What is Network Model Calibration and why is it useful?
What is Model Calibration?
When referring to the calibration of our MV-LV network-wide power flow model, we mean the ability to compare a snapshot of modelled power flow results against real-world (historical) actuals.
The calibration process utilises a snapshot model run. This approach limits the primary unknowns within the model to its impedance parameters and the off-load tap changer positions. This focused approach allows users to gain a clear view of modelled versus measured voltages at the energy consumer (customer) level.
Typically, this calibration is employed to determine or validate off-load tap positions, which are often unknown or unverified within most Distribution Network Service Provider (DNSP) network datasets. By modelling various historic timestamps, confidence in the inferred tap positions can be established, and model uncertainty can be measured across a range of historical network states.
Why is Model Calibration important?
This comparison is vital for several reasons:
- Identify Model Uncertainty: Calibration helps quantify the level of uncertainty inherent in the network model.
- Comparative Accuracy Assessment: It allows for an evaluation of model accuracy across different sections of the network.
- Refine Model Unknowns:
- Where off-load tap positions are unknown, calibration can help infer the currently in-service positions, thereby refining a significant model uncertainty.
- Where off-load tap positions are known, significant ΔV can indicate areas with inaccurate impedance or connectivity modelling.
What does a well calibrated model provide?
- An Accurate Baseline: Essential for reliable operational decision-making, including subsequent transformer tap setting adjustments.
- Enhanced Voltage Analysis: Improved accuracy in voltage profile studies and the assessment of voltage regulation strategies.
- Validated Platform: A trustworthy model for network planning, "what-if" scenario testing, and predicting the impact of network changes.