|Title||Evaluation of a Technical Basis for Setting Firebox Pressure Oscillation Alarm in Fired Heaters|
|Contributor||Guinon, Bruce and Miller, John|
|Spatial Coverage||Kauai, Hawaii|
|Subject||AFRC 2013 Industrial Combustion Symposium|
|Description||Paper from the AFRC 2013 conference titled Evaluation of a Technical Basis for Setting Firebox Pressure Oscillation Alarm in Fired Heaters by Jamal Jamaluddin|
|Abstract||The Rosemount 3051S pressure transmitter was tested to verify if the Standard Deviation signal (Std-Dev) could be utilized as a means to define an alarm point for Draft Oscillation in fired heaters. This test included peak-to-peak pressure pulses from 0.5 in-w.c. to 2.0 in-w.c., and frequencies from 1 Hz to 10 Hz. The Std-Dev signal provided by the pressure transmitter via HART is converted to 4-20 mA signal. The Std-Dev signal was shown to meet the intent of this application, and not likely to cause nuisance alarm concerns if due diligence is used in selecting the alarm point. The purpose of the firebox pressure alarm is to provide an early alert that the burners are outside of the desired, or historical, operating envelope. If this value becomes a nuisance alarm, then it can be increased by one standard deviation. The recommended action for this alarm is to check the operating parameters, such as the O2, damper position, firebox pressure, fuel gas pressure and fuel gas flow. A proactive response to such an alarm would be to automatically reduce the firing rate - an approach that has been implemented at some locations. If the operating parameters are normal, then a visual check of the burners should be conducted to see if one or more burners are becoming unsteady. The affected burner may need to be cleaned.|
|Rights||No copyright issues|
Evaluation of a Technical Basis for Setting Firebox Pressure Oscillation Alarm in Fired Heaters A.S. (Jamal) Jamaluddin and Bruce K. Guinon1 John Miller2 AFRC Annual Meeting September 22 - 25, 2013 Sheraton Kauai Resort, Hawaii 1Shell Global Solutions (US), Inc. 2Emerson Process Management Summary The Rosemount 3051S pressure transmitter was tested to verify if the Standard Deviation signal (Std-Dev) could be utilized as a means to define an alarm point for Draft Oscillation in fired heaters. This test included peak-to-peak pressure pulses from 0.5 in-w.c. to 2.0 in-w.c., and frequencies from 1 Hz to 10 Hz. The Std-Dev signal provided by the pressure transmitter via HART is converted to 4-20 mA signal. The Std-Dev signal was shown to meet the intent of this application, and not likely to cause nuisance alarm concerns if due diligence is used in selecting the alarm point. The purpose of the firebox pressure alarm is to provide an early alert that the burners are outside of the desired, or historical, operating envelope. If this value becomes a nuisance alarm, then it can be increased by one standard deviation. The recommended action for this alarm is to check the operating parameters, such as the O2, damper position, firebox pressure, fuel gas pressure and fuel gas flow. A proactive response to such an alarm would be to automatically reduce the firing rate - an approach that has been implemented at some locations. If the operating parameters are normal, then a visual check of the burners should be conducted to see if one or more burners are becoming unsteady. The affected burner may need to be cleaned. 1. Introduction Positive pressure in the firebox of a process heater is known to be detrimental to heater operation. If and when the pressure goes too positive, a loss of flame can potentially occur. This is considered to be a serious event, since loss of flame can lead to an explosion if the fuel suddenly ignites. The development of positive pressure in the firebox of a heater, and consequent loss of flame, is often preceded by draft fluctuation in the firebox. It is plausible, therefore, that if the operator can be alerted of an unacceptable degree (frequency and amplitude) of draft fluctuation developing inside the firebox, operational changes might be possible to restrict the progress of such fluctuation, and thus avoid a potential flame-loss. Low-NOx burners, installed in natural draft heaters, are particularly vulnerable to draft fluctuation and flameout caused by high firebox pressure, since the technology relies on re-circulating flue gases into the combustion zone to lower the flame temperature, and thereby reduce thermal NOx formation. As a result of this recirculation, the flammability envelope for the combustible mixture is narrowed. In order to validate that the Std-Dev signal could be used to detect unstable flame conditions, a series of tests were conducted at the Rosemount Facility in Minneapolis-St. Paul. The findings from those tests are summarized in this paper. 2. Advanced Pressure Diagnostics Smart field devices are becoming increasingly more prevalent in the process industry. These devices provide not just the basic process measurement (e.g. pressure, differential pressure, draft, temperature, or flow) but they also provide a wide range of additional information, such as second and third process variables, device identification and health, and device diagnostic information. The Rosemount 3051S Advanced Diagnostics Pressure Transmitter provides a large amount of diagnostics information including statistical process monitoring, loop current/power monitoring, and process variable and event logging. The advanced diagnostics capability is available with the DA2 option in the 3051S family of pressure transmitters, including absolute and gauge pressure, differential pressure (DP), DP flow, DP level, and draft. Statistical Process Monitoring (SPM) technology allows a plant engineer or operator to see the variation (or noise) in a process that is normally filtered out when the variable is measured and displayed in the Distributed Control System (DCS). The pressure transmitter samples the pressure at a much faster rate than the DCS, and turns process variation into valuable information. The diagnostics transmitter calculates mean and standard deviation of the pressure. The standard deviation value quantifies the variation in the pressure signal. The statistical values are sent to the host system via digital HART communication. SPM technology is also available on 3051S Foundation Fieldbus pressure transmitters. Figure 1 illustrates the architecture and implementation of a pressure transmitter with Advanced Diagnostics SPM capability. Figure 1 - Architecture of Advanced Pressure Diagnostics The Pressure Sensor measures the pressure at 22 Hz, applies the normal PV Damping (default damping = 0.4 seconds), and then sends the measured pressure to the Control System via 4-20 mA. On a parallel path, a High Pass Filter and Statistical Calculations module takes the un-damped pressure and calculates mean and standard deviation. The statistical values are made available to the control system as secondary variables via HART I/O. If a control system does not have HART I/O capability, the statistical values can still be trended via a HART Tri-Loop (e.g. Rosemount 333 HART Tri-Loop) which converts the HART data to 4-20 mA. Alternatively, a wireless HART adapter (e.g. Rosemount Smart Wireless THUM adapter) can be used to access the advanced diagnostics information, and transmit the information wirelessly to a gateway (e.g. Rosemount Smart Wireless Gateway), which then integrates the diagnostics information with the Control System. Figure 2 further illustrates Statistical Process Monitoring technology. On the left, is the process variable as measured by the pressure transmitter, with different levels of process variation (Normal, Noisy, and Quiet). On the right are the variables that the advanced diagnostics pressure transmitter sends to the control system, the Process Variable (via 4-20 mA) and the Standard Deviation (HART I/O). Figure 2 - Statistical Process Monitoring in a Pressure Transmitter Note that the Process Variable is relatively flat, because the PV damping has removed all of the process variation. In addition to PV damping in the pressure transmitter, there may also be damping or filtering in the DCS, which further removes any variation in the pressure. The standard deviation is a measure of how much variation is in the process. The standard deviation is higher when the process is noisier, and it is smaller when the process is quieter. The SPM standard deviation is not affected by transmitter damping, DCS filtering or slower sampling rates of a DCS. A high-pass filter is used prior to the Statistical Calculations module. The purpose of the high-pass filter is to ensure that the SPM analyzes only the process variation at higher frequencies (approximately 1 Hz and higher). When there is just a slow change in the pressure measurement (such as the heater draft), this should not be interpreted as an increase in the process variation. 3. Frequency Response Test Setup Figure 3 shows a schematic diagram of a system used to test the dynamic response of a pressure transmitter. The purpose of the test system is to create varying pressure signals of known characteristics to evaluate how the pressure transmitter responds. Figure 3 - Dynamic Pressure Response Test System Pressure pulsations are created by a Signal Generator. A Voltage-to-Current Interface is needed to convert the voltage output of the Signal Generator to a current input for the I/P (a Fisher Controls 546) which receives a 4-20 mA current input and outputs 3-15 psi. The pressure output is connected to both a High Speed Reference Pressure Sensor, and to the high side of the pressure transmitter being tested (Device Under Test or DUT). The analog-pneumatic-mechanical design of the I/P allows it to reproduce a smooth sine-wave from the signal generator without digital sampling artifacts that would result when using some microprocessor-based I/P devices for this frequency response testing. A Data Acquisition System is used to simultaneously read the analog outputs of the Reference Sensor and the DUT, along with HART data, including the SPM standard deviation. Figure 4 shows an example of the pressure signals produced by the I/P and the static pressure controller and the differential pressure signal that is applied to the DUT. The average pressure is approximately 126 inH2O, but the actual pressure fluctuates above and below this value. The amplitude of the pressure pulsations changes from 1 inH2O to 3 inH2O. The second plot (b) is an example of the corresponding output of the Static Pressure Controller. The third plot (c) shows the differential pressure that is measured by the DUT. The amplitudes of the pressure pulses are the same as in (a), but the average value is zero. The bottom plot (d) shows the resulting SPM standard deviation value as the magnitude of the pressure pulses changes. When the pulse amplitude increases 3 times, the standard deviation also increases 3 times. Signal Generator (0.01 to 15 Hz) Current-to-Pressure Converter (I/P) (3 to 15 psi) Static Pressure Controller Device Under Test (DUT) Data Acquisition 4-20 mA HART 1-5 V High Speed Reference Pressure Sensor Voltage to Current Interface 0 1 2 3 4 5 6 7 8 9 10 122 124 126 128 130 a.) I/P Output inH2O 0 1 2 3 4 5 6 7 8 9 10 125 125.5 126 126.5 127 b.) Static Pressure Controller Output inH2O 0 1 2 3 4 5 6 7 8 9 10 -4 -2 0 2 4 c.) Differential Pressure inH2O 0 1 2 3 4 5 6 7 8 9 10 0.5 1 1.5 2 2.5 d.) Standard Deviation inH2O Time (seconds) Figure 4 - Pressure Signals Produced by Components of the Dynamic Pressure Response Test System 4. Test Data A series of tests were performed with the dynamic pressure response test system. The purpose was to simulate pressure pulsation patterns that are typical of draft measurements in heater operation, both under normal heater operating conditions, and under conditions of flame instability. During each test, the value of the SPM standard deviation was recorded. The collective goal of all the tests is to determine and validate which factors affect the SPM standard deviation value, in order to get insight for setting a flame instability operator alarm. The tests were performed on a Rosemount 3051S Range 2 (-250 inH2O to +250 inH2O) DP transmitter with the DA2 option. The following sections describe the test parameters, and their effect on the SPM Std-Dev signal. a. Amplitude and Frequency The standard deviation is directly proportional to the amplitude of the pressure pulses, while the effect of frequency is related to both the time response of the pressure sensor, and the high-pass filter used in the SPM algorithm. The high-pass filter causes the standard deviation to be near zero at frequencies near 0 Hz. The first-order time response of the pressure sensor causes the standard deviation to decrease as the frequency of the pressure pulses increases to the higher range (8 to 10 Hz). The middle frequencies (4 to 7 Hz) have the highest values of SPM standard deviation. Figure 5 shows how the SPM standard deviation changes with both the frequency and the amplitude of applied pressure pulses. Figure 5 - Frequency and Amplitude Response of SPM Standard Deviation Prior testing within Shell has suggested that draft pulsations of 2 Hz frequency and higher, and with a Peak-to-Peak variation of 1.0 inH2O or higher, are indicative of flame instability in the burner(s) and/or heater. Draft pulsations with Peak-to-Peak variation of 0.5 inH2O or less are not considered to indicate flame instability. Based on these observations, together with the above data taken at the Rosemount facility, it appears that a standard deviation value of 0.08 inH2O is a good operator alarm limit for detecting flame instability. b. Effect of SPM Sampling Period The SPM feature has a user-configurable sampling period. The sampling period is configurable anywhere between a minimum of 1 minute and a maximum of 60 minutes. Generally, a shorter sampling period will give a faster response of the SPM standard deviation to an abnormal condition, and a longer sampling period will produce a smoother trend of standard deviation. The first test compared the SPM standard deviation values when changing the SPM sampling period from 1 minute to 2 minutes. The results of this test are shown in Figure 6. The SPM sampling period makes no difference in the standard deviation values. Figure 6 - Effect of SPM Sampling Period on Standard Deviation Values c. Time Response of SPM Standard Deviation For any alarm signal to be useful, it must have a response time that is consistent with the process safety time. To test this response time, the peak to peak sine wave input signal was changed from 0.11 inH2O to 1.0 inH2O and the time to achieve 90% of the final standard deviation value was measured. This process was repeated from 1 Hz to 8 Hz. The result was that at a 1-minute sampling period, the 90% of final value was achieved consistently 30 seconds after the amplitude change. This is equivalent to a first order time constant (63% of final value) of approximately 13 seconds. Figure 7 shows an example of this first-order time response of the SPM standard deviation. The test was also repeated with a sampling period of 2 minutes, resulting in reaching the 90% of final value in twice the time (60 seconds). Figure 7 - Example Time Response of SPM Standard Deviation d. Process Condition Change It is important that a nuisance alarm does not occur due to control changes or other process changes not related to pulsation from an unstable burner. To test standard deviation response to different process disturbances, the Signal Generator was set to a frequency of 0.01 HZ and the peak-to-peak amplitude changed from 0.1 inH2O to 1.6 inH2O to simulate a significant process step change. The standard deviation remained low (<0.018 inH2O) even with a 1.6 inH2O sine wave Peak-to-peak amplitude. This is approximately equivalent to a ramping change of 1.6 inH2O over 50 seconds which would be a very large control move or process change. This behavior is a result of the high-pass filter which removes the effect of slow or gradual process changes such as seen in this test. Figure 8 shows the SPM standard deviation values observed during a slow process condition change, in comparison with the selected alarm limit. The standard deviation signal remains low during control moves or process disturbances. It is not expected to be a nuisance due to control or process disturbances. Figure 8 - SPM Standard Deviation Values during Slow Process Condition Change e. Effect of Nominal Pressure and Analog Calibration The tests also looked at conditions that were not expected to affect the SPM standard deviation: Changing the Nominal (Median) Pressure Changing the Transmitter Analog Calibration The effect of the nominal pressure was compared at values of 0 inH2O (the typical nominal pressure for all tests) and -0.5 inH2O (a more common heater draft setting). The analog calibration of the transmitter was changed from [-2.5 to +2.5 inH2O] to [-1.25 to +1.25 inH2O]. This would not be expected to affect the SPM standard deviation because the 4-20 mA analog output in the pressure transmitter is independent of the SPM calculations. Neither of these factors had any measurable effect on the SPM standard deviation. 5. Observations from a Test Furnace and Operating Heaters a. Test Furnace Data In 2006, Shell conducted a single burner test at Callidus Technologies. During this test, Emerson was able to field-test the Rosemount 3051S, Range 2, Standard Deviation signal. From this burner test, Std-Dev signals ranging from 0.03 to 0.22 were measured during burner instability periods, as shown in Table 1. The unusually high Std-Dev value (during field tests, Std-Dev values consistently remained below 0.1) shown in the table for sub-stoichiometric combustion may be an artifact of excessively high fuel gas pressure having been used to attain sub-stoichiometric conditions. High fuel gas pressures are known to produce higher pressure pulses within the firebox; an unstable flame with high fuel gas pressure is, therefore, likely to generate high Std-Dev signals. Under normal operating conditions, sub-stoichiometric combustion conditions are avoided; and therefore, Std-Dev value as high as 0.2 is unlikely to be encountered. Table 1: Test Data from Tests at Callidus Technologies Peak-to-Peak Pressure Pulses (in-wc) Std-Dev 3051S Range 2 Frequency (Hz) Condition 0.8 0.05 1 Flamed-Out 1.4 0.22 3.4 Substoichiometric Figure 9 depicts the pressure pulses and the Std-Dev during this test. Flame out occurred at a standard deviation of approximately 0.052 Std-Dev, with the flame going out and relighting repeatedly. The horizontal coordinate shows the elapsed time during the test. b. Data from an Operating Heater Rosemount 3051S Range 0 (-3 to +3 inH2O) draft transmitters were installed in an operating heater in one of Shell's operating sites in April 2010 Figure 10 shows the start-up values of the Std-Dev and the box pressure. Both the Std-Dev and fuel gas pressure spiked during the light off but remained below 15 percent. The data provided in Table 2 suggests setting an at 0.047 Std-Dev (Range 2). It is important to note, that this maximum Std-Dev value was achieved under pre-startup conditions. The normal operating values post-startup averaged around an Std-Dev value of 0.008 Std-Dev in Range 2(Range 0 has a lower Std-Dev value due to slower response time). Thus, an alarm value of 0.047 Std-Dev in Range 2 would not result in a nuisance alarm since the normal Std-Dev signal is well below this value. Table 2: Instantaneous Data Through 8 Days Following Startup Minimum Value Maximum Value Average Standard Deviation Firebox Pressure, in w.c. -0.63 0.20 -0.39 0.14 Excess Oxygen, % 2.0 10.0 5.1 2.0 Fuel Gas Pressure, psig 0.4 61.6 11.0 9.1 Std-Dev (Range 0) 0.0001 0.014 0.003 0.002 Std-Dev (Range 2) 0.0004 0.044 0.008 0.006 Alarm Set-Point 0.047 0.033 Flame Instability - pressure -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 200 400 600 800 1000 Figure 9: Draft measurement under flame-out conditions during the Callidus test. Figure 10: Effect of firebox pressure and fuel gas pressure on the Std-Dev signal. 6. Recommendations The purpose of this evaluation is to suggest an alarm point for the Std-Dev signal that is above the historical operating conditions so that the operator may be alerted of impending flame instability of a degree that may cause a flame-out. The intent of this alarm is to provide an early alert that the burners are outside of the historical operating envelope. As a first step, the alarm point can be set at or around a Std-Dev of between 0.04 and 0.08. Since the alarm point will vary from heater to heater, we recommend that the operation of the heater is monitored for three to six months to arrive at the proper alarm point for the subject heater. If this value becomes a nuisance alarm (e.g., more than one false alarm over 3 months), then it can be increased by another standard deviation in the fluctuating Std-Dev signal. The recommended action for this alarm is to check the operating parameters such as the O2, damper position, firebox pressure, fuel gas pressure and fuel gas flow. A proactive response to such an alarm would be to automatically reduce the firing rate - an approach that has been implemented at some locations. If the operating parameters are normal, then a visual check of the burners should be conducted to see if one or more burners are becoming unsteady. The affected burner may need to be cleaned. 7. Conclusion The Std-Dev signal is considered adequate for detecting flame instability over a large range of pulse frequencies. It can help provide an alarm in a timely manner for operator action, and can become a robust layer of protection for detecting unstable flames at all draft pressures.