Electric Metering & AMI · Lesson 3 of 5

Designing a representative AMI pilot

A pilot is a controlled test of the future operating model. A demonstration proves that equipment can work under selected conditions. A representative pilot tests whether the proposed system, people and processes work across the conditions that could control rollout: dense and sparse areas, difficult terrain, varied service types, integration exceptions and field-response realities.

13minutes3learning objectives3check questions
LEARNING OBJECTIVES

After this lesson, you should be able to:

  1. Separate a demonstration from a representative pilot
  2. Define measurable success criteria across technical and operating layers
  3. Identify the evidence required before scale
EXPLANATION

What is physically and operationally happening?

A pilot is a controlled test of the future operating model. A demonstration proves that equipment can work under selected conditions. A representative pilot tests whether the proposed system, people and processes work across the conditions that could control rollout: dense and sparse areas, difficult terrain, varied service types, integration exceptions and field-response realities.

Pilot design begins with decisions, not device quantity. If the intended benefit is billing-read availability, the pilot must define the required window, denominator, validation point and treatment of inactive accounts. If the use case is outage visibility, it must define how events reach the outage workflow, how false positives are handled and who confirms restoration.

01

Represent

Select geography, service types, coverage difficulty and customer conditions that resemble rollout.

Coverage of risk
02

Define

Set exact metrics, denominators, time windows, owners and acceptance thresholds.

Testable criteria
03

Operate

Run normal reads, events, commands, integrations and exception workflows over sufficient time.

Repeatability
04

Decide

Classify pass, conditional pass, redesign or stop with traceable evidence.

Scale gate
TECHNICAL VISUAL · SYSTEM-SPECIFIC MODEL

Pilot evidence funnel

A credible pilot narrows uncertainty from field behavior through utility operations.

ENGINEERING VIEW · ILLUSTRATIVE

Pilot and readiness evidence profile

%
Observed chainReference trend

Illustrative percentage of the defined test population remaining inside the validated evidence chain. Values are educational, not Metro Power project-performance claims.

READING NOTEIllustrative only. Percentages are not Metro Power performance claims and must be replaced with project criteria.
OPTIONAL ENGINEERING DEPTHPilot engineeringStructure representative coverage, integration, operational and acceptance evidence before rollout.
DESIGN RELATIONSHIPEnd-to-end success = field exchange × transport × platform processing × downstream acceptance
  • Define the eligible population and time window
  • Measure the result at the utility-use boundary
  • Preserve identity, timestamp, status and acknowledgement
ENGINEERING CHECKS FOR THIS LESSON
  1. 01Trace a reading and a return command across every interface
  2. 02Segment results by topology, terrain and service type
  3. 03Freeze KPI denominators, exclusions and acceptance gates
FAILURE ANALYSIS
Observed signalPossible causeDiscriminating test

HES read exists but billing data is missing

Mapping, validation or export failure

Trace one interval and its acknowledgements

Aggregate KPI is strong but exceptions cluster

Topology or service-condition bias

Segment performance and compare cohorts

PROJECT EVIDENCE TO COLLECT
  • Time-aligned HES and MDMS extracts
  • Device-to-service identity map
  • Exception ageing and closure register
WORKED EXAMPLE

A pilot that appears successful

Two hundred meters in a town center achieve 98% read availability, but the rollout area includes mountains and island communities.

  1. List rollout conditions absent from the pilot
  2. Identify which architecture assumptions could fail there
  3. Add representative locations and integration cases
  4. Retest the criteria with documented exceptions
INTERPRETATION

The 98% result is valid only for the tested population and workflow. It is not sufficient evidence for the full rollout architecture.

COMMON FAILURE OR MISCONCEPTION
A high read percentage automatically means the pilot passed.

The percentage matters only when the population, time window, validation point, exclusions and operating response are defined and representative.

RETRIEVAL PRACTICE · 3 QUESTIONS

Check what you can explain without looking back.

Choose an answer and report your confidence. The confidence signal is stored only until you submit this page.

OBJECTIVE · Distinguish pilot quality

What makes a pilot representative?

How confident are you?
OBJECTIVE · Define a metric

Which read-availability definition is strongest?

How confident are you?
OBJECTIVE · Make a scale decision

A dense-area pilot omits difficult terrain. What is the responsible conclusion?

How confident are you?
Answer every question and confidence prompt.