PTQ Q3 2024 Issue

Cost dierence

Delta (2.4 – 2.5)

3.5

2.5 2.4

3.0

Steam FG

2.5

2.0

1.5

1.0

0.5

0.0

-0.5

9.7

9.9

10.1

10.3

9.7

9.8

9.9

10.0

10.1

10.2

10.3

Hydrogen capacity(ton/h)

Hydrogen capacity (ton/h)

Figure 6 Differences in natural gas (NG), fuel gas (FG), and steam amounts at different S/C ratios

Assessment of model performance Analytical models underwent detailed validation, including train-test, time-series, and holdout validation methods. These approaches ensured robustness and reliability by evaluating model performance on unseen data, accounting for temporal trends, and testing generalisability. Successful models then entered a hot trial phase for live observation, confirming their effectiveness in real-world conditions. These models were deployed for business unit use upon proving their efficiency, demonstrating a meticulous yet practical approach to model validation and application. Following the rigorous validation stages, the models were subjected to a hot-trial phase, where they were monitored live before being accepted. Evidently, the models effectively captured the overall trend, a critical aspect in the optimisa- tion phase. This step in the S/C optimisation project at the HGU unit underscores the importance of real-world testing in validating the effectiveness of the optimisation module’s models. Results After running the models, we were able to see how both the total unit cost per hydrogen and unit variables such as fuel gas, steam, and natural gas change at different S/C ratios. Although methane slip increases as the S/C ratio decreases, the cost lost by increasing natural gas is lower than the cost gained by decreasing fuel gas and increasing steam. Therefore, it is more cost-effective to operate with low S/C under these conditions. Although the cost differ- ence between the case where the S/C ratio is 2.4 and the case where it is 2.5 changes over time, depending on the other unit variables and unit capacity, it has always been negative since the commissioning of the model. The lower cost per hydrogen produced at a low S/C ratio, which we see in the model outputs, is also confirmed by the actual data. As can be seen from the graph, operating at a low S/C ratio leads to hydrogen production at a lower cost than operating at a high S/C ratio on average. Since the unit variables and conditions may be different and the unit cost may vary according to the hydrogen capacity, we see

variations in the cost data. In order to make the right com- parison, it is necessary to compare the periods with similar hydrogen production and make the comparison based on average values. In this way, when we compare the past one-year data with the new three-month data, it is seen that operating at low values is quite profitable at the speci - fied hydrogen capacities. While the total cost per hydrogen varies with unit capacity, the impact of the S/C ratio on the variables also varies with capacity. As capacity increases, natural gas demand rises faster, steam production decreases faster, and fuel gas decline slows down more at low S/C than at high S/C. The differ- ences in natural gas, fuel gas, and steam amounts when operating at the same capacity at low and high S/C ratio are shown in Figure 6 . As unit conditions, capacity, and utility costs change, the model is continuously run within constraints in order to maximise objective function and rec- ommend the optimum ratio accurately. Mert Akçin is a Data Science and Analytics Senior Specialist at SOCAR Turkey, with experience developing soft sensors for business units in refinery and petrochemical complexes. He holds BSc degrees in both industrial engineering and business administration. İbrahim Bayar is a Process Monitoring and Optimisation Manager at SOCAR Turkey, where he is responsible for process activities at hydro- processing units and aromatic complexes. He holds a Masters in chem- ical engineering. Berkay Er is a Senior Digital Transformation Specialist at SOCAR Turkey, where he is responsible for implementation of digital solutions to enhance operational efficiency. He holds a BS degree in chemical engineering and has experience in process control applications. Gizem Kayar Öcal is a Lead Process Engineer at SOCAR Turkey, where she is responsible for hydrocrackers, hydrogen generation, kerosene hydrotreaters, and diesel hydrotreater units. She holds a BS degree in chemical engineering and has experience in process monitoring and optimisation. Muratcan Özpınar is a Process Chief Engineer at SOCAR Turkey, where he is responsible for the hydrocracker and HGU. He holds a BS degree in chemical engineering and has experience in process moni- toring and optimisation.

PTQ Q3 2024