Tiza Group

Home / Guides / Studies and design

Functional Specification for Protection and Control

How to go from a general statement to a testable requirement: structure, an example with times and a matrix tying each requirement to its test.

By Shadai García, CEO and protection engineer at Tiza Group.

115 kV busbarTC52to the lineProtectionmain1test setsecondary injection21 = relay output contact (R-07, 30ms) 2 = breaker opening (R-08, 50 ms)requirement-test matrix (5 of 12 rows)R-07trip in 30 ms or lessfactory testR-08opening in 50 ms or lesssite testR-09lockout of closing after tripfactory testR-10trip in the event logsite testR-11alarm to the control centerno test11 of 12 requirements with assigned test: 91.7 %ChainRequirementCondition and valueAssigned testResult© tizagrp.com115 kV busbarTC52to the lineProtectionmain1test setsecondary injection21 = relay output contact (R-07, 30ms) 2 = breaker opening (R-08, 50 ms)requirement-test matrix (5 of 12 rows)R-07trip in 30 ms or lessfactory testR-08opening in 50 ms or lesssite testR-09lockout of closing after tripfactory testR-10trip in the event logsite testR-11alarm to the control centerno test11 of 12 requirements with assigned test: 91.7 %ChainRequirementCondition and valueAssigned testResult© tizagrp.com
Functional specification: where the time of each requirement is measured and matrix linking each requirement to its test. Location in the system and operating principle. Illustrative: real values come from the nameplate and the manufacturer's manual.
Short answer

The functional specification describes what the protection, automation and control system must do, without saying with which device. Each requirement is written with a condition, an action, a value and a way to check it, and is tied to a factory or site test in a traceability matrix. That way, before buying, it is already known how each function will be accepted.

Short answer

The functional specification of a protection, automation and control system describes what it must do, not with which device. Each requirement is written with four pieces of data: the condition (what happens), the action (what the system does), the value (how much, in what time) and the way to check it (which test, with what criterion). A matrix ties each requirement to a factory or site test, so that before buying you know how each function will be accepted. CIGRE technical brochure 926 (2024) proposes a domain-specific language for this, independent of technology and meant so that the result can later be implemented with IEC 61850 or other standards.

Where it fits

The protection and control philosophy sets the general rules: redundancy, maximum times, backup. The functional specification turns them into individual requirements. From it come the equipment specification, the test plan and the contract. If this step is missing, the supplier interprets the philosophy in the way that suits it and the differences show up in site tests, when correcting them already costs time and money.

Structure of a requirement

A useful requirement answers five questions:

  1. Identifier. A unique code (for example, R-07) that does not change even if the text is reordered.
  2. Condition. The exact situation: type of fault, location, previous state of the system.
  3. Action. What the system must do: trip, block, send a signal.
  4. Value and tolerance. The time, the current, the threshold. With its unit and where it is measured from.
  5. Verification. The test that checks it (factory or site), how the condition is applied and what result counts as a pass.

A general text becomes a testable one like this:

  • General statement: “The line protection must be fast”.
  • Requirement: “R-07. For a three-phase fault at 80% of the line length, the main protection issues the trip in 30 ms or less, measured at the output contact from the instant the fault is applied. Verification: secondary injection in a factory test; it passes if the measured time is 30 ms or less in all three repetitions”.

What types of requirement there are

  • Protection functions: what protects what, in what time and with what selectivity.
  • Control and interlocking: commands, blocking conditions, sequences.
  • Signalling and communications: which signals are sent, to whom and with what maximum delay.
  • Recording: which events are recorded, with what time stamp and for how long.
  • Behaviour when the system itself fails: what happens if a relay, a battery or a channel fails.

Each family is written separately so it can be reviewed with the person who knows it: the protection specialist, the control specialist, the communications specialist.

Example: from requirement to test

Example data (assumed): 115 kV line at 60 Hz. The clearing budget in the philosophy is 100 ms for the main protection. Relay: 25 ms to the trip order and 5 ms for the output relay. Breaker: 3 cycles. The specification has 12 requirements; 4 fault types are tested (single-phase, phase-to-phase, two-phase-to-ground and three-phase) at 3 locations (20%, 80% and 100% of the line) with 3 repetitions each, and each injection lasts 4 minutes between preparation, application and recording.

1. Breaker time. 3 cycles at 60 Hz are 3 / 60 = 50 ms.

2. Total clearing. 25 + 5 + 50 = 80 ms.

3. Margin. 100 − 80 = 20 ms. Requirement R-07 is met with margin.

4. Pass criterion of the test. The relay, up to its output contact, has 25 + 5 = 30 ms: that is what is measured in the factory. The breaker’s 50 ms are checked on site by measuring the opening time. That is why R-07 asks for 30 ms “at the output contact”, and another requirement, R-08, asks for the breaker to open in 50 ms or less. Separating the two avoids blaming the relay for a breaker delay.

5. Test cases. 4 fault types × 3 locations = 12 cases.

6. Injections. 12 cases × 3 repetitions = 36.

7. Duration. 36 × 4 min = 144 min, that is, 2.4 hours.

8. Coverage. Of the 12 requirements, 11 have an assigned test and 1 does not: 11 / 12 = 91.7%. The one that does not (R-11, on the alarm to the control centre) is marked in the matrix as pending assignment and is not taken as met.

The traceability matrix

It is a table with one row per requirement, with the columns: identifier, short text, device or function, factory test, site test and status. It serves three purposes:

  • Before buying: check that every requirement has a test and that every test answers a requirement.
  • During testing: record the result of each test against its requirement.
  • In an audit or a claim: show what was asked for, what was tested and what was found. Acceptance tests are described in commissioning of protection relays.

A matrix that is kept up to date also protects the project against late changes: when a requirement changes, the matrix shows at once which tests must be repeated and which equipment is affected, instead of finding that out during site tests.

What changes when the system uses IEC 61850

If the substation is built with communications according to IEC 61850, the signalling and interlocking requirements later become messages between devices, described in SCL files. The functional specification does not say which message to use: it says what information must arrive, to whom and in how much time. That separation lets you change supplier or technology without rewriting the requirements, and, when reviewing the trip matrix, it makes clear which requirement justifies each trip.

Common mistakes

  • Mixing in one sentence two requirements with different tests.
  • Writing values without saying where they are measured from (from the fault, from pickup, at the output contact).
  • Naming a product instead of describing the function.
  • Not assigning a test to communications and recording requirements.
  • Changing the text without keeping the identifier, which breaks traceability.
  • Leaving vague words such as “fast”, “adequate” or “reliable” where a number and a test are needed.

Checklist

  • Each requirement with a unique identifier.
  • Condition, action, value and tolerance written.
  • Measurement point of each time defined.
  • Factory or site test assigned to each requirement.
  • Requirements for failures of the system itself included.
  • Traceability matrix with no empty rows.
  • No product names in the functional requirements.
  • Review by protection, control and communications.

References

  • CIGRE, Technical Brochure 926 (2024), Methods for specification of functional requirements of protection, automation and control, working group B5.64.
  • Calculations and example data: our own.

Need this study?

Writing the functional specification for protection and control for your project, with the matrix that ties each requirement to its test, is part of our engineering design and studies service. We do it for projects in any country. Write to us.

Frequently asked questions

What is the difference between a functional requirement and an equipment requirement?

The functional one says what the system must do in a given situation: for example, trip the breaker within a certain time for a certain fault. The equipment one says what the device must have: inputs, outputs, supply voltage. They are written separately because the functional requirement does not change when the supplier changes, and the equipment one does.

How many requirements should a specification have?

As many as it takes so that no function goes uncovered, and none that cannot be checked. It is better to have 40 testable statements than 200 generic ones. A requirement counts when it has an identifier, a value and an assigned test.

Do you have to use a formal language to write it?

It is not mandatory. CIGRE technical brochure 926 proposes a domain-specific language for describing protection, automation and control functions, but a sentence template with condition, action, value and verification method delivers a good part of the benefit with ordinary tools.

WhatsApp