| OCR Text |
Show AFRC 2017 INDUSTRIAL COMBUSTION SYMPOSIUM September 18th, 2017 Houston, Texas, USA Estimating Fired Heater Coils Safety Time Using Transient Analysis Sultan Ahamad, Mohit Thakur, Mahmood Izadi, Rimon Vallavanatt Bechtel Corp., Houston, TX of temperature with time, and calculate the safe time for coil before it fails or product degrades. Introduction During operation of a process fired heaters, there are potential for sudden interruption of process flow. This could be due to pump trip or accidental closing of a valve in the inlet line. When process flow is tripped, the fuel must be cut off to protect the process coils. The heat input to the unit must be stopped immediately to prevent over-heating and degrading the fluid inside the coil and possible tube failure due to over pressure and temperature. The control loop has some inherent time delay before the fuel shut off valve to the heater is fully activated and thus stop additional fuel input to the system. It is critical to ensure that these coils do not fail or the process products get degraded when process flow is stopped. In this study, the results of the semi-analytical method proposed by Ahamad et al. [1] are compared with the ones obtained from numerical simulation of a similar case using Dynamic Simulation software. Comparing the results will help to find a more accurate estimated Safety Time. Process Safety Time Process Safety Time is defined as the time period between a failure occurring in the process or its control system and the occurrence of the hazardous event if the safety function is not performed. Safety Time is quite difficult to determine precisely. It can be estimated for each pair of cause & consequence in many cases. A dependable Safety Time estimation is the first step in identifying the time potentially available for response time of the system. Response time is the time taken from the input to the sensor associated with safety function being set, to the output device completing its required action. An adequate control system must be designed to ensure that the fuel gas is isolated within estimated Safe Time. Safety Time estimation is crucial for control loop design of equipment in the operating plant. It should be noted that numerical simulation of a fired heater is cumbersome due to complex geometry and high operating temperatures and pressures. Furthermore, considering the radiation effect for overall heat transfer rate in fired heaters makes the simulation more complex and expensive. This forces engineers to use simplifying assumptions to reduce complexity of simulation while these assumptions reduce accuracy of results. Fired Heater Radiant Section Coil The process fluid is heated in fired heater coils by heat provided by combustion of fuel in the radiant section. A typical fired heater consists of a radiant and convection sections. In radiant section, heat is transferred mostly by radiation. The convection section recovers heat from the flue gas leaving the radiant section and transfers it to the cold process fluid in the tubes. Heat transfer in fired heater is through combination of radiation and convection. Most of the previous work on fired heater is simulated as steady state. However, in this case, a transient heat transfer calculation is required. The intent of this transient analysis is to analyze and obtain analytical solutions for transient heat transfer considering the effect of radiation as well as convection heat transfer Predict the variation The process fluid outlet temperature is the parameter which controls the firing (fuel flow) rate in the heater. A set of shut-off valves are provided on fuel gas line to close the firing in the fired heater. There are multiple interlocks provided for safe operation of fired heaters. These interlocks are 1 Estimating Fired Heater Coils Safety Time Using Transient Analysis provided for safe operation of the heater and to prevent the potential degrading of process fluid, fouling of heating coil and failure of coils. The details of these interlocks are not discussed in this paper. Many of these interlocks, if initiated, close the shutoff valves both at inlet as well as at outlet of the heating coils. Fired Heater Coils Safety Time Calculation Methodology Fired heater heat transfer rate is highest in the radiant section of the heater. The radiant and convection sections are schematically shown in Figure-1. The calculation shall be performed at the tube which is exposed to the highest heat flux and maximum metal temperature in the radiant section, as this is the tube which will fail first. Average Radiant Heat Flux is defined as: = (1) Figure-1: A Typical Fired Heater Configuration The heat flux on the tubes is not uniform. Heat flux varies along the circumference of the tube as well as height of the firebox. Figure-2 illustrates the heat flux variation along circumference of a tube in a fired heater [2]. The heater designers use an average heat flux to size the radiant section of the fired heater. However, it is the peak (maximum) heat flux which controls the peak metal temperature. Hence, calculation shall be based on the peak heat flux tube zone. • • Peak (Maximum) Radiant Heat Flux at any point in a coil can be estimated from following equation: qm = Fc Fl Ft qa + qc (2) the coil through natural convection and conduction. To simplify the calculation, this small transferred heat is ignored. Tube wall thickness is small in these applications. Therefore, it is assumed that tube metal temperature shall be uniform along the thickness of the tube at the point of consideration. The tube wall temperature shall be highest at the point of peak heat flux in radiant section of fired heater. Therefore, transient heat transfer calculation shall be carried out at the point of peak heat flux. Heat balanceFactors Fc, Fl, and Ft in Equation-2 are complex functions of tube geometry and heat flux distribution [5]. In order to reduce the complexity of analytical method, a factor representing the ratio of peak heat flux to average heat flux will be used for calculation. ℎ = " " " " Typically, methods common in industry for calculating heat transfer in radiant section use empirical correlations, developed based on experimental data. The heat transfer in the radiant section is calculated with widely used methods proposed by Wimpress [3] and Lobo-Evans [4]. This method assumes complete flue gas mixing in the firebox such that there are no longitudinal or Basis and assumptions for fired heater calculations: • Heat is transferred to tube which increases the tube wall temperature. A very small portion of this heat will be transferred to stagnant fluid in 2 Estimating Fired Heater Coils Safety Time Using Transient Analysis transverse temperature gradients. The equation for radiant heat transfer is as follows: = ∝ − + ℎ − + ℎ ( − ) Dynamic simulation: The Dynamic Simulation tool is used to verify the safety time calculations. A fired heater unit operation is selected in the dynamic simulation tool to model the heater. The radiant coil section of the fired heater is simulated in detail. The radiant section tube inner and outer diameter, tube length, number of passes, physical properties and heat transfer properties of both tube and shell are defined identical to the case study. (3) Considering for peak heat flux and using the heat balance during transient case, it will result into following equation: ∝ − = (4) Case Study: Integrating from time t = 0, at which TSt = TS0 to time ‘t' gives following after rearranging: ∝ =∫ − 1 + ℎ − A summary of detailed safety time calculation study is presented here for a vertical cylindrical, single fired, hot oil fired heater. The results of the proposed semi-analytical method as well as the ones obtained from the dynamic simulation tool are presented in this section. The input design values for this fired heater are as follows: (5) This equation is difficult to solve due to being complex and non-linear. There are several mathematical steps requiring approximation and rearrangement of the terms to get a solution. Ts0 = 536°F Tst = 770°F Tg = 1,478°F For the selected fired heater, other parameters (e.g. α, fr, F, etc.) and constants are estimated based on fired heater configuration and operating conditions. The results of the proposed semi-analytical method are plotted in figure-3. The maximum allowable metal temperature is calculated based on fired heater operating condition and tube material properties [5]. The maximum allowable metal temperature calculated for this case is 770°F. When flow is stopped the tube metal temperature at the time of The final solution is presented in Equation-6 which can be used for estimating the Safety Time of a fired heater coil. Where, = 2 3 +7 ; = = ∝ 2 .thk.ρ 3 (6) +7 Single Fired Coils Double Fired Coils Figure-2 : Typical radiant heat flux variation along circumference of the radiant tube 3 Estimating Fired Heater Coils Safety Time Using Transient Analysis 800 Maximum Allowable Metal Temperature 700 600 500 0 50 100 150 200 250 Time, Seconds Figure-3: Fired heater Safety Time Estimated by SemiAnalytical Method Failure Time 900 800 Maximum Allowable Metal Temperature Safety Time (Start of degradation) Tube Metal Temperature, °F 1000 It should be noted that the tube will start degrading for temperatures above the maximum allowable metal temperature, 770°F. The tube rupture may happen at ≈1,000°F for Carbon Steel tubes [5]. This is shown on Figures 3 and 4 as Failure Time. For the current case study, the Failure Time is estimated to be around 200 seconds. 700 600 500 0 50 100 150 200 250 Time, Seconds Figure-4: Fired heater Safety Time Estimated by Dynamic Simulation The total time it takes to isolate the fuel gas to the fired heater in this case is 25.2 seconds. The control loop is designed so that the time required to isolate the fuel gas is lower than the estimated Safety Time. The fired heater refractory re-radiating heat is a significant contributor, which might increase the tube metal temperature even after fuel gas is being cut-off. Therefore, the controls design is safe for this application as the estimated Safety Time by both methods is higher than the Process Response Time. An operating loop is required to be capable of closing the fuel gas supply to the burners in the fired heater before reaching this temperature. It should also be noted that the fired heater walls and floor are refractory lined. The refractory, depending on type of refractory and thickness, retain a good amount of heat. This should also be considered when designing the operating loop. The Process Response Time, calculated for the control loop, in this case study includes the following components. The following data is taken from a fired heater in an LNG plant. The details of the data are not reproduced in this paper due to confidentiality. Components Sensor Delay Logic Solver Interface Logic Solver Final Element Total Process Response Time 900 Safety Time (Start of degradation) Tube Metal Temperature, °F The results of the numerical simulation using the Dynamic Simulation tool are shown in figure-4 for the same input design values listed above. The hot oil control valve is modeled with a typical actuator closing time of 10% per second. For the case study, the hot oil flow to the heater is stopped by malfunction of the control valve (fail close). It is assumed that the fuel requirement in the heater does not change as the hot oil flow reduces. As the hot oil flow to the heater decreases to eventually no flow situation, the tube metal temperature in the radiant section increases quickly to the maximum allowable metal temperature (770°F). The Safety Time estimated by the fired heater Dynamic Simulation model is 108 seconds which is close to the one calculated from the semi-analytical model. Failure Time 1,000 incident is 536°F as can be seen in Figure-3. In this case, it takes around 93 seconds for the metal to reach the tube maximum allowable temperature. Conclusion In this study the estimated Safety Time calculated by the semi-analytical method proposed by Ahamad et al. [1] is compared to the Safety Time obtained from the Dynamic Simulation tool. For both methods, a case study is provided based on real working control loop data of an LNG plant. Safety Time is determined for the case study by crossing the tube wall temperature and tube maximum allowable temperature and compared to the Process Response Time calculated for the control loop. Time, sec 0.5 20 0.75 0.2 0.75 3.0 25.2 4 Estimating Fired Heater Coils Safety Time Using Transient Analysis of ASME, Jul. 1956, pp. 1103-1111 [3] R. N. Wimpress, Rating fired heaters, Hydrocarbon processing and petroleum refinery, Oct 1963, Vol 42, No-10 pp. 117-126 [4] Lobo, W. E., Evans, J. E.; Heat Transfer in Radiant Section of Petroleum Heaters, Trans. Am. Inst. Chem. Engrs. 35, pp. 748-778, 1939 [5] API-530- Calculation of heater-tube thickness in petroleum refineries This study shows that the estimated Safety Time calculated by the semi-analytical method is close to the one obtained from the Dynamic Simulation tool. The comparison of the estimated Safety Times resulted from two different methods confirms that the proposed semi-analytical method is reliably capable of estimating the Safety Time for fired heaters in an economical way while the method offers a reasonable accuracy due to its analytical nature. Sultan Ahamad is a Sr. Fired Heater Specialist at Bechtel Corp. in Houston, Texas. He has more than 19 years of experience in the design, engineering and troubleshooting of fired heaters and combustion systems for the refining, petrochemical and LNG industries. He graduated from the Indian Institute of Technology in Roorkee, India, with a degree in chemical engineering. He is a member of API SubCommittee on Heat Transfer Equipment and contribute to the development of API standards and recommended practices. He is also a member of American Institute of Chemical Engineers (AIChE), American Society of Mechanical Engineers (ASME) He has published and presented several papers on fired heaters and related subjects. He can be reached by email: sahamad@bechtel.com For the case study, Process Response Time is less than estimated Safety Time which means the control system is capable of safely stopping the equipment from operation without tube failure or degrading of the process fluid. Also, a Failure Time is estimated for the current case study. Based on the results obtained from both models, failure may happen only about 100 seconds after the estimated Safety Time which shows the high criticality of this study. Nomenclature A = heat transfer area of the tube, ft2 C1, C2 = constants cp = specific heat of tube, Btu/lb.°F dT = small change in tube metal temperature, °F dt = small change in time, hr F = overall exchange factor in radiant section of fired heater fr = peak to average heat flux correction factor in radiant section Fc = circumferential heat flux variation factor Fl = longitudinal heat flux variation factor Ft = heat-flux variation factor due to tube metal temperature h = flue gas heat transfer co-efficient, Btu/hr.ft2.°F hc = convective heat transfer co-efficient in radiant section, Btu/hr.ft2.°F m = mass of the tube section, lb qa = average radiant heat flux, Btu/hr⋅ft2 qc = average convective heat flux in radiant section, Btu/hr⋅ft2 qm = maximum radiant heat flux, Btu/ft2.hr qr = total heat absorbed in radiant section, Btu/hr Tg = radiant section average radiating flue gas temperature, °R Ts = tube metal temperature, °R TS0 = initial tube metal temperature, °F TSt = tube metal temperature at time t, °F t = time, hr thk = tube thickness, ft = Stefan-Boltzmann's constant ρ = density of tube material, lb/ft3 αAcp = equivalent cold plane area in radiant section of fired heater, ft2 Mohit Thakur is a Sr. Process Engineer at Bechtel Corp. in Houston, Texas. He has more than 10 years of experience in Process and equipment design, engineering, and dynamic simulations in the Oil, Gas and Chemicals industries. He graduated with a thesis in Chemical Engineering from Lamar University, Beaumont, TX. He can be reached by email: mthakur@bechtel.com Mahmood Izadi is a Sr. Packaged Equipment Specialist at Bechtel Corp. in Houston, Texas. He has more than 15 years of experience in the oil, academia, and manufacturing. He graduated with a Ph.D. in mechanical engineering from Clarkson University, Potsdam, NY. He has publications in fouling mitigation in heat exchangers, power plant optimization, and heat exchanger design. He can be reached by email: mizadi@bechtel.com Rimon Vallavanatt is Sr. Principal Engineer at Bechtel Corp. in Houston, Texas. He has more than 40 years of experience in the design, engineering and troubleshooting of fired heaters, thermal oxidizers, boilers and flares. He graduated from the University of Kerala in India with a degree in mechanical engineering. He also received a degree in industrial engineering from St. Mary's University in San Antonio, Texas. Mr. Vallavanatt is a registered professional engineer in the state of Texas, and he has served on the American Petroleum Institute's subcommittee on heat transfer equipment for the past 30 years. He can be reached by email: rvallava@bechtel.com References [1] S. Ahamad, M. Izadi, R. Vallavanatt, ‘Applying transient analysis to estimate fail safe operating time of heating coils in LNG plant', Proceedings of IMECE2015, 2015 ASME International Mechanical Engineering Congress and Exposition, November 13-19, 2015, Houston, Texas, USA [2] L. A. Mekler, Process design of tubular heaters, Transactions 5 |
