Inspection Data validation using Kastner analysis Concept

Inspection is an interpretive process involving the collection and analysis of data, regardless of the objective or type of data being handled. The primary focus of any inspection process is to determine whether the data is accept or reject but there are most important question which is the basic foundation for our assessment process which is Data validity —that is, whether it represents true values or realistic measurements or nor.

Determining validity is not a simple "yes" or "no" decision. Instead, it depends on the required level of accuracy. In other words, the sensitivity of the assessment plays a significant role in defining validity. This is why it is critical to establish criteria to evaluate whether the data is acceptable based on the required accuracy (tolerance to true value)

we will discuss in brief the data validity only coming from NDE regardless the P.O.D (assume we select the highest P.O.D technique based on API-571,581,ASME PCC-3,ASME V,API-577or any other reference)

The Kastner Analysis Concept provides a structured approach to evaluate the validity of collected inspection data using the following summarized steps:

Step 1: Collect and Verify Data

1.Take a sample of your inspection data and perform verification using a secondary, more precise technique to obtain true reference values.

Example: Compare measurements from Magnetic Flux Leakage (MFL) or Inline Piping Thickness Pigging with Spot Thickness Measurements.

Step 2: Study Data Pattern and data performance related to actual values-

2.Use the collected data to calculate the following parameters (here we establish the process performance (MFL) variables)

a-Calculate the local Error at every spot (Spot reading – MFL reading)

b-Calculate the Absolute average error for all the collected data (give you indication to what extent we are far from the true in other words to what extent MFL represent about the true thickness values).

c-Calculate the Numerical average error (Bias) (give you indication the used technique give you over estimated measurements or under estimation measurements).

d-Calculate Std Deviation (give you indication about the measurement process variability I mean the measured value near to the true value with narrow bans or wide band).

Step 3: Establish the Validity Acceptance criteria -

e-Here the critical step (here we will establish acceptance criteria for our data) , what is the confidence factor which you need to rely on this data if this data is not critical may be confidence factor 75% will be enough (that is mean 75% of data will be reflect the actual case) for most cases conf factor is 95% and based on this required confidence level the tolerance of the error will be established using the equation

Confidence Intervals (error tolerance based on the confidence level target)-

Step 4: Automated Data visualization using tables and chart -

after apply the equation and establish the validity acceptance criteria so

Based on this Results we have to decide the MFL need to be adjusted or the operator skills may be not as the required level or the whole NDE process shall be changed with more accurate one.

Note-

Conclusion

The Kastner Analysis Concept provides a systematic approach to validate inspection data by quantifying its accuracy, bias, and variability. By applying these calculations, inspectors can determine whether the collected data meets the established accuracy criteria and can be considered valid for the intended purpose.

The process helps ensure that critical decisions are based on reliable measurements, enhancing the safety and integrity of assets.

This Approach are inherent approach in most of the AIMS/IDMS/CIMS software including condition assessment or risk assessment.

Decision making based on inspection mainly depending on the collected data validity