Distress Initialisation

When the model initialises (start of model run), values for AADI, IV and T100 needs to be estimated for each model segment. This process proceeds in two stages:

Stage 1 – Estimate AADI, IV and T100

At the start of the model run, the expected surface life is determined using the expected life value held in Juno/RAMM. Next, the probability of observing distress is calculated using the relevant predictor variables such as ADT, heavy vehicles, Urban/Rural class, etc. With the expected life and the probability of observing the distress now calculated, the “expected” values for AADI, IV and T100 are calculated and stored against specific keys in the Cassandra JFunction model.

Stage 2 – Calibrate based on observed data

At the start of the model run, we have the advantage of knowing what the current surface age and percentage distress is for each segment. This provides the opportunity to calibrate the expected values calculated in Phase 1 as explained above.

For example, if the expected AADI is 8 years and the initial surface age is 9 years without any sign of distress, then the expected AADI can be adjusted to take the lack of distress into account. Similarly, if the expected AADI is 8 years but the segment already shows distress at 4 years’ age, then the expected AADI needs to be adjusted.

This calibration based on the known values at the start of the run can be automated by means of the Cassandra JFunction known as an “S-Curve Calibrator”. This function takes as inputs:

  • observed surface age
  • observed percentage distress
  • expected values for AADI, IV and T100 (from Phase 1)
  • minimum and maximum values for AADI, IV and T100

Juno Cassandra will then automatically determine if calibration is necessary and adjust the expected values for AADI, IV and T100 to fit the observed data while constraining the values within the minimums and maximums provide. Details of how the S-Curve Calibrator function works in Cassandra can be found here.