Intro:

The University of Maryland is home to over 800 fume hoods. Roughly 200 of these hoods have the potential to create fume-hood-driven lab exhaust in their respective spaces; 80 hoods in teaching labs, 120 in research labs. To begin a behavior change pilot program with the goal of energy reduction, 4 labs were chosen, nearly identical in size, each with 4 fume hoods, and all on the same corridor of the Chemistry Building.

Lab exhaust represents lost energy through the conditioned air. This is calculated based upon $5 per CFM per year. Fume hood sash height is correlated with exhaust and is used as the key performance indicator for this project (KPI). In a normal operating system of a fume hood driven exhaust, the relationship between sash position and CFM of the exhaust is represented by y=11x+200; for every 1% increase in sash position, there is an increase of 11 CFM of the exhaust with a base exhaust around 200 CFM for a closed hood. This relationship is true for VAV hoods at UMD and was modeled using over 150,000 observations of multiple hoods which were sampled on campus over the course of a year.

Two strategies were implemented at different times to understand the relative impacts of each: Shut the Sash stickers, developed by UC Davis; and a Shut the Sash Competition with $125 worth of gift cards to on-campus dining as an incentive for the winning lab. The stickers were placed on the fume hoods on September 19th. Three weeks later on October 9th, the competition was announced over email to all PI’s and lab workers in the four labs. The competition began immediately. To not have biases for differing research taking place across labs or across time, the KPI of fume hood sash height was only taken during the night hours of 20:00 through 08:00. Secondly, the data was monitored to check for instances when the fume hoods were being used past 20:00. The data was consistently captured by the building automation system (BAS) in 30-minute intervals of average-data for each interval. The lab members were given weekly updates on their position in the competition, including a graph of all labs’ behavior, as well as forecasted annual savings of energy, money, and carbon of their improved behavior relative to business as usual.

Data:

FumeHoodCompFinalGraph2.png

fig. 1 Graph of Results

(open image in new tab to enlarge)

 

Capture

fig. 2 Chart of Results

 

19 Fume Hoods

Percent Reduction from Baseline (each)

CFM Reduction Energy Reduction Financial Reduction Carbon Reduction

During Competition

-31.6%

6,600

36 MWh $3,800 18 tonnes

Ideal Annual Savings

-31.6%

6,600 314 MWh $33,000

157 tonnes

Realistic Annual Savings -20% 4,200 200 MWh $21,000

100 tonnes

fig. 3 Table of Results

 

Figures one and two above represent the same data in different ways. Figure one shows the collection of the four labs, each lab represented by a unique color, while the curved line is the smoothed average for that labs’ four fume hoods. Figure three represents the competition results as well as forecasting savings beyond the competition itself.

The stickers alone resulted in a 10% decrease in average fume hood height in the Research Labs (fig. 2). This represents a maximum potential of 110 CFM of reduced exhaust per hood; almost $550 per year in energy costs. A great return for a $3 sticker. Implemented across all 120 Research Lab fume hoods that have energy savings potential represents a potential savings of $65,000 per year under ideal ventilation.

In Teaching Labs, the data shows that the stickers also had an impact on the behavior in those labs. The pre-sticker average heights for all 80 hoods averages at 13.2%, and after sticker height average of 8.6%. This 4.6% improvement in hood height correlates to a savings of over $20,000 annually for the 80 Teaching Lab hoods. The behavior of teaching labs started better than the faculty hoods, and the impact of the sticker implementation was proportionally lower.

Through the first month post-competition, the improved behavior is holding steady. If the average height plateaus at a 20% reduction from baseline we could expect a savings of $21,000 for the 16 hoods in the competition plus 3 auxiliary hoods in the same labs compared to the baseline (fig. 3). If the competition results were replicated across all 120 faculty hoods it would result in a savings of $150,000 annually (fig. 4).

 

 

Percent Change Compared with Baseline Potential Savings Per Hood Potential Savings All Hoods Energy Savings All Hoods
Sticker impact per hood Research Lab 9.8% $540 $65,000 616 MWh
Sticker impact per hood Teaching Lab 4.6% $250 $20,000 193 MWh
Impact if competition is applied to all faculty hoods 23.7% $1,250 $150,000 1489 MWh

fig. 4 Table of potential results for expansion to the entire campus under ideal ventilation

 

Discussion:

During interviews with researchers across labs, it was a lack of knowledge about the energy-intensity of fume hoods which led to the original poor fume hood behavior. In the weekly emails to the labs in competition, an update on how much less carbon dioxide was being admitted to the atmosphere was included and this resonated with the lab workers more than energy, and much more than money saved to the university. Breaking down the climate impact resolution to a per hood (or per person) impact gives the end user the sense of agency and ability to change their behavior knowing how their impact helps slow climate change.

The competition lasted 6 weeks, which was found to be an appropriate length. After developing the framework in RStudio to quickly produce graphs and analyze the BAS data on a weekly basis, it took approximately 3 to 4 hours per week to maintain this competition.

Each fume hoods’ historical data were analyzed independently and it was found that two hoods’ data had consistent errors, showing +20% height of the actual hood heights. This was identified by plotting a scatterplot of observations, as well as using Tableau (see appendix). This issue was confirmed by communicating with a lab member to identify a time where the hood was closed and check the BAS. These wonky hoods were left out of the analysis.

 

Conclusions:

The Fume Hood Stickers on their own have shown to have a ¼ to ⅓ improvement in behavior here at UMD compared with the baseline. The ROI is quick and effortless, and the stickers should be employed on all hoods resulting in over $80,000 in annual savings alone across 200 hoods. The Fume Hood Competition was very effective and shows that the impact is outliving the competition itself. From interviews, the historic poor fume hood behavior is most likely due to researchers not knowing the energy & carbon impacts of fume hoods. Because of this, it is recommended that there is an annual educational outreach, potentially in the form of posters or flyers delivered to each lab, or posted in the fume hood-intensive buildings.

The competition has a large value on its own, but the lack of access to further data in the BAS inhibits employing this beyond the pilot competition. An annual competition repeating the pilot competition should continue to keep up awareness of fume hood energy impact; there are 16 hoods plus 3 auxiliary hoods representing a large chunk of the target fume hoods.  As a result of this program, the 19 hoods impacted will save the University of Maryland over 200,000 kWh in energy per year and 120 metric tonnes of carbon dioxide entering the atmosphere. Through the replication of these results to all targeted faculty research labs, there is an opportunity to save over 870 metric tonnes in reduced carbon emissions, 1.5 GWh of energy, and $150,000 in cost savings.
Resources: 

To reproduce this competition at your university or institution, feel free to start by using the script in R that I developed with the included sample dataset and ReadMe:

https://github.com/emery123/FumeHoodCompetition

 

Appendix:

Visualizations such as this one made in Tableau are useful for looking at all fume hoods at once to identify abnormalities in the data. For instance, in lab 3356, the hood represented by the orange bar never dips below a certain level and was later confirmed to be pulling inaccurate data.

FHCOmpt (1).png

Assumptions:

Electricity cost: $0.105 per kWh

Carbon cost: 500g per kWh

Exhaust cost: $5 per CFM per year

Numbers may not sum due to rounding and significant digits

Ideal ventilation assumes an air change rate (ACH) of 8 or less per space