Mar-2021

Increasing hydrogen plant capacity

A validated model enabled analysis of five potential revamps for an increase in hydrogen production

DIYAR KILIÇ MERT and EYÜP AZIZOG˘LU, Tüpras Izmit Refinery
EDA BAYRAKTAR, ERHAN ÖZSAG˘IROG˘LU, GÖZDE DÖNMEZ and SULTAN ORMAN, Tüpras Headquarters

Article Summary

The need for more hydrogen in refineries increases in response to new regulations or for more flexibility in processing heavier feeds. This need can be compensated with hydrogen rich gas (the platformer unit’s net gas) and with 99.9% pure hydrogen. As hydrogen rich gas (HRG) cannot be used in hydrocrackers, a revamp of the hydrogen unit is required in such cases. There are several capacity revamp options for steam methane reformers. This study analyses the effects of these options on hydrogen production capacity and on the equipment of the unit.

Basic hydrogen sources in refineries are steam methane reformers and platforming/reforming processes. For high pressure operation, high purity hydrogen is needed. The main route is steam methane reforming (SMR) of natural gas followed by shift reactions. These shift reactions can take place at high temperature (HTS), medium temperature (MTS), or low temperature (LTS). For these reactions, hydrocarbon feed is combined with HRG for pretreatment and steam.

Capacity revamp alternatives
There are several revamp alternatives for the SMR process. These can be for improving the reforming process and shift reactions, or for enhancing hydrogen recovery. To improve the reforming process, the endothermic reaction temperature can be increased or the reforming volume can be increased. For the shift reactions, a more active low temperature catalyst can be used.

Simulation and model validation
Simulation of the hydrogen unit
A commercial, rigorous simulation program was used to model the hydrogen unit. The configuration of the model covers reactors, heat exchangers, air coolers, fired heaters, columns, drums, pumps, valves and pipe segments.

The first step in simulation of the hydrogen unit is to carry out a number of selections and identifications. To be able to estimate the properties of hydrocarbon, water, and steam, appropriate property packages should be used. In this study, Peng Robinson and ASME steam fluid packages were used. The stoichiometry and equilibrium constant (Keq) of the SMR reactions and shift reactions are defined in the reactor section.

In the second step, the unit model is configured using field data (distillation, flow rates, operational data) and equipment process data sheets. When cresting a simulation model, the following steps and assumptions are required.

SMR
A rigorous fired heater is used to model the convection section of the SMR. An equilibrium reactor is selected to model the radiant section of the SMR since all reactions take place in the radiant section. The following parameters are defined or selected in the reactor model.
- Stoichiometry is defined.
- Approach to equilibrium (ATE), which is a parameter to measure catalyst performance, is selected according to actual SMR performance.
- Keq is selected from the program library.

Shift conversion reactor
An equilibrium reactor was selected for this study. The following parameters are defined or selected in the reactor model.
- Stoichiometry is defined.
- ATE is selected according to actual HTS catalyst performance.
- Keq is selected from the program library.

Waste heat boiler (WHB)
In the selected program, only a TEMA type heat exchanger can be modelled rigorously. Since the WHB is a non-TEMA type heat exchanger, a simple HEX model was used.

Pressure Swing Adsorption (PSA)
In the existing program, PSA cannot be modelled rigorously. Therefore, a component splitter was used. The components’ split ratios were selected according to actual plant efficiency.

Model validation
Simulation models can be validated based on unit design data (such as heat and mass balance) or field data. The main steps during validation are summarised below. Each validation step is performed according to the procedures detailed in the simulation.

Creating the design model
First, the simulation model is generated using the unit design data (heat and mass balance of the unit) and equipment process data sheets; afterwards, this model is checked with simulation outputs according to predetermined verification limits.

Generation of plant model
Before evaluating alternatives for a capacity revamp, the unit base model is created according to the eight-hour average values of a day when product and charge analyses are made and production was stable. The following assumptions are used in the model:
- The ratios of the radiant and convection section duties to the heater total duty are considered to be fixed. In normal operation, several factors such as ambient temperature and insulation properties affect this ratio.
- The SMR outlet temperature is fixed in order to match actual methane slip. To account for the heat loss between the temperature measurement location in the field and the catalyst, the SMR exit temperature is assumed to deviate from the actual temperature by 30-50°C.
- ATE = 0°C for SMR and ATE = 10°C for HTS.

Validating the model results
The critical parameters of a hydrogen unit for model validation are methane and CO slip, hydrogen production, steam to carbon (S/C) ratio, tail gas production, steam production and temperature, PSA efficiency, fuel gas consumption, boiler feed water (BFW) consumption, flue gas and bridge wall temperatures.

Acceptance of deviations between simulation models and actual site data
Comparing simulation model outputs with design and field data, results are acceptable if the difference is ±5%. Acceptance criteria are based on possible measurement errors and general refinery applications.

Checking fouling resistance in heater convection tubes
The first step in rigorous modelling of the convection chamber of the furnace is adjustment of the fouling resistance values in accordance with the field temperature profile (such as the outlet temperature values of the convection bank and flue gas). Convection zone tubes are 100% clean only when the fouling resistance value is equal to zero. It can be concluded from the temperature profile and the rate of steam production rate whether the convection tubes are clean.