optimize efficiency with thermal iq and advanced …...optimize efficiency with thermal iq™ and...

OPTIMIZE EFFICIENCY WITH THERMAL IQ™ AND ADVANCED ANALYTICS

Wyatt Culler Jayson Perdue John Parker

TABLE OF CONTENTS

Abstract ......................................................................................................................................................................................... 2

Description of site issue ................................................................................................................................................................ 3

Troubleshooting and analysis ....................................................................................................................................................... 5

Conclusions ................................................................................................................................................................................... 8

Appendix. Siegert Combustion. Efficiency versus Thermal Efficiency .......................................................................................... 9

1

OPTIMZIE 2

ABSTRACT

This whitepaper shows how Thermal IQ™ and advanced analytics can be used to troubleshoot equipment and optimize efficiency. One of the plant heating boilers at the Honeywell Thermal Solutions campus in Muncie, Indiana had persistent issues with short-cycling. The short-cycling caused additional wear via thermal fatigue stress and excessive valve actuation, as well as reductions in efficiency. Thermal IQ allowed a number of potential causes of the short-cycling to be narrowed to a single root cause that could be addressed. Addressing the root cause reduced the short-cycling of the boiler and resulted in an efficiency gain of about 5%. Note that the techniques and sensors used to optimize this small heating boiler are also applicable to larger process steam boilers where they could result in larger savings.

OPTIMIZE EFFICIENCY WITH THERMAL IQ™ AND ADVANCED ANALYTICS 3

Description of site issue

The manufacturing facility for Honeywell Thermal Solutions in Muncie, Indiana utilizes a natural gas fired boiler, pictured in Fig. 1, for plant heating. This is used as a low-pressure boiler with a nominal pressure setpoint around 10 PSI. Throughout 2019 and into 2020 this boiler would frequently short-cycle, as shown in Fig. 2. The time traces in Fig. 2 show (from top to bottom) the commanded firing rate, fuel flow, stack O2 percentage, and boiler pressure. All traces exhibit sinusoidal variations with period around 4 minutes. This unstable operation is not conducive for boiler efficiency or longevity. Some potential causes for the rapid cycling are identified in the list below. However, without adequate instrumentation it was not possible to isolate these potential root causes into an actual root cause.

1.1 Potential Root Causes

• Low condensate return

o A low condensate return could be caused by a leaky steam trap or a hole in a heat exchanger.

o Symptoms:

▪ Frequent boiler filling

▪ Frequent cycling of makeup water

▪ Temperature swings in feedwater

o Potential root causes:

▪ Frequent cycling of city water valve

▪ Steam leak necessitating frequent feedwater top offs

• Excessive feedwater flow

o Introducing too much feedwater too quickly can reduce the boiler pressure. The burner control compensates by quickly ramping up the throttle.

o Symptoms:

▪ Rapid cycling of burner throttle

▪ Rapid cycling of feedwater pumps

▪ Rapid changes in boiler pressure

o Potential root causes:

▪ Oversized feedwater pumps

▪ Faulty on/off floats

1

Figure 1. Muncie Boiler


Figure 2. Data traces showing boiler short cycling

1.2 Instrumentation and Control Upgrades

The boiler controls were upgraded to Honeywell SLATE in 2015, but the

upgrade was limited to adding the burner control system, new actuators, and

limited fuel, air, and pressure metering. The SLATE upgrade confirmed the

severity of the short-cycling issues.

In January 2020, the boiler was upgraded with additional sensors to measure

feedwater flow, freshwater flow, feedwater temperature, and combustion air

temperature.

Figure 3. SLATE Control Panel


Troubleshooting and analysis

Data collection with the new sensors was started by January 25th, 2020. Figure

4 shows representative time-series data from the new sensors. Very little

makeup water was added to the boiler (5th plot from top), indicating the rapid

cycling was not caused by a low condensate return. Additionally, the feedwater

temperature (7th plot from the top) does not appear to be low or to oscillate

significantly. However, the feedwater flow (6th plot from top) shows frequent,

high amplitude oscillations that lead the oscillations in the firing rate and fuel

flow plots (top and second from the top plots). This indicates that that

excessive feedwater flow is the root cause of the rapid oscillations in the firing

rate because it happens before the firing rate and fuel flow oscillations. The

root cause was addressed by reducing the maximum feedwater flow using a

gate valve, as the installed feedwater pumps had no adjustment on pump

pressure. The transition for the flow decrease is shown by the red region on Fig.

3. After the adjustment period the rapid oscillations in boiler firing rate were

greatly diminished. Additionally, the boiler pressure setpoint was much better

maintained after the feedwater flow adjustment.

Figure 4. Data traces before and after feedwater flow adjustment

The savings benefit of the feedwater flow adjustment is quantified by

comparing the boiler thermal efficiency both before and after the

adjustment. Note that calculating the actual thermal efficiency using

2


estimated steam flows1 is more accurate than using a Sievert combustion

efficiency calculation2. Figure 5 shows the thermal efficiency of the

boiler, where the amount of steam generated is inferred from the

feedwater flow. A time series of the daily average thermal efficiency in

black overlaid on the firing rate in red is shown in Fig. 5, while Fig. 5b

shows a histogram distribution of the daily thermal efficiencies. Before

the feedwater flowrate adjustment, the daily efficiency was around 75%

and the boiler rapidly cycled from high fire to low fire as indicated by the

nearly solid-looking red trace firing rate trace in Fig. 5. After the

feedwater flow was decreased, the thermal efficiency increased to above

80 percent. This efficiency gain translates to about a 5% savings in fuel

per year.

Figure 5. Comparison of efficiency before and after feedwater flow adjustment

Figure 6 shows a summary plot of the daily fuel cost of running the

boiler, assuming a natural gas price of $0.80 per therm. Figure 6a shows

the daily fuel cost of the boiler in black on top of the plotted firing rate in

1 It was not possible to install a steam flow meter, so steam flow was inferred using a mass balance on the feedwater flow. 2 Elaborated in the Appendix


red. Fig. 6b shows a histogram of the daily costs. It is important to note

that the daily cost depends on both boiler demand and boiler efficiency.

Although the daily fuel cost increased after the boiler feedwater flow

was optimized, Fig. 5 indicates this increase was due to increased

demand rather than lower efficiency. Assuming an average daily cost of

$300, the roughly 5% efficiency gain translates to an annual savings of

more than $5,000 per boiler.

Figure 6. Comparison of daily cost before and after feedwater flow adjustment


Conclusions

This case study has demonstrated how advanced sensors and the

capability of Thermal IQ can be used to optimize an industrial boiler.

The additional instrumentation allowed the root cause of the short-

cycling of the boiler to be identified as excessive feedwater flow, and to

be quickly remedied.

The instrumentation also allowed the thermal efficiency to be calculated,

showing how this adjustment resulted in about a 5% efficiency gain.

Finally, this case study has shown how Thermal IQ coupled with fuel flow

sensing can be used to both track the daily fuel cost on a per-asset basis

and to quantify gains in terms of dollars saved. Even greater cost savings

can be possible when applying these techniques to larger process steam

boilers.

3


Appendix: Siegert Combustion Efficiency versus Thermal Efficiency

Siegert Combustion Efficiency versus Thermal Efficiency

Efficiency in indirect heating applications such as boilers is sometimes

modeled using the Siegert equation (also known as the combustion

efficiency equation or similar variants). This semi-empirical equation

takes the form of equation 1, using the assumptions of equation 2,

where A and B are constants that vary by fuel. For natural gas, A is taken

as 0.66 and B is 0.009. This equation only models the stack heat loss and

assumes that all other heat losses (such as radiation) are negligible.

𝜂 = 100% − 𝑄𝑆𝑡𝑎𝑐𝑘(%)

𝑆𝑡𝑎𝑐𝑘𝐿𝑜𝑠𝑠 = (𝑇𝑆𝑡𝑎𝑐𝑘 − 𝑇𝐶𝑜𝑚𝑏𝑢𝑠𝑡𝑖𝑜𝑛𝐴𝑖𝑟) (𝐴

21 −%𝑂2+ 𝐵)

Figure A.1. shows the combustion efficiency of the Muncie boiler

calculated using equation 2. This equation yields an efficiency around

90%. This is significantly higher than the thermal efficiency shown in Fig.

4 and the efficiency estimates provided by the boiler manufacturer.

Thus, the Siegert method of calculating combustion efficiency over-

predicts the actual thermal efficiency of a boiler.

Figure A.1. Comparison of efficiency before and after flow adjustment

4

References [1] Ministry of Fuel and Power. The Efficient Use of Fuel. H.M. Stationery Office. 1944. [2] TSI Incorporated. Combustion Analysis Basics. Tech. Rep. P/N 2980175, Rev. B, TSI. 2004.

For More Information

www.ThermalSolutions.Honeywell.com

Honeywell Process Solutions

Honeywell Thermal Solutions 2101 City West Blvd

Houston TX 77042

WP-20-03-ENG I 03/20

© 2020 Honeywell International Inc.

http://www.honeywellaidc.com/

optimize efficiency with thermal iq and advanced …...optimize efficiency with thermal iq™ and...

Documents