Filter
Search...
Asset Type
Asset Name
Asset Type
Simulation Status
Days below...
Last Simulati...
Tuning Status
Average Total Simulati...
Last LP Health...
Avg Total LP Error
Avg Data Quality...
LP Status
Calibration
!
Tuning Required !
Needs
Update !
Simulation Requires Calibration
01-15-2018 Required
67
FCCU 101
FCC
12 years ago 23
3.7
18
2.2
Insu cient Number of Cases
Calibration
!
N/A ?
89
FCCU 401
FCC
Tuned
66
1.2
N/A
N/A
Required
LP Submodel Requires Update
Calibration
!
Tuning Required !
Needs
Update !
01-15-2018 Required
67
FCCU 501
FCC
11 years ago 21
4
18
2.1
Calibrated After
1
1 - 3 of 3 items
Filter
Clear All
Figure 1 Asset overview dashboard showing model health scores, calibration requirements, and data-quality indicators
a new generation of automated maintenance applications integrates process simulation, machine learning (ML), and cloud computing to ensure that digital twins transform from a periodic, labour-intensive recalibration task into a contin- uous, self-monitoring lifecycle management system that keeps models remain aligned with live plant conditions. This automated approach creates a continuous feedback loop that monitors model ‘vital signs’, such as prediction accuracy and data reliability. This allows engineers to inter- vene before deviations compromise performance. Unlike purely statistical black-box tools, this framework applies first-principles rigour. This preserves engineering credibility while ensuring application-guided calibrations accelerate tuning without losing the physical meaning of the process variables. Cloud-based execution further enables this architecture to scale across multiple units and data sources. As shown in Figure 1 , the asset overview dashboard con- solidates the status of multiple process models to display calibration status, health indices, and data-quality scores. Engineers can quickly identify which units are drifting out of tolerance and intervene before performance degradation occurs.
To complement this high-level view, Figure 2 illustrates how specific process variables impact overall model health over time. Each colour band represents a key process var- iable contributing to the health score, allowing users to visualise which parameters are degrading and when recali- bration is required. The automation process follows a structured method- ology from configuration to execution, beginning with the configuration of model performance indicators (MPIs), data quality parameters (DQPs), and key performance indica - tors (KPIs), along with site-specific error tolerances. Once deployed, the system performs two primary functions: Autonomous degradation detection : Health indices quan- tify deviations between simulated and measured data. When thresholds are exceeded, the system triggers automatic alerts, shifting the engineer’s role from data analyser to action taker. Automatic calibration and tuning : Leveraging parame- ter optimisation, sensitivity analysis, and cross-prediction, the software evaluates calibration quality across multiple datasets. Parallel cloud computing allows rapid tuning cycles, while open interfaces support scalable deployment across multiple units or sites.
FCCU 101 FCC
LATEST DATA QUALITY
SIMULATION STATUS
LP SUBMODEL STATUS Needs Update !
31/12/2013 77
Last Calibration Date: 15/01/2018 Needs Calibration !
AVERAGE DATA QUALITY 02/10/2013 to 31/12/2013 67
SIMULATION HEALTH SCORE 23
LP SUBMODEL HEALTH SCORE 18
Analyse Calibration
SIMULATION ERROR SCORE 3.69 !
Performance Overview
Health Score Explorer Health Score Settings Asset Settings
Message Centre
Risk Contribution to Simulation Health Score
Filter
This chart shows the risk score of the Simulation Model where 100 is the maximum risk. It is recommended to recalibrate if the total risk exceeds the risk threshold.
Clear
Time Range
100
Start Case
80
End Case
60
MPI Plots
Clear
40
Sort By
MPI Risk
20
10 # of plots Display
0 03.32.03 003
Nov
Dec
8
15
22
29
8
15
22
29
8
15
22
29
Filter
Clear All
Aggregate Risk Score MCB Mass Flow
Remaining MPIs value
Risk Score Threshold
LCO Sulphur Content
FCC Standard FCC Bottoms
FCC C Standard Naphtha
C + Gasoline Mass Flow FCC Apparent Conversion
FCC Propene
LCO Mass Flow FCC Standard Cycle Oil
FCC Coke
Figure 2 Detailed health-score dashboard showing data contributions to model performance and threshold tracking over time
54
PTQ Q2 2026
www.digitalrefining.com
Powered by FlippingBook