Bao Zhenyu; Wang Ning; Zhang Hongfei; Duan Yongfeng; Yu Fengchang

(R&D Center of Luoyang Technology, SINOPEC Engineering (Group) Co., LTD., Luoyang 471003)

Abstract: Ammonium chloride corrosion in the reactor effluent system remains to be a barrier for the safe operation of the ebullated-bed hydrogenation unit as impurity content is higher compared with that of the ordinary hydrogenation units.In this research, a Sinopec envisaged case study was conducted on feed oil containing 2.92 μg/g of Cl and 0.38% of N,because the impurity content of feed oil was representative in residue oil. The deposition patterns in heat exchangers were investigated by changing process variables, and then water wash strategy was optimized in view of the relative humidity to obtain a minimum water flowrate, and finally the process optimization suggestions concerning the operation of heat exchangers were proposed. Results show that with the measured content of nitrogen and chlorine in the feed, the NH4Cl deposition temperature of hot high-pressure vapor and hot low-pressure vapor was 223.4 °C and 173.7 °C, respectively,and the minimum water wash flowrate for heat exchangers of hot high-pressure vapor with mixed hydrogen and hot low-pressure vapor with cold low-pressure oil was 38.0 t/h and 5.4 t/h, respectively. Water wash should be carried out intermittently upstream of the heat exchanger tube passes. In consideration of energy consumption, it is recommended to reduce the tube pass outlet temperature of the above heat exchangers to 240 °C and 190 °C, respectively.

Key words: ebullated-bed hydrogenation, corrosion, ammonium salt, water wash, deposition temperature

1 Introduction

Ebullated-bed hydrogenation (EB-RDS) process,designated to solve the problem of heavy oil utilization, is becoming prevailing in China’s re fineries[1]. However, the equipment plugging and corrosion caused by ammonium chloride (NH4Cl) gradually become the main problems that trouble the EB-RDS unit, owing to the poor feed oil properties[2-3].

Up to now, researchers have conducted extensive studies on the corrosion behavior of NH4Cl. Toba, et al.[4-5]claimed that titanium alloys and alloys with a pitting resistant equivalent number (PREN) of 40 or higher were recommended in dealing with concentrated NH4Cl solution, especially under the “boil-dry” condition. Owing to the fact that the reactor effluent inevitably contains a certain amount of water[6], the NH4Cl deposit may absorb water in the vapor phase when the relative humidity(R.H.) is greater than 10%[7-8], and can cause the underdeposit corrosion[9]. Metal corrosion caused by NH4Cl is most severe at a relative humidity of around 60%, when the temperature is 80 °C. Lin and Lagad[10]found the relationship between corrosion and the critical relative humidity (R.H.) at a higher temperature (204 °C), and they argued that severe corrosion could still be initiated when the relative humidity was below the critical R.H. of NH4Cl, because deliquescence of FeCl2might occur.

Corresponding countermeasures for preventing NH4Cl corrosion include source control of chlorides, process design optimization, and materials selection. Shargay,et al.[11]suggested to optimize the operation of water wash and temperature control, and proposed several methods to reduce the source of chlorides in the feed,such as optimizing the electric desalting and dewatering operation[12-14], and improving the performance of the dechlorination agent. Shargay and Marciniec[8]illustrated the criterion for the selection of water wash method. For“dry” reactor effluent streams with a relative humidity of less than 10%, intermittent water wash may be sufficient, while for “wet” reactor effluent streams (with a relative humidity of more than 10%) and cases that need frequent intermittent water wash (more than once in every 3 months), continuous water wash should be used.When upgrading the materials, alloys with high pitting resistance are recommended[15].

However, there are few reports in the literature describing corrosion problems coupled with the suggested mitigation strategies for EB-RDS. Owing to the specific operating conditions, deposition of NH4Cl in EB-RDS displays different features as compared to traditional hydrogenation units, and the countermeasures for prevention of plugging and corrosion caused by NH4Cl remain vague and insufficient. This research focuses on one case of EBRDS used in Sinopec Corp., Ltd., and mainly deals with the deposition pattern of NH4Cl, water wash strategy,and corrosion prediction. It is known for the first time that related research has been done, which attempts to provide a reference for the process optimization and safe operation, so that it is meaningful for technicians to maintain equipment integrity and to improve the overall operating cycle of the process unit.

2 Basic Information

The EB-RDS studied in this paper has four separators for the reactor effluents, with the process scheme shown in Figure 1. E105 and E108, the heat exchangers of hot high-pressure vapor (HHPV) with mixed hydrogen (H2)and hot-low pressure vapor (HLPV) with cold lowpressure oil (CLPO), respectively, are selected as the research objects, as they are prone to be affected by NH4Cl. Relevant process stream information is listed in Table 1, where chlorine content (wCl) and nitrogen content(wN) in the feed are obtained from laboratory analyses.Corresponding temperature and pressure data for E105 and E108 are listed in Table 2 and Table 3, respectively.

Table 1 Relevant process stream information

3 Analysis Procedures and Methods

In this paper, research is conducted on the features of NH4Cl salt formation and the water wash process, and the research procedure is shown in the following flowchart(Figure 2), in which NH4Cl deposition process calculation and corrosion rate estimation are illustrated in subcharts(Figure 3 and Figure 4).

Figure 1 Process scheme of reactor effluent system in the FB-RDS

Table 2 Temperature and pressure data for E105

Table 3 Temperature and pressure data for E108

Figure 2 Flowchart of research procedure

Figure 3 Subcharts for NH4Cl deposition calculation

3.1 NH4Cl deposition temperature (TNH4Cl) calculation

TNH4Cl(°C) calculation is conducted based on the given relationship provided by API RP-932B[7]:

Figure 4 Subcharts for corrosion rate estimation

It can be seen from Eq. 1 thatTNH4Clis determined by the partial pressure of NH3(pNH3, kPa) and HCl (pHCl, kPa),therefore, eitherwClorwNis investigated to fully figure out the characteristic of deposition. During this process,the conversion rates of N and Cl in the feed are reckoned as 41% and 100%, respectively, according to design speci fications. Under different conditions, the conversion rates of N and Cl may vary, depending on the feedstock and catalyst property.

The values ofpNH3andpHClare obtained from the simulation results provided by the Pro II 8.0®process simulator, while the thermodynamic equation of Grayson-Streed is applied.

3.2 NH4Cl deposition rate (ΔnNH4Cl) calculation

ΔnNH4Clis calculated by the following equation[16]:

where,nVis the total vapor molar flowrate, (mol/h);Pis the total pressure, (mmHg);Kis the reaction equilibrium constant, which is only determined by the temperatureT(°C), and can be expressed as the following equation:

3.3 Minimum appropriate flowrate of injected water (M)

The value ofMis obtained with the help of Pro II 8.0 software. According to API RP-932B[7], the rule-ofthumb water wash flowrate should have at least 25% of remaining water after the injection point, so that sufficient free water is available for dissolution. Therefore, liquid water proportion is to be investigated with the variation of water wash flowrate. A theoretical estimation equation for the value of M is also provided by API RP-932B, whereas Pro II simulation is adopted in this part for a higher degree of accuracy.

3.4 Selection of water wash strategy

In this research, the values of TNH4Clare first compared with the tube pass outlet temperatures of the heat exchangers (Tout), and then the R.H. values of the streams at Toutare compared with the value of 10%. The speci fic criteria for the selection of the water wash strategy are listed as follows.

·When TNH4Cl＜ Tout, no need for water wash;

· When TNH4Cl≥ Tout, and R.H.≥ 10%, a continuous water wash is suggested;

· When TNH4Cl≥ Tout, and R.H.＜ 10%, an intermittent water wash is suggested;

The value of R.H. is calculated by using Eq. 4:

where, yH2Orepresents the H2O mole fraction; P represents the pressure of the stream, MPa; and psrepresents the saturated vapor pressure of water, MPa.

3.5 Concentration of NH4Cl in aqueous solution(cNH4Cl)

When the values of wCland wNare fixed, the value of cNH4Clis dependent on the water wash strategy and water wash flowrate. Generally, the water wash strategy applied in re fineries can be divided into two types, i.e., continuous and intermittent. For continuous water wash, cNH4Clcan be directly calculated according to Eq. 5. For intermittent water wash, c'NH4Clis an average value obtained by Eq.6, which in fact is a formula based on the actual water wash flowrate and the total ammonium salt accumulation during the water wash interval.

For Eq. 5 and Eq. 6, M is the liquid water flowrate, t/h; t is the water wash interval, d; ΔnNH4Clis the deposition rate of NH4Cl, g/h; and t' is the duration of water wash, h.

3.6 Corrosion rate estimation

By using several parameters related to the stream downstream of the water injection point, the corrosion rate of heat exchanger tubes ASTM A335 P12[17]can be estimated by Eq. 7, which is derived from regression using the actual production data of the refinery as well as the laboratory corrosion evaluation data of NH4Cl solution.

where, Rcorris the corrosion rate of A335 P12, mm/a;cNH4Clis the concentration of NH4Cl in the stream, %,(suitable for the scope of application: 0＜ cNH4Cl＜ 5%); T is the stream temperature, °C, (suitable for the scope of application: 160 °C≤ T≤ 200 °C); v is the stream flow velocity after water wash, m/s, (suitable for the scope of application: 2 m/s ＜ v＜ 7 m/s).

Based on the type of the heat exchangers: E105 (DIU1400-18.9/20.3- 360- 6/19- 2) and E108 (BIU900- 4.0- 215-6/19- 2I), and the thickness of the heat exchanger tubes(2 mm), the value of v can be calculated by the volumetric flowrate and the total cross-sectional area of the tubes.

4 Results and Discussion

4.1 TNH4Cl calculation

According to the connection between facilities, a simulation model (shown in Figure 5) is established to obtain the molar flowrate of NH3and HCl in the gas phase. Herein, the stream S5 and stream S8 are to be focused when the NH4Cl deposition position for HHPV and HLPV is ascertained, respectively.

4.1.1 E105 (heat exchanger of HHPV with mixed H2)To clarify the relationship between TNH4Cland wCl,calculation is performed with fixed molar flowrate of NH3in vapor, with the results shown in Table 4, in which the values of TNH4Clare subsequently calculated and added.

It can be seen from Table 4 that within the range of the wClinvestigated, the value of TNH4Cllies between 204 °C and 240 °C, which means that the initial deposition position is inside the E105 tube. Therefore, a water injection point needs to be added before the E105 tube pass inlet.

Figure 5 Simulation model for TNH4Cl calculation

Table 4 Calculation results of pNH3, pHCl and TNH4Cl in HHPV

To further figure out the correlation ofTNH4ClandwN,calculation is implemented with a fixed HCl molar flowrate in vapor (corresponding to the value ofwCl:.9 2 μg/g), with the results presented in Table 5.

Similar trend can be observed from Table 5 thatTNH4Clrises with the increase ofwN, provided thatwClis fixed.

For comprehensive observation of howTNH4Clis in fluenced by bothwNandwCl, Figure 6 is drawn. It can be seen that with the increase ofwNandwCl, the value ofTNH4Clkeeps rising, but the rising trend gradually slows down. Compared withwN, the value ofwClis fairly low, which makes it a key factor affecting the value ofTNH4Cl. Moreover, within the investigated ranges ofwNandwCl, the value ofTNH4Clfalls in between 190—260°C. It means that the initial salt formation is most likely to take place inside the tubes of E105. Therefore, the data in Figure 6 can be used to determine the probable location of salt formation at differentwNandwCl, and can further provide a guide for impurity control or design improvement.

Figure 6 Relationship between TNH4Cl in HHPV vs. wN and wCl

4.1.2 E108 (heat exchanger of HLPV with CLPO)

The relationship betweenTNH4ClandwClis shown in Table 6, in which the molar flowrate of NH3in vapor is fixed.

It can be seen from Table 6 that within the range ofwClinvestigated, the value ofTNH4Clin HLPV lies between 158 °C and 187 °C, which means that the initial deposition position is mainly inside the E108 tube. Therefore, a water injection point needs to be added before the E108 tube pass inlet.

For a fixedwClvalue of 2.92 μg/g, the variation ofTNH4ClwithwNis investigated, with the results shown in Table 7.Table 7 illustrates thatTNH4Clincreases from 165 °C to 193 °C when the value ofwNis increased from 0.1% to 2%.Binary interactions ofwClandwNon the value ofTNH4Clcan be illustrated by Figure 7. Based on the ranges ofwClandwNstudied,TNH4Cllies between 145—205 °C,i.e., except for the cases in whichwClandwNare simultaneously low (the dark blue and navy blue areas in Figure 7), the ammonium salt will form inside the tubes of E108.

4.2 ΔnNH4Cl calculation

4.2.1 E105

Figure 7 Relationship between TNH4Cl in HLPV vs. wN and wCl

As HHPV flows through the heat exchanger tubes, its temperature drops gradually. During this process, the NH4Cl salt will crystallize whenTNH4Clis reached, and the value of ΔnNH4Clis obtained as 25.70 mol/h (1.37 kg/h)using Eq. 2 and Eq. 3. Herein, calculation is performed at the outlet temperature of the tube pass, so that the total amount of salt in the heat exchanger can be determined. The value ofwClandwNis taken as 2.92 μg/g and 0.38%, respectively.Calculation result shows that the deposition phenomenon of NH4Cl in the E105 tubes is fairly prominent.

4.2.2 E108

Based on the same method with the above calculation,the value of ΔnNH4Clis 63.1 g/h in the E108 tubes, which is prominently smaller than that of E105. This result meansthat during the operation of the hot high-pressure separator(HHPS), most of the NH3and HCl species enter the gas phase. In contrast, low flowrates of NH3and HCl are found in the hot high-pressure oil (HHPO), which enters the hot low-pressure separator (HLPS). Consequently, a relatively smaller value of ΔnNH4Clis acquired in E108 due to less NH3and HCl species in HLPV.

Table 7 Calculation results of pNH3, pHCl and TNH4Cl in HLPV

4.3 Water wash calculation

A mixer and a water injection line are factitiously added at the tube pass entrance of E105, in order to make clear the in fluence of water wash process on HHPV. Similarly,the same is done for E108 when investigating HLPV.

4.3.1 Water wash calculation for HHPV

Figure 8 displays the simulation model built, where

·‘WATER’ is the water injection line;

·‘S5.’ is the tube pass inlet line of E105, also known as HHPV, which is equivalent to the stream ‘S5’ in Figure 5;

·‘S11’ and ‘S12’ are the mixed stream.

Figure 8 Model for water wash calculation of HHPV

The calculation result is shown in Figure 9. It is shown that the water wash flowrate can satisfy the above requirement at a value of 38.0 t/h. The temperature of‘S11’ is 214.4 °C.

Figure 9 In fluence of water wash flowrate to liquid water proportion and ‘S11’ temperature■—Liquid water proportion; ●—Temperature

4.3.2 Water wash calculation for HLPV

Figure 10 displays the simulation model built, where

·‘WATER1’ is the water injection line;

·‘S8.’ is the tube pass inlet line of E108, also known as HLPV, which is equivalent to the stream ‘S8’ in Figure 5;

·‘S13’ and ‘S14’ are the mixed stream.

Figure 10 Model for water wash calculation of HLPV

The calculation result is shown in Figure 11. It can be seen that the water wash flowrate can satisfy the above requirement at a value of 5.4 t/h. The temperature of ‘S13’is 170.0 °C.

Figure 11 In fluence of water wash flowrate to liquid water proportion and ‘S13’ temperature■—Liquid water proportion; ●—Temperature

4.4 Selection of water wash strategy

4.4.1 E105

The R.H. of stream S6 (see Figure 5) at the outlet temperature of the tube pass is calculated, and the result is 1.13%, which indicates that the stream in the tubes of E105 is “dry”, and the ammonium salt will not absorb moisture, so it is suggested that intermittent water wash is applied in front of the E105 tube pass inlet.

4.4.2 E108

The R.H. of stream S9 (see Figure 5) at the outlet temperature of the tube pass is calculated, and the result is 0.45%. Similarly, intermittent water wash is suggested to be applied ahead of the E108 tube pass inlet.

4.5 cNH4Cl calculation

In this part,c'NH4Clis calculated under the circumstance that the water injection point should be located in front of the tube pass inlet of E105 and E108. The calculation results are important for estimating the corrosion rate.

4.5.1 E105

According to the values of ΔnNH4Cland M obtained previously, the value of c'NH4Clis calculated using Eq. 6 when intermittent water wash is adopted, and the calculation result is 5.58%. However, it is rather changeable due to different operating habits. Upon assuming that water wash is conducted every 3 months(90 d), and the operation lasts for 3 h each time, the calculation is demonstrated as follows.

4.5.2 E108

Referring to the above calculation method, the value of c'NH4Clis obtained, and the calculation result is 1.80%.

4.6 Rcorr estimation

The value of Rcorris estimated using Eq. 7, and it is based on the situation that no corrosion inhibitor is added and no deoxygenation operation is done. In fact, the severity of the corrosion problems will be higher than the predicted value when intermittent water wash is adopted,owing to the under-deposit corrosion of NH4Cl caused by remaining water after the water wash.

4.6.1 E105

During intermittent water wash, corrosion of the tubes in NH4Cl solution can be described by an average value of Rcorr, which is obtained as 19.77 mm/a.

Corrosion happens each time when water wash is implemented upstream of E105, so an equivalent annual corrosion rate R'corris proposed to show the thickness reduction rate as a whole, and the value of R'corris 0.027 mm/a. In this process, no corrosion is supposed during the period without water wash upstream of E105. It can be seen that the value of R'corris rather low, however,attention should be paid to the under-deposit corrosion.

4.6.2 E108

For E108, intermittent water wash seems to be relatively easy as the value of ΔnNH4Clis much lower than that of E105. Corresponding values of Rcorrand R'corrare estimated as 10.89 mm/a and 0.015 mm/a, respectively.

4.7 Process optimization suggestions

4.7.1 E105

Based on the measured value of wCl(2.92 μg/g), the inlet temperature of the E105 tube pass (316.68 °C) is much higher than the value of TNH4Cl(223.4 °C). Water wash in front of E105 introduces considerable high-grade energy loss. Since the salt formation in the tubes is unavoidable,it is recommended to use an additional heat exchanger or optimize the heat exchange network to reduce the tube pass inlet temperature of the E105 to 240 °C, so that the energy loss, the water flowrate, and the operating load can be simultaneously reduced. According to the calculation,the minimum water wash flowrate (with 25% of liquid water remaining) can be reduced to 20.2 t/h, which means that 46.8% of the water and 15.11 MW of the energy are saved.

4.7.2 E108

With regard to the above water and energy conservation strategy, the inlet temperature of the E108 tube pass(340 °C) is suggested to be lowered to 190 °C. By adopting this strategy, the minimum water wash flowrate is lowered to 1.8 t/h, which indicates the possibility that 66.7% of the water amount and 2.19 MW of the energy are saved.

5 Conclusions

This research deals with the NH4Cl deposition behavior,the corrosion prediction, as well as the process design optimization in EB-RDS unit. The main conclusions are drawn below:

(1) NH4Cl is likely to deposit inside the tubes of both HHPV and HLPV last stage heat exchangers if N content and Cl content in the feed are greater than 2 μg/g and 0.4%, respectively. Once the NH4Cl deposition occurs,the deposition rate is high enough to cause the plugging of heat exchanger tubes.

(2) A water injection point should be added at the tube pass inlets of the “potentially risky” heat exchangers, and it is suggested that the Grayson-Streed thermodynamic equation be applied when calculating the minimum water wash flowrate to ensure 25% of liquid water remaining at the water injection point. The water wash strategy should be prudently adopted by referring to the relative humidity of the stream at the tube pass outlet.

(3) For both safety and energy-saving considerations, the temperature of the water injection point should be 15—20 °C higher than the NH4Cl deposition temperature. It is recommended that the heat exchanger network should be optimized before adding the water injection point.

Acknowledgements: This research is supported by the Scienti fic and Technological Development Project of Sinopec:‘Comprehensive Processing and Related Technology Research of Heavy Oil Ebullated-Bed Hydrogenation’, Contract No.118013-4.