On this tab you define the transition probabilities. The system has generated the required lines for you already.
For each transition probability record enter the probability that the equipment moves from one state (the From State) to another state (the To State). For each line the total must add up to 1.
For the weight record enter the perceived value you would like to attach to the equipment in that state. State 1 represents the best condition and usually gets a 1 therefore. The highest state in your matrix represents the worst condition and thus should get a lower weight.
Once the matrix is complete click the Generate Curve Points button. This button will now generate a new set of curve points based on the data entered in the matrix. The system will calculate the fraction of equipment in each state and determine the health index for each iteration, i.e., curve point, generated. The system will then transform the health index for each curve point into a condition score.
If your matrix has different zones, you can add and remove zones as required. Click Add New Zone to create the next zone for your matrix and click Remove Highest Zone to take one zone away. Each zone represents 5 iterations for the decay curve. If you have a decay curve for 30 years and the Curve Point Interval is set at 1, then Zone 1 represent the first 5 years, zone 2 represents year 6 through 10, zone 3 year 11 through 15, etc. Each zone has its own matrix to support an increase or a decrease in the rate of decay. The last zone entered determines all remaining curve points. Meaning that if zone 3 is the last zone, then the matrix for zone 3 determines the curve points from year 11 until year 30. Once you are done adding or removing zones and updating the matrix, click the Generate Curve Points button to refresh the curve points.
It is also possible to have the matrix automatically generated, but you must have collected enough condition score history for equipment for trustworthy results. Based on the provided history the system will calculate the matrix and the curve points automatically if you click the Generate Matrix and Curve Points button. This button will start the Regression Based Non-Linear Optimization process.
This process only works with 5 states and only for period or usage-based decay curves.
The system must have at least one history record for each iteration of the first 5
iterations, i.e., the matrix for zone 1 can be calculated.
Once started the Regression Based Non-Linear Optimization process will do the following:
-
Determine History Records Included - System will include history ranking records as requested on the Ranking History Selection tab of this screen.
-
Equipment is selected based on Class, Category or Equipment Code.
-
If Include Withdrawn is selected, then equipment with system status ‘Withdrawn’ is included in the history.
-
If Include Out of Service is selected, then out of service equipment is included.
-
History record Type must match your selection on the Ranking History Selection tab.
-
Condition Score on the history record falls on or between the Condition Score Start and the Condition Score End of the decay curve.
-
If Include Linear is selected, then history records where From Point and To Point are entered are included in the selection.
-
If any of the UDFs on the Ranking History Selection record are filled in, or selected / unselected for the check boxes, they must match with the corresponding UDF on the equipment history ranking record.
-
If the decay curve Type is usage based, the history records must reference the same meter UOM as the decay curve.
-
-
System generates new matrix. The system will:
-
Delete all Transition Probability records belonging to the decay curve.
-
Set the Weight record. If the current Weight record has data in one or more of the first five state fields and at least one weight is greater than zero and not all five weights are the same, then the system will leave the Weight record as is. Otherwise, the system will set the Weight record with the following values for state 1 through 5: 1.0, 0.84, 0.69, 0.46, 0 (zero).
Note that these are the default settings which were determined for power transformers. The only exception is State 5 (0.29). HxGN EAM will use 0 instead, because the ending condition score is defined on the decay curve and the process needs to trend towards that number. Setting the last weight to 0 achieves that. -
Start zone 1 Regression Based Non-Linear Optimization process for the first 5 iterations.
-
Add calculated zone 1 Transition Probability records to the matrix.
-
Repeat for zone 2 through a maximum of 99 zones if there is history data present for the zone’s corresponding iterations.
-
-
System generates new curve points automatically after the matrix is complete and the sets the header data to indicate the process is not running anymore and updates the Average Percentage Error. The lower this number the better match the system found between the history records and the resulting curve points.