
FIGURE 1 IS THE STATISTICAL DATA FOR ANALYSIS
THESE NUMBERS ARE NOT BASED ON REAL DATA, SHOWN FOR EDUCATIONAL PURPOSES ONLY.

FIGURE 2 Data is Converted to Histogram

FIGURE 3 Histogram is Converted to Probability Distribution Function

FIGURE 4 Probability Distribution Function Shown in 3D Bubble Graph with Focus Bushing that has change of power factor from nameplate that is an outlier from the 128 bushing testing session history
Figure 4 above is using the visual display analysis looking only for an outlier bushing from the population.
If there is no outlier , it is hard to list all the bushings that are behaving differently from the rest of population. Central tendency method using mean and standard deviation will be the next logical method to find the list of bushing that need more attention.
Next step after having the list of bushing that needs more attention is using the trending analysis of power factor measurement, capacitance measurement, and watt measurement.
Finally additional tool such as Weibull Analysis of bushing failure will be helpful to prioritize the bushing overall strategy.
IMPORTANT: DON'T FORGET TO SEEK ADVICE FROM YOUR BUSHING VENDOR AND POWER FACTOR TEST INSTRUMENT VENDOR THEY ARE YOUR PARTNER.
FIGURE 5 Trending Analysis of Capacitance Reading Valid For OIP Bushing Only
FIGURE 6 Trending Analysis of Watt Reading Valid For OIP Bushing Only
By automating the data collection and statistical analysis, transformer engineer will have the necessary tool to monitor new Power Factor test measurement of transformer bushing for any outlier or listing of 2 or 3 standard deviation from their population of transformer bushing fleet to generate the list of bushing for more detailed analysis using trending method. This insight in turn will trigger maintenance strategy, operation strategy, replacement strategy, specification update, vendor management, purchasing strategy, and financial planning.
INVBAT.COM is helping students, teachers, and employees to create their own customized voice calculator website that will help them to enjoy and not fear a fast coming digital workplace like industrial internet of things and internet of things powered by machine learning, deep learning, and artificial intelligence. Imagine now you can solve similar problem as shown in above example in less than a minute by using your smarthphone, tablet, laptop or desktop computer. This computational speed is not possible before internet and cloud technology.
You might be interested to read more:
So what if I can't remember formula and forget how to use calculator?