pt q&a
More answers to these questions can be found at www.digitalrefining.com/qanda
Q What methodologies are scaling up to detect deviations in plant operations? A Philippe Mège, Digital Factory Services Manager, Philippe.mege@axensgroup.com, and Pierre-Yves Le-Goff, Global Market Manager, pierre-yves.le-goff@axensgroup. com, Axens Data quality is essential for reliable decision-making, devia- tion detection, and process optimisation. As highlighted by a recently published Axens white paper, robust data cleaning and normalisation are critical first steps. Machine learning models, such as random forests, can estimate missing vari- ables needed for accurate normalisation. Once normalised, statistical methods like Mahalanobis distance and Hampel filters effectively identify outliers, which can be visualised on original time series for easy interpretation. Monitoring moving averages and applying normality tests helps detect gradual drifts in process param- eters. Dimensionality reduction techniques like Principal Component Analysis (PCA) reveal changes in parameter correlations over time, while clustering methods such as KMeans, guided by the elbow method, extract meaningful patterns from time series data without requiring labelled inputs. The unsupervised nature of these techniques enables rapid deployment and actionable insights without extensive data labelling. Integrating these methods, available on Axens Connect digital platform, within automated control systems and combining them with domain expertise and AI enhances scalability, accuracy, and reduces false positives, supporting proactive plant operation management. Connect is a mark of Axens. Q What strategies can be optimised to prevent margin leakage in plant operations? A Melvin Berrios-Soto, Product Marketing Manager, Emerson’s Aspen Technology business, melvin.berrios- soto@aspentech.com Margin leakage in plant operations can result from a range of factors, including operational inefficiencies, suboptimal feedstock utilisation, and limited coordination between production units. To address these challenges and improve profitability, dynamic optimisation solutions such as Aspen’s proprietary GDOT can be strategically implemented on top of an advanced process control (APC) layer. APC solutions have traditionally been used to improve unit performance by operating closer to process constraints. These systems help maintain throughput and product qual- ity while managing energy consumption. However, APC applications often function independently, relying on delayed feedback from other units. Manual adjustments to APC limits based on this feedback can lead to misalignment, creating
Planning / scheduling
Month/ weeks/
Updated model constraints True plant constraints
days/ hours
Aspen GDOT™
Realtime
APC Crude unit 1
APC Crude unit 2
APC Hydro treater
DCS platform
Running to the limits of performance - 24x7, minute-by-minute
Figure 1 Dynamic optimisation solution architecture
a gap between planned and operations that contributes to margin leakage. To enhance coordination and reduce these inefficien - cies, dynamic optimisation technologies like GDOT can be introduced (see Figure 1 above). The technology facilitates real-time alignment across multiple process units by auto- matically generating optimal targets for APC systems. It uses reconciled plant data and economic drivers to ensure that control decisions reflect current operating conditions and market dynamics. This closed-loop approach supports more consistent performance and enables faster responses to changes in demand, feedstock quality, or equipment sta- tus. In practice, implementations of GDOT have shown mea- surable financial benefits, including margin improvements of $4-$10 million annually in middle distillate refining and $6-$10 per ton of ethylene in olefins production. Integrating GDOT with APC systems can help reduce operational silos and improve overall plant coordination. This combined strategy supports more efficient operations and may contribute to improved margin retention, particularly in environments characterised by variability or complexity. GDOT is a mark of Aspen. A Philippe Mège, Digital Factory Services Manager, Philippe.mege@axensgroup.com, Pierre-Yves Le-Goff, Global Market Manager, pierre-yves.le-goff@axensgroup. com, and Romain Roux, Vice-President, Decarbonisation & Consulting, Axens, romain.roux@axensgroup.com In today’s volatile refining landscape, margin preservation is not just a financial imperative; it is a strategic necessity. One of the most effective levers to prevent margin leakage is the early detection of performance deviations. By continu- ously monitoring unit performance and comparing it against high-fidelity models, Axens’ digital platform, Connect’In, enables refiners to detect subtle degradations, such as
5
PTQ Q4 2025
www.digitalrefining.com
Powered by FlippingBook