# From one load schedule to the whole power graph

> A list of consumers goes in, a phase-balanced power graph comes out. Balancing phases by hand is work that people do worse than machines.

Source: https://naasson.com/en/blog/one-load-schedule-the-whole-power-graph/
Published: 2026-04-01
Language: en
Direction: Power by Naasson (https://naasson.com/en/power/)
Publisher: Naasson — https://naasson.com

---
The input is a load schedule: a list of consumers with their ratings and requirements. The output is a power graph balanced across phases, plus a design book with a package for the electrician.

Between input and output lies work that is usually done by hand in a spreadsheet, with errors.

## Why phase balancing is hard to do by hand

The task looks like arithmetic: spread the consumers across three phases so that the loads come out even. With twenty consumers it really is arithmetic.

Then the constraints arrive. Some consumers must be on the same phase. Some must be on different phases. Some need uninterruptible power and go to a separate group. Some have inrush currents that must not stack at one point.

Each constraint is simple on its own. Together they produce a problem where a person finds a solution, not a good one. And the person does not notice when that solution stops being a solution after three more luminaires are added to the list.

## The hard part is recalculation, not the first calculation

The first balance will get done by hand. The problem is the second one.

The load schedule changes all the time: equipment is added, a rating changes, a consumer moves to another room. In a spreadsheet this is a one-line edit, and nobody rebalances after it, because rebalancing means redoing the whole job.

Six months later the panel is built to a document that is no longer consistent. This is discovered at commissioning.

When the graph is built from the schedule, a change to the schedule means a new graph. Recalculation is a command, not a project.

## What else comes from the same schedule

Design loads. Cable cross-sections. Short-circuit and arc-flash currents. Earthing loops. Selectivity. Load flow.

All of them depend on the same input set. Calculating them separately in different files is a reliable way to make them disagree, because usually only one file in five gets updated.

## The design book and the electrician's package

Both are printed from the same source. This matters. The installer's document and the model the calculations ran on are not two things that need to be kept in sync. They are two views of one thing.

The electrician needs specific facts in the package: which cables, where they run, what cross-section, what label, in what order. The engineer needs the model. The manager needs a report. One source, different views, generated deterministically.

## Where the person stays

In the load schedule. It is written by someone who knows the site.

And in checking the result. The calculation is advisory: it produces the distribution and flags risks, but it does not confirm compliance and does not select protective devices. Choosing the circuit breaker, confirming the room category and signing the conclusion stay with a qualified engineer.

The split is simple. The machine does the volume work, where people make mistakes because of sheer quantity. The person does the work that needs accountability.