single

exhaust an, balancing dampers listed in accor-

dance with the UL 710 standard can be installed

on each exhaust hood section to maximize the

energy savings o a DCV system. This is because

each hood needs to have the ability to indepen-

dently regulate airow based on its state (o,

idle or cooking). To illustrate the energy savings

that can be achieved with dampers installed, a

site is evaluated with both confgurations.

The examined site is in Seattle, and is a 24/7operation. The only time the kitchen exhaust

hoods are shut down is or a daily water-wash

operation (approximately 15 minutes). The ex-Figure 5: Case study with balancing dampers installed.

120

100

80

60

40

20

0

ExhaustFanSpeed(%)

Average Airfow = 73% o Design

5/13/12 5/14 5/15 5/16 5/17 5/18 5/19 5/20 5/21 5/22

Exhaust Fan VFD Speed, %

Figure 4: Examined site exhaust hood conguration.

haust hoods are installed as back-to-back island style canopy

hoods and are connected to a single exhaust an (Figure 4).

Each hood is ftted with a balancing damper at the exhaust col-

lar. The DCV system operates with cooking activity sensors

installed on all hoods. The design airow or the site is 11,290

cm (5328 L/s). Figure 4 shows monitored data or exhaust

an speed. On average, the exhaust airow rate was 73% o

design. Figure 5 shows the system rarely operated close to

design airows because the our hoods did not have cookingoccurring at the same time.

Figure 6shows the exhaust an speed or the same DCV sys-

tem and time period without the dampers installed. To model

the system without dampers installed, hood status was also

monitored with the an speed and exhaust airow data. These

ags are generated by the control algorithm based on the inputs

rom the cooking activity, space and duct temperature sensor.

I one o the our hoods was in cooking state, that an speed

would increase to 100% to reach design airow or the par-

ticular exhaust hood. I status was o or idle or all hoods,

an speed was assumed to be equal to that o the system with

balancing dampers.


12/13



13/13

4 8 A S H R A E J o u r n a l a s h r a e o r g N o v e m b e r 2 0 1 2

and the exhaust airow is controlled based on a

hoods exhaust temperature (more accurately the

temperature dierence between hood exhaust

and space temperature) a dilemma is revealed:

which exhaust temperature (or hood) should be

used as a control signal or DCV without damp-ers?

The hood with the highest exhaust temperature

would be the saest bet, but this would require

a more sophisticated control algorithm (not the

case or many DCV suppliers) and will still end

up with a higher total exhaust airow compared

to DCV with dampers. In some cases, a fxed

leading hood is assigned, and its exhaust tem-

perature is used to control exhaust airow or the

whole system in DCV systems without dampers.

Future of DCV Systems

As evidence shows, the cooking activity sensor

is an important component o an efcient DCV

system. However, this is not the most eective

way to identiy appliance status. Taking a signal

directly rom the cooking appliance is a more e-

ective way to detect appliance status (cooking,

idle or o). Most modern cooking appliances are

equipped with programmable logic controllers

Compared to DCV systems with cooking activity sensors,

systems that use only temperature sensors can have signifcant

lags in response time; more than two minutes in the evaluated

cases o the open-vat ryer and griddle. Not detecting cooking

in a timely manner results in loss o C&C, which allows heatand cooking euent to spill to the kitchen space. Any sav-

ings associated with an energy can quickly be oset by an

increased load on cooling and heating equipment.

When temperature only systems are used, setpoints must

be calibrated or a given application (appliance combination).

This is key to ensuring the system operates as intended. In-

appropriate setpoints can result in hoods that run at design

airow constantly (an expensive exhaust-only hood) or idle

continuously (allowing spillage to occur). The setpoints also

should be reset or winter and summer to account or variation

in kitchen space temperature, unless a space temperature sen-

sor is used or automatic reset.Using automatic balancing dampers listed per UL Standard

710 or DCV systems with multiple hoods connected to a single

exhaust an signifcantly improves system energy efciency. The

energy savings or a our-hood system can be double when com-

pared to an identical DCV system without balancing dampers.

References

1. NFPA. 2011. NFPA Standard 96-2011, Standard for Ventilation

Control and Fire Protection of Commercial Cooking Operations.

National Fire Protection Association.

2. ICC. 2012. 2012 International Mechanical Code. International

Code Council.

Figure 6: Case study without balancing dampers installed.

120

100

80

60

40

20

0ExhaustFanSpeed(%)

Average Airfow = 86% o Design

5/13/12 5/14 5/15 5/16 5/17 5/18 5/19 5/20 5/21 5/22

Exhaust Fan VFD Speed Without Dampers

Table 3: Energy savings comparison with and without balancing dampers.

Estimated Savings

SystemHeating

(Therms)

Cooling

(kWh)

Exhaust Fan

(kWh)

Supply Fan

(kWh)

DCV With

Dampers1,133 6,435 32,554 10,851

DCV Without

Dampers623 3,539 15,697 5,232

Dierence 510 2,896 16,857 5,619

(PLCs) that already know appliance status. Establishing com-

munication between the appliance and DCV controller is all

that is needed.

As noted previously, cooking equipment and CKV are a

kitchens primary energy consumers. The term demand con-trol ventilation implies that hood exhaust is modulated based

on demand by cooking appliances under the hood. Cooking

appliances defne overall kitchen energy consumption because

CKV energy consumption is, to a large extent, driven by ap-

pliances being used, and their status defning DCV exhaust

airow. However, DCV doesnt optimize the energy consump-

tion o the source: the cooking equipment.

The next step in the development o an energy-efcient

kitchen is implementing a demand-controlled kitchen (DCK)

strategy, where appliances are controlled based on cooking

demand and communicate their status to DCV to minimize

CKV energy consumption. Indeed, how many times have youseen a range with all burners on and no pots on it or a triple-

stack conveyer oven with all stacks on and just one conveyer

being used? When we implement a DCK strategy with energy-

efcient cooking appliances that are integrated with a DCV

system (controlled based on cooking schedule and demand),

we will have a truly energy-efcient kitchen.

Conclusions

Commercial kitchen DCV systems can oer great energy

savings to the end user when properly implemented. Care

should be taken to ensure the proper DCV system and sensor

type are selected or a given appliance lineup.

demand-controlled ventilation for commercial kitchens

Documents