Arch. 6241 Building Simulation Design Practice Instructor: Godfried Augenbroe
Energy Plus + EPC Calibration Study Smithgall Student Service Building Presented by: Tianyu Feng
INTRODUCTION Built Year: 1990 Floor area: 3957 m² Major Feature: i> 2 story-tall ii> mainly brick wall, metal roof finish iii> a 2 story tall atrium at NE corner closed with curtain wall iv> mostly office, one auditorium, and one atrium v>: district heating & cooling Ferst Art Center
Student Center
kWh/m2/yr
Simulation result
resource
real
E+ with DC/DH
EPC
E+ with own HVAC
Consumption
157.38
28.11
144.30
144.30
% off to real
0%
82.14%
8.77%
8.37%
kWh/m2/yr
Simulation result
resource
real
E+ with DC/DH
EPC
E+ with own HVAC
Consumption
89.29
89.01
55.20
59.60
% off to real
0%
0.31%
38.18%
8.37%
kWh/m2/yr
Simulation result
resource
real
E+ with DC/DH
EPC
E+ with own HVAC
Consumption
102.74
82.55
121.91
82.56
% off to real
0%
19.64%
18.67%
19.64%
Simulation result Start point •
Real data are quite different with simulation results
•
The difference of heating, cooling and electricity are not proportional, or no reasonable pattern at all
Energy Plus Calibration
I. Methodology
a. Select uncertain input and categorize uncertain level
b. Calibrate 3 times with different level of inputs
Calibration 1
Cooling setpoint T (2) Heating setpoint T (2) Infiltration; People activity People density; Light; Equipment
More uncertain
wall U roof U Win. U
Calibration 2
Calibration 3
II. Result
kWh/m2/yr
Energy Plus Calibration
kWh/m2/yr
Heating Real
E+
Cal1
Cal2
Cal3
157.4
28.1
82.2
125.9
134.0
20.0%
14.9%
0.0%
82.1% 47.8%
kWh/m2/yr
Cooling Real
E+
Cal1
Cal2
Cal3
89.3 0.0%
89.0 0.3%
114.4 28.1%
115.4 29.3%
123.7 38.5%
Electricity Real
E+
Cal1
Cal2
Cal3
102.7 82.5 71.6 0.0% 19.7% 30.3%
96.2 6.3%
79.7 22.4%
Energy Plus Calibration
kWh/m2/yr
II. Result
TOT.
Real
JAN.
FEB.
MAR
APR
MAY
JUN
JUL
E+ with DC/DH
57.9%
56.2%
48.9%
50.2%
37.2%
33.0%
30.6%
19.0% 32.2% 42.0% 53.4% 53.8%
42.9%
Cali. 1
25.6%
15.6%
8.6%
36.3%
31.0%
29.4%
26.2%
13.1% 28.7% 31.1% 18.9% 15.9%
23.3%
Cali. 2
0.1%
10.9%
20.6%
7.6%
8.3%
25.9%
21.1%
10.0% 15.9%
2.8%
8.9%
14.8%
3.4%
Cali. 3
3.2%
16.1%
26.8%
9.8%
12.1%
29.1%
24.6%
12.8% 20.4%
4.8%
12.7% 17.1%
3.4%
AUG
SEP
OCT
NOV
DEC
kWh/m2/yr
EPC Calibration and Comparison
kWh/m2/yr
Heating Real
E+
Cal1
Cal2
157.4 0.0%
143.6 8.8%
65.4 58.4%
129.6 17.7%
Cooling Real
E+
Cal1
Cal2
89.3 0.0%
55.2 38.2%
165.0 84.8%
221.9 148.5%
kWh/m2/yr
Electricity Real
E+
Cal1
Cal2
102.7 0.0%
121.9 18.7%
110.9 8.0%
107.3 4.4%
Total Real 349.4 0.0%
E+ 320.7 8.2%
Cal1 341.3 2.3%
Cal2 458.7 31.3%
Calibration Sensitivity in EPC Question What parameters are most sensitive to the result change in calibration ? a. Stp 1 Htg COP, Ctg COP, Infil., Appliance, Lighting, Occ, wall U, roof U, window U,
example
1. Only change 1 parameter at a time 2. Each must change to same energy usage
Each parameter get a range, which could allows them reach same energy level
1< Htg COP< â&#x2C6;&#x17E;
Reduce Del E by 3%
1< Htg COP< 3.33
1< Clg COP< â&#x2C6;&#x17E;
Reduce Del E by 3%
1< Clg COP< 2.66
b. Stp 2 Input all 9 parameters from step 1
In EPC, USE binary, Limit the amount of variable chosen to reach best possible calibration result e.g. when # of variable is 1, the most sensible para will be chosen
Generate a ranking
Calibration Sensitivity in EPC 1 variable
CLG COP
2 variable
CLG COP
Lighting
3 variable
CLG COP
Lighting
HTG COP
4 variable
CLG COP
Lighting
HTG COP
Appliance
5 variable
CLG COP
Lighting
HTG COP
Appliance
What if we set the range each parameter in a more realistic condition CLG COP Lighting HTGof COP Appliance roof U Infiltration with different capacities in influencing the results? 7 variable CLG COP Lighting HTG COP Appliance roof U Infiltration Window U 6 variable
8 variable 91variable variable
CLG COP CLG CLG COP COP
Lighting Lighting
2 variable
CLG COP
HTG COP
3 variable
CLG COP
HTG COP
Appliance
4 variable
CLG COP
HTG COP
Appliance Lighting
5 variable
CLG COP
HTG COP
Appliance Lighting
Infiltration
6 variable
CLG COP
HTG COP
Appliance Lighting
Infiltration
7 variable
CLG COP
HTG COP
Appliance Lighting
8 variable
CLG COP
HTG COP
Appliance Lighting
HTG COP HTG COP
Appliance Appliance
roof U roof U
Infiltration Infiltration
Wall U
Wall U
Swing Parameters Infiltration Wall U Infiltration
Window U Window U
Wall U
Window U Window U Roof U
Occupancy
Calibration Sensitivity in EPC Limitation • Other parameters in this model in for Smithgall, and it can not be used by most buildings. • The calibrated parameter range may not be applied to other projects due the some additional conditions • The result is about QUALITY (ranking) not QUANTITY. • The weather/ climate zone is also an very important factor in this case. Conclusion • This study has been limited to very specific condition, and most likely won’t conclude certain type of situation •
but at least, it still shows some level of sensitivity of these uncertain parameter
kWh/m2/yr
kWh/m2/yr
kWh/m2/yr
Problem faced
â&#x20AC;˘ Before start calibrating, check the input and original model IN DETAILS â&#x20AC;˘ If the original data has more than 10 times, fundamental problems might exist
Conclusion • The data from meter may not always be accurate • More calibrated input parameters doesn’t mean more accurate result • A reasonable range for each calibrated input parameter is important to the result • The resolution we want could determines the accuracy level of the same calibration result. = if (target is del. E =1, combine htg+ clg +ele, analyze each E) = if (target is utility bill =1, calibrate with unit price, consider Es)