Guide

What this tool does

You upload a spreadsheet of voltage readings taken from the historian. The tool checks how well each bulk supply point stayed inside its allowed voltage range over seven days, and produces the Section 9 page of the NITS Performance Report as a PDF.

Voltage is recorded once every minute. Seven days of one minute readings comes to 10,080 readings for each station. Every single one of those readings is compared against the allowed range. That comparison is where all three numbers in the report come from.

The allowed range, or band

Every station has a nominal voltage. That is the voltage it is meant to sit at. On the Ghanaian transmission network that is 69, 161, 225 or 330 kV.

The Grid Code accepts that the real voltage will never sit exactly on the nominal figure, so it allows it to drift up to 5% above or below. That window is called the band.

A reading sitting exactly on 152.95 or exactly on 169.05 counts as inside the band. The limits themselves are allowed.

The band does not come out of your spreadsheet. It is worked out purely from the nominal voltage you pick on the Stations screen. It is not the highest and lowest reading in the file. This is deliberate. The band is the standard the readings are being judged against, so it cannot be taken from the readings themselves. If it were, every station would pass its own band automatically and the report would be meaningless.

The three numbers

The report gives three figures for every station. Each one answers a different question about the same week.

PTVRGB: how much of the week was good

The full name is Percentage of Time Voltage Remained within the Grid Code Prescribed Band. Count how many readings were inside the band, divide by how many readings there were, and turn it into a percentage.

Higher is better. 100% means the voltage never once left the band.

NVE: how many times things went wrong

NVE stands for Number of Voltage Excursions. An excursion is a spell of time when the voltage was outside the band. NVE counts how many separate spells there were.

The word separate matters. If the voltage sits below the limit for forty minutes without a break, that counts as one excursion, not forty. A second excursion is only counted once the voltage has come back inside the band and then left it again.

A higher NVE means the voltage was misbehaving more often.

CDVE: how long the problems lasted

CDVE stands for Cumulative Duration of Voltage Excursions. It is the total time, in minutes, that the voltage spent outside the band over the whole week. Because readings are one minute apart, it is simply a count of how many readings were out of range.

Why three numbers and not one

Each one catches a different kind of poor performance. Imagine two stations that were both outside the band for 300 minutes in a week, so both have the same CDVE. One had a single fault that lasted five hours. The other flickered in and out of range sixty separate times. The first station has an equipment problem. The second has an instability problem. PTVRGB alone would rate them identically, but their NVE figures tell you they are completely different situations.

The worst performing BSP

The report has to name one station as the worst performer. The tool picks the station with the lowest PTVRGB, because that is the station whose voltage was acceptable for the smallest share of the week.

If two stations end up with the same or very nearly the same PTVRGB, the tie is settled by looking at NVE first and then CDVE. The station with more excursions, or with more total time out of range, is treated as the poorer performer.

A station whose signal is frozen is never picked, for the reason explained in the next section.

The Stations screen

Nominal voltage

The tool reads the nominal voltage out of each station's tag name. A tag reading AKOSOMBO 161 A7F VOLTAGE tells it that station is 161 kV, so it preselects 161 for you.

Check each one before you continue. If a nominal voltage is wrong then the band is wrong, and every figure calculated for that station is wrong with it.

The Signal column

This column is a health check on the incoming data, not on the voltage. It describes the week you have currently selected, so it updates the moment you pick a different start date. There are three things it can say.

Frozen is the one to take seriously. It usually means the real signal is not reaching the historian and a fixed placeholder value is being recorded instead. A station like that would otherwise score a perfect 100% and look like your best performer, so the tool marks it clearly and keeps it out of the worst performer decision.

The two switches

Drop 0 kV readings as lost telemetry

This is switched on by default.

Sometimes the historian records a voltage of 0 kV. A busbar genuinely at zero volts would mean a total outage, and those do not quietly last for hours at a time. In practice a run of zeros means the measuring equipment stopped reporting, not that the voltage collapsed.

When this switch is on, those readings are set aside. They are not counted as good and they are not counted as bad. They are simply left out, as though that minute was never recorded.

Two things follow from that, and both matter.

First, a gap does not create a second excursion.

If the voltage really had recovered in the middle, that would be a different story. A reading back inside the band always splits one excursion into two.

Second, the readings that were set aside leave the total. PTVRGB is then worked out from the minutes that were actually measured, rather than from a flat 10,080. The Signal column on the Stations screen is where you see how much of the week was missing, so check it before you produce the report. The PDF itself does not record this, which is why the Signal column is worth a look every time.

What happens if you switch it off:

With the switch off, zero is treated as a real voltage. Zero is far below any band, so all 2,258 dropped readings become excursions. Bolgatanga looks about nineteen points worse and its recorded time out of range nearly triples. Every one of those extra minutes is a communications failure, not a voltage failure, which is why the switch is on by default.

Drop readings the historian flagged **

This is switched off by default.

Next to every voltage column in the spreadsheet there is a small column that is usually blank but sometimes contains two asterisks. The historian puts it there. It is some kind of note about the quality of that reading.

The switch is off by default because the flag covers a lot of data. Almost a third of Techiman's readings carry it. Throwing away that much data based on a guess would do more harm than leaving it in.

Once GRIDCo tells you what the flag means, switch it on if those readings should be excluded. Nothing else about the tool needs to change.

The Report screen

Current week and previous week

Every figure is shown twice. The Current week column is the week you selected. The Previous column is the seven days immediately before it, so you can see whether a station is getting better or worse rather than judging one week in isolation.

If your spreadsheet does not reach back far enough to cover all seven days before your selected week, the previous column is left blank. A partial week compared against a full one would be misleading, so nothing is shown rather than something wrong.

The bar and the warning marker

The bar beside each PTVRGB figure is there to make the column quick to scan. A longer bar means more of the week was spent inside the band. The worst performing station is drawn in red and its figures are shown in bold, matching the band above the table.

A small amber warning triangle beside a station name means that station's signal was frozen for the whole week. Its figures are shown greyed out, because they describe a fixed placeholder rather than a real measurement, and the station is left out of the worst performer ranking.

Choosing the week

Pick any start date. The week runs seven days from whatever you choose. It is not fixed to Monday through Sunday. GRIDCo's own published week in the sample report runs from a Friday to the following Thursday, so the tool does not force a calendar week on you.

Some dates near the end of your data are greyed out and cannot be picked. That is because there are not seven full days of readings after them. The data is there, it just cannot start a complete week.

What goes on the PDF

The PDF is headed NITS Performance Report and follows the standard Section 9 layout, with the indices listed under Quality of Supply Indices:

• 9.1 gives PTVRGB for every station, as a percentage.
• 9.2 gives NVE for every station, as a count.
• 9.3 gives CDVE for every station, in minutes.
• 9.4 names the worst performing station.

Each of 9.1, 9.2 and 9.3 carries two columns, one for the current week and one for the previous week, and 9.4 names the worst performing station for both. Where the spreadsheet does not reach back far enough, the previous week cells are left blank.

The Remarks column is left empty for you to write in. Nothing is printed below the table.

Nothing is stored

Your spreadsheet is never written to disk. It is held in memory only for as long as it takes you to finish, and is discarded after 2 hours. There is no database, no report history and no accounts. If you come back to an old link and it has expired, upload the file again.