2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and 25 February 2016 1 Authors Liam McLaughlin Vilnis Vesma Luis Marques Almanza Acknowledge support of the Austrian Energy Agency and Marco Matteini of UNIDO 2
125
Embed
Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
2/18/2016
1
Energy Performance Measurement
and indicators
Liam McLaughlin
Luis Marques Almanza
Tehran, UNIDO office
24 and 25 February 2016
1
Authors
Liam McLaughlin
Vilnis Vesma
Luis Marques Almanza
Acknowledge support of
the Austrian Energy Agency
and Marco Matteini of UNIDO
2
2/18/2016
2
Scope of the training
Focus is individual organisations
• Industry
• Large Buildings
• Public sector
Not dealing the policy level
• National EE
• Sectoral EE
• Sectoral benchmarking
3
WHY ARE WE HERE?
4
1
2/18/2016
3
Why are we here?
Purpose Importance
Stop climate change
Reduce energy cost
Use less energy
Improve energy performance
“My boss sent me”
Other
5
Requires performance improvement
• Only ISO management system standard that requires this
Requires monitoring of performance
Is that what organisations require?
Role of ISO 50006 (Energy Performance Indicators - EnPIs
and Energy Baselines - EnBs)
6
Energy Management System - ISO 50001
2/18/2016
4
How do you measure energy
performance now? Actual cost compared with budget?
kWh last month compared with the same month last year?
kWh/m2 compared with another facility
kWh/unit of production
Moving total of 12 months kWh
More complex method
7
Is this good performance?
8
kW
h o
f N
atu
ral G
as p
er
year
2/18/2016
5
Objective support for decision making
• Often subjective reasons
How much energy we are consuming?
Is consumption increasing or decreasing?
Is performance improving or not?
• Energy Performance indicators (EnPIs) & Baselines (EnB)
Are we meeting targets?
Can we verify savings from improvements?
Are we meeting budgets?
How to allocate costs
Purpose of energy metrics
9
Performance measurement options 1. M&V of a project or operational improvement
2. Critical operating parameter showing effect of
an operational change
• For example, combustion analysis results
3. Observation after awareness training
• For example, number of PCs switched off
4. Normalised whole facility indicator
5. Other normalised indicators
The last 2 are the main focus of this training10
2/18/2016
6
11
EnMS - ISO 50001 simplified
kWh(€ + CO2)
Commit to
change
Plan the
changes
Make the
changes
Check the
results
12
EnMS – conceptual cycle
BUILD COMMITMENT
Decision making and
support
Reporting
DEVELOP INFORMATION
AND PLANS
Opportunity list and action
plan
Technical audits and
operation control review
Collect energy bills and
sub-meter data
Analyze past and present
energy consumption
Forecasting: Targets and
budgets
Develop (or review)
baselines and EnPis
Identify and quantify SEUs
Identify relevant variables
and collect past data
IMPLEMENTATION
Procurement and Design
Operational control
Training
Implement action plan
CHECKING
Verify results of action plan
Compare actual and target
(or expected) consumption
Investigate and correct
significant deviationskWh(€ + CO2)
2/18/2016
7
13
EnMS – conceptual cycle
BUILD COMMITMENT
Decision making and
support
Reporting
DEVELOP INFORMATION
AND PLANS
Opportunity list and action
plan
Technical audits and
operation control review
Collect energy bills and
sub-meter data
Analyze past and present
energy consumption
Forecasting: Targets and
budgets
Develop (or review)
baselines and EnPis
Identify and quantify SEUs
Identify relevant variables
and collect past data
IMPLEMENTATION
Procurement and Design
Operational control
Training
Implement action plan
CHECKING
Verify results of action plan
Compare actual and target
(or expected) consumption
Investigate and correct
significant deviationskWh(€ + CO2)
Terminology
Energy use v energy consumption
Energy driver, factor, relevant variable, independent
variable
Expected energy consumption
Energy performance, saving, efficiency, conservation
Energy Budget
14
2/18/2016
8
Related topics
EnMS (ISO 50001)
EnMS as a metering system
Monitoring and Targeting (M&T)
Measurement and Verification (M&V)
Building Management System (BMS)
Building Energy Management System (BEMS)
15
Relevant standards
ISO 50001 (EnMS) and ISO 50004 (EnMS Guidance)
ISO 50006 (Baselines and EnPIs)
ISO 50015 (M&V of an organisation’s savings)
ISO 17747 (Calculating energy savings)
UNIDO Energy Management Capacity Building Program
16
2/18/2016
9
Is it easy to improve?
17
Discussion
What is your previous experience or views on
energy performance measurement?
18
2/18/2016
10
BUILD COMMITMENT
19
DELUSIONS AND BARRIERS
(TO IMPROVEMENT)
Build Commitment
20
2
2/18/2016
11
21
EnMS – conceptual cycle
BUILD COMMITMENT
Decision making and
support
Reporting
DEVELOP INFORMATION
AND PLANS
Opportunity list and action
plan
Technical audits and
operation control review
Collect energy bills and
sub-meter data
Analyze past and present
energy consumption
Forecasting: Targets and
budgets
Develop (or review)
baselines and EnPis
Identify and quantify SEUs
Identify relevant variables
and collect past data
IMPLEMENTATION
Procurement and Design
Operational control
Training
Implement action plan
CHECKING
Verify results of action plan
Compare actual and target
(or expected) consumption
Investigate and correct
significant deviationskWh(€ + CO2)
"How many managers have been told by their
staff that bad coal consumption was due to low
output? How is it possible for them to judge
whether this is an excuse or a reason?”
These are the opening words from a fuel efficiency bulletin, published in
1943 by the Ministry of Fuel and Power, which criticises the "ton of coal per
ton of output" metric as a misleading indicator of fuel efficiency.
The author was Oliver Lyle, managing director of the eponymous sugar
refinery, a very knowledgeable and eminent engineer who had no time
whatever for the Specific Energy Ratio. Any works engineer today will know
that SERs vary continuously for reasons nothing to do with energy efficiency.
22
2/18/2016
12
Typical example
23
Which was the worse energy
performance?
Foundry industry
WorseBetter
Typical example
24
0.805
0.870
Foundry industry
2/18/2016
13
Energy per unit of production
100
126
99 95
0
20
40
60
80
100
120
140
2008 2009 2010 2011
25
Car assembly industry
Which is right?
26
-16.74 %
+2.19 %
-8.94 %
Brewing industry
2/18/2016
14
Uses for SEC
Cost allocation
Legal or corporate compliance
Benchmarking (?)
Not useful for energy performance measurement
• Except if negligible baseload and only one relevant
variable
27
Uses for absolute energy trends
Annualised view is good for setting future budgets
Annualised view is good for monitoring spending
against budget
Good overview
Not useful for energy performance measurement
• Except if no relevant variables
28
2/18/2016
15
Uses for normalised models
Typically the only effective way to know if
performance is improving or not
Are targets being met
Whole facility
Individual SEUs, buildings, departments
29
30
Basic terminology
Energy performance indicator
Energy Baseline
Energy Target
Energy Improvement
Refe
rence E
nP
Ivalu
e
(baselin
e p
erio
d)
Curr
ent
EnP
Ivalu
e
(report
ing p
erio
d)
Energy Baseline
Energy Target
Target
Achieved !
Actual value
Source: Adapted from ISO 50006
2/18/2016
16
Energy performance indicators
31
EnPI type Example Problems
Measured energy value Annualised
consumption
Misleading results
Do not use variables
Do not measure energy efficiency
Ratio kWh per unit Misleading results
Does not account for baseload
and non-linear effects
Regression Y=mX+C
X: variable value
C: baseload
Complex if it is not linear
Uncertainty
Must be maintained and adjusted
Engineering model Energy simulation Complex
Must be maintained and adjusted
Source: ISO 50006
Energy performance indicators: Criteria
Attributes we need:
• Only responds to changes in energy performance
• Unaffected by weather, production outputs or other
relevant variables
• Direction and magnitude of change consistent with
change of performance
32
2/18/2016
17
33
But …
“MWh/tonne” does not meet these criteria
Two or more relevant variables? We cannot even
calculate such a ratio
ISO 50004 advises against Specific energy consumption
unless there is no baseload and only one variable
• Has anyone an example of such an organisation?
34
Variable
kWh
Using SEC shows
not only non-precise results (YELLOW)
but usually contrary results (RED)
REGRESSION
TREND
SEC: kWh / UNIT
baseload
baseload
Regression vs SEC
2/18/2016
18
Is it easy to show improvement?
35
Where are we going?
36
But…Are we really improving?
How can we be sure??
YES!!!
I’ll show you!!
We want to:
• Develop a model for expected performance.
• Compare actual with expected
• Quantify performance, +ve or -ve
• React to deviations
• Communicate to build commitment
2/18/2016
19
Discussion
Is Specific energy consumption (SEC)
useful?
37
MANAGEMENT INFORMATION
Build Commitment
38
3
2/18/2016
20
Management Information
Communication
Support
Commitment
Decision making
Reporting
39
DEVELOP INFORMATION
AND PLANS
40
2/18/2016
21
PAST AND PRESENT
CONSUMPTION
Develop information and plans
41
4
42
EnMS – conceptual cycle
BUILD COMMITMENT
Decision making and
support
Reporting
DEVELOP INFORMATION
AND PLANS
Opportunity list and action
plan
Technical audits and
operation control review
Collect energy bills and
sub-meter data
Analyze past and present
energy consumption
Forecasting: Targets and
budgets
Develop (or review)
baselines and EnPis
Identify and quantify SEUs
Identify relevant variables
and collect past data
IMPLEMENTATION
Procurement and Design
Operational control
Training
Implement action plan
CHECKING
Verify results of action plan
Compare actual and target
(or expected) consumption
Investigate and correct
significant deviationskWh(€ + CO2)
2/18/2016
22
Analyze Energy Use & Consumption
Collect past and current monthly consumption data at the
facility level (energy bills).
Determine what other data may be available for analysis.
• Sub-meter data
• Interval data
• Equipment information
• Other data
Determine PAST and CURRENT consumption by use.
Note: The time period for data collected will depend on
your organization and what data is available.
43
What are my energy sources,
uses and consumption levels? Electrical, natural gas, propane, hydro, wind?
What facilities, systems or equipment are using energy?
What data do we have? Where?
What data do we need? Where?
How much energy are we consuming?
How much did we consume in the past?
What are energy consumption trends for the future?
44
2/18/2016
23
Moving total of previous 12 months (or 52 weeks, etc)
• Removes seasonal effects
• Gives a real view of comparison v budget
• Effects of a change stay for next 12 periods
• Absolute numbers
• No allowance for changing activity levels
Very useful for forecasting, you can quickly judge what
next 12 months use will be
• You need to correct for known changes in output or other activity
Annualised trends
45
What does this tell us?
46
Food industry
0
200000
400000
600000
800000
1000000
1200000
1400000
1600000
1800000
2000000
01/2
009
02/2
009
03/2
009
04/2
009
05/2
009
06/2
009
07/2
009
08/2
009
09/2
009
10/2
009
11/2
009
12/2
009
01/2
010
02/2
010
03/2
010
04/2
010
05/2
010
06/2
010
07/2
010
08/2
010
09/2
010
10/2
010
11/2
010
12/2
010
01/2
011
02/2
011
03/2
011
04/2
011
05/2
011
06/2
011
07/2
011
08/2
011
09/2
011
10/2
011
11/2
011
12/2
011
01/2
012
02/2
012
03/2
012
04/2
012
05/2
012
06/2
012
07/2
012
08/2
012
09/2
012
10/2
012
11/2
012
12/2
012
01/2
013
02/2
013
03/2
013
04/2
013
05/2
013
kW
h p
er
month
2/18/2016
24
Is this good information?
47
0
200000
400000
600000
800000
1000000
1200000
1400000
1600000
1800000
2000000
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
2010
2011
2012
2013
Electricity data in annualised view
48
12/2
009
01/2
010
02/2
010
03/2
010
04/2
010
05/2
010
06/2
010
07/2
010
08/2
010
09/2
010
10/2
010
11/2
010
12/2
010
01/2
011
02/2
011
03/2
011
04/2
011
05/2
011
06/2
011
07/2
011
08/2
011
09/2
011
10/2
011
11/2
011
12/2
011
01/2
012
02/2
012
03/2
012
04/2
012
05/2
012
06/2
012
07/2
012
08/2
012
09/2
012
10/2
012
11/2
012
12/2
012
01/2
013
02/2
013
03/2
013
04/2
013
05/2
013
15000000
15500000
16000000
16500000
17000000
17500000
18000000
18500000
19000000
19500000
20000000
kW
h p
er
year
(ELE
C)
2/18/2016
25
Actual annualised electricity
usage and costs
49
1000000
1050000
1100000
1150000
1200000
1250000
1300000
1350000
1400000
1450000
1500000
12/2
009
01/2
010
02/2
010
03/2
010
04/2
010
05/2
010
06/2
010
07/2
010
08/2
010
09/2
010
10/2
010
11/2
010
12/2
010
01/2
011
02/2
011
03/2
011
04/2
011
05/2
011
06/2
011
07/2
011
08/2
011
09/2
011
10/2
011
11/2
011
12/2
011
01/2
012
02/2
012
03/2
012
04/2
012
05/2
012
06/2
012
07/2
012
08/2
012
09/2
012
10/2
012
11/2
012
12/2
012
01/2
013
02/2
013
03/2
013
04/2
013
05/2
013
15000000
15500000
16000000
16500000
17000000
17500000
18000000
18500000
19000000
19500000
20000000
Euro
per
year
(ELE
C)
kW
h p
er
year
(ELE
C)
Consumption
Cost
Common mistakes
1. Year-to-date reporting
• Inaccurate near start of year
• Moving annual totals or averages better
• Calendar has no significance
• Why waste information from prior periods?
• Long-term history gives superior analysis
50
2/18/2016
26
Exercise 01
This data comes from the same plant
as the previous data.
• 1. Calculate and represent the annualised
trend for 2013
• What is the annual consumption in the year
ending July 2013
• 2. Which is the % change in consumption in
2013 compared to 2012?
51
Exercise 01 - Solution
52
2/18/2016
27
SIGNIFICANT ENERGY USES
(SEUS)
Develop plans and information
53
5
54
EnMS – conceptual cycle
BUILD COMMITMENT
Decision making and
support
Reporting
DEVELOP INFORMATION
AND PLANS
Opportunity list and action
plan
Technical audits and
operation control review
Collect energy bills and
sub-meter data
Analyze past and present
energy consumption
Forecasting: Targets and
budgets
Develop (or review)
baselines and EnPis
Identify and quantify SEUs
Identify relevant variables
and collect past data
IMPLEMENTATION
Procurement and Design
Operational control
Training
Implement action plan
CHECKING
Verify results of action plan
Compare actual and target
(or expected) consumption
Investigate and correct
significant deviationskWh(€ + CO2)
2/18/2016
28
What are energy uses?
• “manner or kind of application of energy1”
• The service provided
• e.g. Light, heat, pump, cool, ventilate, convey, etc.
Significant energy uses
• Large energy uses
• Uses with good potential for savings
SEU is a central and key concept of an
EnMS
1Source: ISO 50001
What are SEUs?
55
Brainstorm:
• What do you think are the large uses?
• Where do you think there are good savings opportunities
List them
Tools:
• Motor list
• Thermal process list
• Lighting list
What to do in a multi-building organisation
• Is it the biggest buildings?
Exercise 02 Identify your SEUs
56
2/18/2016
29
Motor List
57
Source: UNIDO EnMS Tools
SEUs
58
Electricity Heat
8,651,145
1,584,839
821,876
1,525,654
809,185
1,063,759
3,838,640
47%
9%
4%
8%
4%
6%
21%
Refrigeration
Comp Air
Lighting
Dryers
Pumps
Ovens
Others
0 5,000,000 10,000,000
kWh per year
4,250,300
6,146,639
9,297,205
22%
31%
47%
Steam
Hot water
Dryers
0 5,000,000 10,000,000
kWh per year
Electricity + Heat
170,012
245,866
371,888
692,092
126,787
65,750
122,052
64,735
85,101
307,091
8%
11%
17%
31%
6%
3%
5%
3%
4%
14%
Steam
Hot water
Dryers
Refrigeration
Comp Air
Lighting
Dryers
Pumps
Ovens
Others
0 200,000 400,000 600,000 800,000
euros per year
2/18/2016
30
5,079,109
7,345,234
11,110,160
20,961,723
3,840,065
1,991,406
3,696,660
1,960,655
2,577,488
9,301,024
7%
11%
16%
31%
6%
3%
5%
3%
4%
14%
Steam
Hot water
Dryers
Refrigeration
Comp Air
Lighting
Dryers
Pumps
Ovens
Others
0 10,000,000 20,000,000 30,000,000
kWh per year
5,079,109
7,345,234
11,110,160
22%
31%
47%
Steam
Hot water
Dryers
0 10,000,000 20,000,000 30,000,000
kWh per year
20,961,723
3,840,065
1,991,406
3,696,660
1,960,655
2,577,488
9,301,024
47%
9%
4%
8%
4%
6%
21%
Refrigeration
Comp Air
Lighting
Dryers
Pumps
Ovens
Others
0 10,000,000 20,000,000 30,000,000kWh per year
SEUs
59
Electricity Heat Electricity + HeatPrimary Energy Primary Energy Primary Energy
799
1156
1748
2855
523
271
503
267
351
1267
8%
12%
18%
29%
5%
3%
5%
3%
4%
13%
Steam
Hot water
Dryers
Refrigeration
Comp Air
Lighting
Dryers
Pumps
Ovens
Others
0 1000 2000 3000
tCO2 per year
799
1,156
1,748
22%
31%
47%
Steam
Hot water
Dryers
0 1,000 2,000 3,000
tCO2 per year
2,855
523
271
503
267
351
1,267
47%
9%
4%
8%
4%
6%
21%
Refrigeration
Comp Air
Lighting
Dryers
Pumps
Ovens
Others
0 1,000 2,000 3,000
tCO2 per year
SEUs
60
Electricity Heat Electricity + HeatCO2 Emissions CO2 Emissions CO2 Emissions
2/18/2016
31
PERFORMANCE MODELS –
ONE RELEVANT VARIABLE
Develop information and plans
61
6
62
EnMS – conceptual cycle
BUILD COMMITMENT
Decision making and
support
Reporting
DEVELOP INFORMATION
AND PLANS
Opportunity list and action
plan
Technical audits and
operation control review
Collect energy bills and
sub-meter data
Analyze past and present
energy consumption
Forecasting: Targets and
budgets
Develop (or review)
baselines and EnPis
Identify and quantify SEUs
Identify relevant variables
and collect past data
IMPLEMENTATION
Procurement and Design
Operational control
Training
Implement action plan
CHECKING
Verify results of action plan
Compare actual and target
(or expected) consumption
Investigate and correct
significant deviationskWh(€ + CO2)
2/18/2016
32
Expected consumption
We can compute expected consumption accurately if we
consider the variables that cause consumption to vary
• Production throughput?
• Weather?
• etc
We must be able to measure these variables
• (Also known as ‘driving factors’, ‘relevant variables (in ISO 50001 and ISO 50006)’,
‘energy factors’, ‘explanatory variables’, ‘independent variables’, or ‘drivers’).
63
Relevant variables
Measurable
Routinely variable
Cause consumption to vary
(or are plausibly correlated)
• Production activity…
• Weather…
• Hours of darkness…
• Distance driven…
• … etc …
64
2/18/2016
33
Static factors
Some things influence consumption but do not routinely
vary…
For example:
• Size of process equipment
• Number of luminaires in a lighting system
• Size of a building
65
Exercise 03
List some possible variables that affect energy
consumption in your organisations
• Do this for one or two SEU’s
Categorise them as variable and static
Do you know an SEU whose energy consumption is not
affected by a variable?
66
2/18/2016
34
Scatter diagram
67
Represent consumption VS
relevant variable
See the trend
Observe the dispersion
Obtain the formula
0
200
400
600
800
1,000
1,200
1,400
1,600
0 10 20 30 40 50 60
kW
h/w
eek
CDD 15
y = 18,572x + 167,84R² = 0,8926 Remember: Y= mX + c
• c and m are constants
• X is a measured “relevant variable” variable
0
200
400
600
800
1,000
1,200
1,400
1,600
0 10 20 30 40 50 60
kW
h/w
eek
CDD 15
y = 18,572x + 167,84R² = 0,8926
Scatter diagram
68
You can also use formulae in
excel c: =INTERCEPT (known_y's,known_x's)
m: =SLOPE (known_y's,known_x's)
R2 =RSQ(known_y's,known_x's)
Remember: Y= mX + c
• c and m are constants
• X is a measured “relevant variable” variable
2/18/2016
35
Understand and interpret results
Intercept:
• What does the intercept mean?: Consumption when all the
variables are 0 at the same time.
• It is the baseload in most of the cases, unless that case is outside
of the model range.
69
R2:
• What does the R2 mean?: % of variation explained by variables
• High R2:
a) If all predicted variables were included:
a) Strong correlation. Not necessarily good performance.
b) If not all predicted variables were included. Think why.
a) The other ones were not really variables.
b) Saving Opportunities in operational control.
• Low R2:
a) There are other variables.
b) Saving Opportunities in operational control.
70
Understand and interpret results
2/18/2016
36
A technical example: boiler
71
y = 1.3761x + 189.84R² = 0.9933
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
0 2000 4000 6000 8000
kW
gas
kW steam
0%
10%
20%
30%
40%
50%
60%
70%
80%
0 2000 4000 6000 8000
eff
icie
ncy
kW steam
• 1,3761 kW of gas to get each kW of steam.
• Standing losses of 189.84 kW of gas
• The efficiency is lower when the output (and input) is lower.
Example: glass furnace
72
Fixed 220,000
kWh per week
Variable
355 kWh
per tonne
2/18/2016
37
Exercise 04
73
Suppose weekly consumption characteristic is
• 220,000 kWh/week + 355 kWh/tonne
What would it be at daily intervals?
• 31,400 kWh/day + 50.7 kWh/tonne
• 31,400 kWh/day + 355 kWh/tonne
• 220,000 kWh/day + 355 kWh/tonne
• 220,000 kWh/day + 50.7 kWh/tonne
Exercise 04
74
Look at these data, from a SEU
(compressed air) in the demo
plant.
• 1. Use a scatter diagram to analyse
the relation between sliced
products volume (t) and
consumption.
• 2. What is the intercept telling us?
• 3. How many kWh will we need to
produce 10 t more each month?
2/18/2016
38
Exercise 04 - Solution
75
PERFORMANCE MODELS -
MORE THAN ONE RELEVANT
VARIABLE
Develop information and plans
76
7
2/18/2016
39
77
EnMS – conceptual cycle
BUILD COMMITMENT
Decision making and
support
Reporting
DEVELOP INFORMATION
AND PLANS
Opportunity list and action
plan
Technical audits and
operation control review
Collect energy bills and
sub-meter data
Analyze past and present
energy consumption
Forecasting: Targets and
budgets
Develop (or review)
baselines and EnPis
Identify and quantify SEUs
Identify relevant variables
and collect past data
IMPLEMENTATION
Procurement and Design
Operational control
Training
Implement action plan
CHECKING
Verify results of action plan
Compare actual and target
(or expected) consumption
Investigate and correct
significant deviationskWh(€ + CO2)
Straight-line models are most common
More complex models may be appropriate
• Curved characteristics
• Multiple relevant variables
• Modelling from first principles
Expected-consumption formulae
78
x - + (
≠ ÷ √
2/18/2016
40
Expected consumption =
• c kWh per week (or per day, month etc)
• + m1 kWh per tonne of product A
• + m2 kWh per tonne of product B
• + m3 kWh per tonne of product C
Multiple relevant variables
79
Discussion
Consider a car
What are all the relevant variables for fuel
consumption?
Which are practical to measure?
Which are economical to measure?
80
2/18/2016
41
81
Have we chosen appropriate relevant
variables?
Use technical knowledge
and common sense
Test the significance of
each factor in the model
82
Testing significance of relevant variables
Use Excel’s regression
analysis tool
2/18/2016
42
83
Testing significance of relevant variables
84
Testing significance of relevant variables
2/18/2016
43
85
Testing significance of relevant variables
P-value < 0.1?
86
Testing significance of relevant variables
2/18/2016
44
87
Testing significance of relevant variables
P-value < 0.1?Coefficients
P-value:
• What does the P-value mean?: Probability of being significant.
• Low P-value:
a) The variable is significant.
• High P-value:
a) The variable is not significant.
b) Some variables are correlated. Colinearity. Check it.
c) The variable is significant but there are other variables.
d) Saving Opportunities in operational control.
88
Understand and interpret results
2/18/2016
45
89
Testing significance of the model
Significance F < 0.1?
Significance F:
• What does this mean?: Probability of being significant.
• Low Significance F:
a) The model is significant.
• High Significance F:
a) The model is not significant.
b) Some variables are non-linear.
• But:
• Low significance F + high P-value
a) Colinearity
90
Understand and interpret results
2/18/2016
46
91
Regression analysis – key points
Regression analysis is only a statistical estimate of the
effect of each relevant variable
Technical understanding of the process is critical
Operational control is an un-measurable relevant variable
• Important concept
• Often very significant
THE IMPORTANCE OF
WEATHER
Develop information and plans
92
8
2/18/2016
47
93
EnMS – conceptual cycle
BUILD COMMITMENT
Decision making and
support
Reporting
DEVELOP INFORMATION
AND PLANS
Opportunity list and action
plan
Technical audits and
operation control review
Collect energy bills and
sub-meter data
Analyze past and present
energy consumption
Forecasting: Targets and
budgets
Develop (or review)
baselines and EnPis
Identify and quantify SEUs
Identify relevant variables
and collect past data
IMPLEMENTATION
Procurement and Design
Operational control
Training
Implement action plan
CHECKING
Verify results of action plan
Compare actual and target
(or expected) consumption
Investigate and correct
significant deviationskWh(€ + CO2)
Weather-related energy demand
94
Energy consumption varies because of the
weather in many industries
Space heating and cooling
Cold stores
Industries with refrigeration as an SEU
Clean rooms in pharmaceuticals,
microelectronics, etc.
Is it feasible to relate energy consumption
directly to outside-air temperature?
2/18/2016
48
What about weather-related demand?
95
We did this for an energy-intensive process…
Can we do something similar for weather-related energy
consumption?
Heating Degree Days (HDD)
and Cooling Degree Days (CDD)
“Base temperature”:
• HDD base: outside temp. above which no artificial heating is required.
CDD base: outside temp. below which no artificial cooling is required.
• Default in the UK & IRL 15.5ºC (Austria is 12C)
• Other countries differ: Lower HDD base in countries with high
standards of weatherisation
• Depends on the building construction and internal heat gains
• Can be calculated in a daily/monthly/yearly basis.
96
2/18/2016
49
How to get HDD and CDD?
www.degreedays.net
97
How to get HDD and CDD?
98
City name and press
Station Search
Choose station
Choose data options
Generate
Wait and download
2/18/2016
50
Temperature-related demand
99
Heating base
temperature
Days
De
gre
es
Temperature-related demand
100
Shaded area is proportional
to heat energy requirement
Days
De
gre
es
2/18/2016
51
Heating ‘degree-day’ figures
101
Weekly gas consumption
Notice similarities between the shapes
Plot energy against degree days
Temperature-related demand
102
Weekly degree-day values
2/18/2016
52
Temperature-related demand
103
Knowing the degree day figure, we can read off expected
gas consumption
Temperature-related demand
104
12.7
900
Expected consumption= 900 kWh/week + 12.7 kWh/degree day
2/18/2016
53
What does outside temperature affect?
Building heating and cooling energy
Refrigeration as an SEU
• Food and drink industries
Industries with critical indoor environmental conditions
• Microelectronics
• Car assembly (painting is a SEU)
Humidity can have a similar effect
105
Changing base temperature
Sometimes another base temperature is needed to get year
round data.
• E.g. many zeros at the traditional base.
• Typical in industries with refrigeration during all the year
106
2/18/2016
54
A typical cooling degree-day history
107
Weather in multivariate regression
108
In general HDD is used in Heating analysis, and CDD in
Cooling Analysis
But in some cases both need to be included
When the same system is used for heating or cooling:• heat pump