ASHRAE Journal - March 2022 - 37

TECHNICAL FEATURE
to boiler operators. Third, steam or hot water may be
delivered to other buildings through tunnels that are
not climate controlled. Therefore, any loss between the
cycles is extra reheating of the pipe in the next cycle.
Condensing boilers have the same possible losses.
However, they have higher effi ciencies, especially
at mid-range loads. A condensing boiler has a heat
exchanger in the fl ue exhaust.8 Condensation forms
on the fl ue, and the latent heat of the condensate is
harvested with the heat exchanger. This reduces losses
up the fl ue, but doesn't affect the short cycling losses. A
condensing boiler has lower effi ciency at high loads due
to the capacity of the heat exchanger.
Boiler life can also be impacted by short cycling.
Temperature changes during cycling cause thermal
stresses in boilers.9 Those stresses are a main factor in
determining boiler life. Reducing short cycling should
have a benefi t of reducing repair requirements. If
energy savings measures were to increase short cycling,
that would be another discouragement to enact them.
Another problem with short cycling is that it is more
likely to lead to more condensation and more corrosion
in noncondensing boilers.
Modulation is often cited as a means to prevent short
cycling. Boilers have minimum operational levels.
Boilers without modulation have two modes: on and off.
However, some boilers can reduce their output to match
the demand. Common modulation ratios vary from 5:1
to 10:1. The lower limit of modulation is controlled by
matching the pump and burner rates.10 Some fi re tubes
might be turned off at lower modulation levels. If the
production of heat can exactly match the demand, the
boiler could go a very long time before cycling.
In large installations, boilers are often arranged in
banks.11 Having multiple boilers has several advantages
such as the ability to repair one boiler while maintaining
service. Another advantage is that a bank of boilers can
be operated as a means to modulate heat production by
turning on all boilers only in peak season.
Methods
Data from three heating systems (Table 1) were collected
by the author and clients from energy bills for two years.
All sites were located in Detroit, Mich., and the data was
for different two-year periods for each system since each
represented a different client.
TABLE 1 Boiler data.
SITE
BUILDING(S)
SIZE OF CLIMATECONTROLLED
SPACE
(FT2)
BOILER SYSTEM SIZE
TYPE/NUMBER
OF BOILERS
MODULATION
12,800
1 MBtu
Conventional/1
None
122,000
10 MBtu
Condensing/1
5:1
1,200,000
100 MBtu
High Effi ciency/2
5:1 Each
Weekly gas meter readings were collected. All data
was pre-COVID, so building use was consistent during
the period. Additionally, heating degree days (HDD) for
Detroit, Mich., were found for the relevant times. Water
heating for occupant use was done in separately submetered
systems for Sites 1 and 2, but for Site 3 the water
heating was combined for all 23 buildings . For Site 3,
boiler steam was used in heat exchangers to produce hot
water. When space heating was off, the base load from
water heating was determined and how it changed with
seasonal changes in building use.
Building models are an appropriate method to size
heating systems. However, the author had diffi culty
getting building models to match actual building performance
within 20%. Models can be made to match by
assuming boiler effi ciencies and adjusting assumed heat
loss values. However, this can be arbitrary and is not
appropriate for determining the total system effi ciency.
When measuring boiler effi ciency, often the only way
the boiler combustion effi ciency is measured is through
combustion gas analysis.2
For our purposes, the relative effi ciency is suffi cient.
This means the fuel demand can be compared to the
HDD to see how the effi ciency varies. It is possible to
match the maximum relative effi ciency found to the
ideal effi ciency of the boiler, but that is unnecessary for
supporting the conclusions of this work. The goal is to
fi nd the relative drop in effi ciency for each system at low
loads. The relative effi ciency is defi ned for this work as
the HDD divided by the fuel use. The relative effi ciency
will not be consistent between different sites because
larger sites have more square footage.
The fuel was natural gas at all sites. Meter readings
sometimes needed to be adjusted. For example, negative
fuel consumptions were found in the data. This impossibility
demonstrates that meter readings sometimes have
M A R C H 2 0 2 2 ashrae.org ASHRAE JOURNAL
37
1
2
1 Offi ce
Building
2 School Buildings
3
23 Climate-Controlled
Buildings for a School
http://www.ashrae.org
ASHRAE Journal - March 2022

Table of Contents for the Digital Edition of ASHRAE Journal - March 2022
