How to use this book

Page status: selective teaching route and reference-library guide.

Begin with the parallel-member lesson. Make a prediction, work through the algebra, and run its short Python calculation. You can then follow the eight parts below, or enter at the first unfamiliar modelling task.

Follow the teaching route

PartStart hereWhat to work out
1. A plausible model gives the wrong answerParallel membersDerive a valid aggregate rating and its state boundary
2. From equipment to equationsSource to canonical modelCarry identities, terminals, orientations and units into equations
3. Conductors, connections, and groundLoad connectionsIdentify the voltage and return path each device uses
4. Graphs for different computationsOne network, many graphsSeparate physical topology from computational coupling
5. Transformations and recoveryPreservation contractsDeclare joint observations and recover eliminated quantities
6. Constraints and decisionsParallel AC decision caseCheck the source constraints after a transformation
7. Evidence for a computationNumerical consequencesDistinguish residual checks, conditioning, model adequacy and reproduction
8. An end-to-end modelling studyStudy workbookRecord and defend a reproducible study

The PDF selects material for these eight parts. Specialist derivations, exhaustive indexes, research logs, and language-model setup remain in HTML. Links to reference pages outside the PDF open the online library. The legacy PDF filename is retained so existing download links continue to work.

Work with the examples

Keep a short record of the source assumptions, predicted result, observed result, and explanation of any difference. Change one condition at a time. Use the computational case guide to choose a command and understand what its output establishes.

Power engineers may move quickly through familiar circuit laws while checking the conventions. CS and OR readers should use the notation reference for terminal and phasor conventions. The general multiconductor case is introduced in stages; it need not be the first model you implement.

Check the evidence

A definition states the convention used here. A derivation supports a result under its stated assumptions. An executable witness records a declared case and tolerance. An independent implementation may still share input data or model assembly. External review requires a recorded human reviewer and scope; the current claims ledger has no entries in that category.

Use the evidence map and knowledge-base index when a result needs closer inspection. A numerical value, a schema-valid certificate, and a physical validation answer different questions.