Lecture 2: Energy Audits and Baseline

Energy Modelling
Building Energy Management and Optimization
CIVE 5708

Araz Ashouri PhD

Adjunct Research Professor
Civil & Environmental Engineering
Agenda
• Energy audits and retro-commissioning
• Condition surveys
• Utility data analysis (change point models, peak load
analysis metrics, CUSUM)
• Auditing equipment and tools
• Standards and guidelines
• Matlab example for change point modelling
• Case study
• Criticism to conventional energy audits
Energy audits
• Determine where, when, why and how energy is being used
• Identify opportunities to improve efficiency, decrease
energy costs and reduce GHG emissions
• Verify the effectiveness of energy conservation measures
(ECMs) after they have been implemented

• Auditing is the process to identify ECMs

• When you implement ECMs and fix issues you identified,
this becomes retro-commissioning
Energy audits
• Data collection and review
• On-site surveys and system measurements
• Observation and review of operating practices
• Data analysis
Energy audits
• They can be carried out by external consultants or
by in-house experts on a single equipment,
process, department, plant, or company
Energy audits
• Macro-audit
• high level in the structure of energy-consuming systems
in the entire site or facility
• Micro-audit
• narrower in scope focusing on a single type of
equipment or a single system
Levels of Energy Audits
1. Conduct a condition survey (Level 1)
• assess the general level of repair, housekeeping and operational
practices that have a bearing on energy efficiency and flag
situations that warrant further assessment as the audit progresses
2. Analyze energy consumption and costs (Level 1 to 2)
• Collect, organize, summarize and analyze historical energy billings
and the tariffs that apply to them (often only with utility bills)
3. Compare energy performance (Level 1 to 2)
• Determine energy use indices and compare them internally from
one period to another, from one facility to a similar one within your
organization, from one system to a similar one, or externally to best
practices available within your industry
4. Dig deeper (Level 3)
• Gather system and equipment level performance data
• Build and calibrate detailed BEMs
Typical structure of an energy audit
• Building description
• Envelope
• Visual assessment of elements
• Qualitative measurements (thermal imaging, tests with smoke
sticks)
• Quantitative measurements (Estimation of thermal
conductance; Blower door tests or tracer gas tests for
airtightness)
Typical structure of an energy audit
• Building description
• HVAC and lighting
• List of equipment (AHUs, chillers, boilers)
• Control sequences
• Building use
• Data availability
• Controls and mechanical drawings
• BAS graphics
• Interviews with the operators and the energy managers
Typical structure of an energy audit
• Energy use analysis
• Utility usage analysis
• Typical energy use and comparison to benchmark data
• Historical trends
• Load analysis
• Baseline energy modelling
• CUSUM analysis
 Load analysis

𝑘𝑊ℎ 𝑢𝑠𝑒𝑑 𝑖𝑛 𝑎 𝑝𝑒𝑟𝑖𝑜𝑑

𝐿𝑜𝑎𝑑 𝑓𝑎𝑐𝑡𝑜𝑟 (%) = 𝑥100
𝑃𝑒𝑎𝑘 𝑘𝑊 𝑥 24 ℎ𝑟𝑠. 𝑝𝑒𝑟 𝑑𝑎𝑦 𝑥 𝑁𝑜. 𝑑𝑎𝑦𝑠 𝑖𝑛 𝑝𝑒𝑟𝑖𝑜𝑑
𝑃𝑒𝑎𝑘 𝑘𝑊 𝑢𝑠𝑒𝑑 𝑖𝑛 𝑎 𝑝𝑒𝑟𝑖𝑜𝑑
𝐷𝑒𝑚𝑎𝑛𝑑 𝑓𝑎𝑐𝑡𝑜𝑟 (%) = 𝑥100
𝑀𝑎𝑥𝑖𝑚𝑢𝑚 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑙𝑜𝑎𝑑
𝑀𝑎𝑥. 𝐷𝑒𝑚𝑎𝑛𝑑 𝑜𝑛 𝑃𝑜𝑤𝑒𝑟 𝑆𝑡𝑎𝑡𝑖𝑜𝑛
𝐷𝑖𝑣𝑒𝑟𝑠𝑖𝑡𝑦 𝑓𝑎𝑐𝑡𝑜𝑟 (%) = 𝑥100
𝑆𝑢𝑚 𝑜𝑓 𝐼𝑛𝑑𝑖𝑣𝑖𝑑𝑢𝑎𝑙 𝑀𝑎𝑥. 𝐷𝑒𝑚𝑎𝑛𝑑
Baseline energy modelling

Intercept
(energy use at
Slope = DQ1/DT1
balance temperature)

Balance temperature
CUSUM
cumulative sum
control charts
Auditing equipment and tools
The auditor’s toolbox

• Electric power meter

• Combustion analyzer
• Thermometer
• Illuminance meter
• Infrared thermometer or imaging camera
• Relative humidity
• Airflow meter
• Ultrasonic leak detector
• CO2 sensor
• VOC detector
• Occupancy loggers
Electric power meter
 Combustion analyzer

Boilers
Unit heaters
Hot water heaters
Furnaces
Ovens
Illuminance meter
Example of measuring lux

Source: ASHRAE
Temperature and RH
Measuring surface temperatures

Infrared thermometer Thermal imaging camera

Source: ASHRAE
 Airflow measurements
Differential pressure meter Capture hood
Ultrasonic leak detector
Ultrasonic flow meter
 VOC sensor
CO2 sensor

250-400ppm Normal concentration in outdoor air

VOC: Volatile organic compounds
Concentrations typical of occupied indoor spaces with proper air
400-1,000ppm
exchange

1,000-2,000ppm Complaints of drowsiness and poor air

Headaches, sleepiness, loss of attention, increased heart rate and

2,000-5,000 ppm
slight nausea

5,000 Workplace exposure limit

>40,000 ppm Exposure may lead to permanent brain damage, coma, even death.
Occupancy loggers

PIR – Passive InfraRed

Building Performance Simulation – and use of BIM in operations

EnergyPlus
TRNSYS
IES
ESP-r
eQUEST
Revit + Insight 360
OpenStudio
Standards and guidelines for
building energy audits
Matlab example
clc; clear; close all;

[num,txt,raw] = xlsread('SpaceHeatingCoolingLoads.xlsx','Cooling Load');

tOut = num(:,1);
qClg = num(:,end);

% change point temperature range to be searched

tChpCool = 1:20;

% sweep through the change point temperature

for i = 1:20
regressors = [(tOut > tChpCool(i)).*(tOut - tChpCool(i))];
mdl = fitlm(regressors,qClg);
store(i,1) = mdl.Rsquared.Ordinary; % store the R2 value
end
[ind,ind] = max(store(:,end)); % change point temp

% construct the model for the best change point model

regressors = [(tOut > tChpCool(ind)).*(tOut - tChpCool(ind))];
mdl = fitlm(regressors,qClg);

% generate predicted temperatures for plotting

tOutTest = (-20:30)';
regressors = [(tOutTest > tChpCool(ind)).*(tOutTest - tChpCool(ind))];
fitResults = predict(mdl,regressors);
% create a plot for the selected heating change point model
fig = figure('units','inch','position',[0,0,3,3]);
scatter(tOut,qClg,1,[0.1 0.1 0.1],...
'MarkerFaceAlpha',.3,'MarkerEdgeAlpha',.3)
hold on
plot(tOutTest,fitResults,'k','LineWidth',3);
xlim([-20 30])
xticks([-20:10:30])
set(gca,'TickDir','out');
% Add title and axis labels
xlabel('Outdoor temperature (^{o}C)')
ylabel('Cooling load intensity (W/m^2)')
box off
% Add a legend in the top, left corner
legend('Measured','Modelled','Location', 'northoutside')
legend boxoff
print(fig,'ChangePointCooling.png','-r600','-dpng');
• Limitations of change point models method
• Older buildings were more envelope dependent. The role of
internal gains have increased with better construction
practices – increased airtightness and thermal insulation.
• Internal gains are stochastic – i.e., they are hard to predict as
they vary from one building to another, as a function of the
time of day, day of week, month of year, and from one
thermal zone to another.
• Modern HVAC equipment (chillers and heat pumps) has
drastically different efficiencies at part-load conditions
• Efficiency of modern equipment cannot be treated as
constant, and the balance-point temperature varies diurnally
or seasonally.
• Effect of solar loads is not accounted for properly – which was
fine as older buildings had much smaller window-to-wall
ratios.
• Bin method partly addresses some of these
limitations
• Basic idea behind the bin method is to divide the
solution space into groups (i.e., bins)
• Develop regression models for each bin separately
• e.g., one could develop different models for different
outdoor temperature ranges
• e.g., one could develop different models for occupied
hours and unoccupied hours
• e.g., one could develop different models for sunny and
overcast days
 Bin method – Example 1

Bsch
Bin method – Example 2
Bin method limitations

• How to bin your data (i.e., size of bins and location

of bins)
• e.g., should I fit four models?: sunny and cold, overcast
and cold, sunny and cool, overcast and cool
• e.g., should I fit eight models?: (sunny, overcast) (cold,
cool) (occupied, unoccupied)
• e.g., what is threshold for sunny / overcast (100 W/m2 or
200 W/m2 or 300 W/m2)?

1. Graphically inspect and identify naturally occurring clusters

2. Employ a classification algorithms to automatically determine
the significant variables and their bin size
 Case study
Analysis of the heating and cooling use data of the federal
buildings in Ottawa
Load duration curve
 Change point models with variable balance temperature

• Piecewise line-fits for heating

and cooling load intensities
were built for 35 buildings
• Each piecewise line-fit has
three degrees of freedom
• Balance temperature Tb
• Intercept for the fit x1
• Slope for the fit x2
• We have fitted six
parameters for each building,
three for heating and three
for cooling
Box-whisker plots presenting the distribution of the change point
models of the 35 buildings for heating and cooling load intensity
Solid-lines are for the mean of the 35 buildings
Distribution of the six parameters predicting the heating and
cooling load intensities for each building

• B7, B17, and B24 have cooling load

intercepts greater than 10 W/m2
• These buildings use lots of cooling
regardless of the outdoor temperature
• Likely no economizer was programmed
• B15, B24, B31 have high heating loads
intercepts greater than 20 W/m2
• These buildings use lots of heating
regardless of the outdoor temperature
• A potential controls sequencing
problem
 Distribution of the six parameters predicting the heating and
cooling load intensities for each building

• B21, B24, B31 have the highest slopes • B10, B13, and B17 have the lowest
with respect to cooling balance temperature for cooling
• B21, B24, and B31 have the smallest • B3, B10, and B23 have the highest
slopes with respect to heating balance temperature for heating
• These buildings have the poorest • Most buildings have lower balance
envelope performance temperatures than 18°C
• Cooling balance temperatures in more than 32 out of 35
buildings were lower than heating balance
temperatures
• Meaning that these buildings heat and cool at nearly
identical conditions
• This situation is the worst in B3, B10, and B17
Criticism to conventional energy audits
 Typical commercial building behavior over time
Energy Consumption

Periodic Re-tuning Continuous Re-tuning

Ensures Persistence Maximizes Persistence

Time
S. Katipamula (2012), Lessons learnt from Building Re-tuning Training,
ASHRAE Conf., Jan
Agami Reddy (2017) Big Data in Building Operations Workshop in Ottawa
PMV (predicted mean vote )
+3 = hot
+2 = warm
+1 = slightly warm
0 = neutral
-1 = slightly cool
-2 = cool
-3 = cold
Limitations of conventional energy audits and
retro-commissioning
• Labour intensive and costly
• Persistence of ECMs
• Access to skilled personnel, equipment and tools

We need automatic or semi-automatic data analysis

tools to provide guidance to in-house operations team
Bibliography
• NRCan (2017) Energy Savings Toolbox – An Energy Audit Manual and Tool
http://www.nrcan.gc.ca/energy/efficiency/industry/cipec/5161
• ASHRAE Guideline 14-2014 – Measurement of Energy, Demand and Water Savings
• ASHRAE 2004 – Procedures For Commercial Building Energy Audits
• Mills, E., Bourassa, N., Piette, M. A., Friedman, H., Haasl, T., Powell, T., & Claridge, D. (2005,
May). The cost-effectiveness of commissioning new and existing commercial buildings: Lessons
from 224 buildings. In Proceedings of the National Conference on Building Commissioning.
• N. J. Bourassa, M. A. Piette, and N. Motegi, "Evaluation of retrocommissioning persistence in
large commercial buildings," ed, 2004.
• S. Y. Cho, "The persistence of savings obtained from commissioning of existing buildings," Texas
A&M University, 2002.
• Haberl, Jeff S., D. E. Claridge, and C. Culp. "ASHRAE's guideline 14-2002 for measurement of
energy and demand savings: How to determine what was really saved by the retrofit." (2005).
• Claridge, D., Haberl, J., Liu, M., Houcek, J., & Athar, A. (1994, August). Can you achieve 150% of
predicted retrofit savings? Is it time for recommissioning?. In Proceedings of the 1994 ACEEE
Summer Study (pp. 5-73).