Sensors, IoT & Telemetry · Engineering example

Electrical Load & Power Quality Intelligence

Use appropriately installed interval and event metering to separate energy use, billing demand, and power-quality events. Connect electrical observations to equipment operation so qualified personnel can investigate the right circuit and test a remedy safely.

The operating challenge

One electrical dashboard cannot explain every interruption.

A monthly bill shows total energy and perhaps maximum interval demand, while operators remember brief trips and quality interruptions. Those are different time scales. A motor start can create a short voltage event without setting a billing-demand peak. A demand peak can occur during steady simultaneous loads without any power-quality fault. Useful analysis must start with the question, metering location, sample resolution, and electrical topology.

Where was it measured?

A utility meter, service entrance, panel, and machine feeder see different parts of the electrical system. Without a one-line map, a waveform has little diagnostic context.

What time scale matters?

Billing intervals describe demand and consumption. Voltage sags, imbalance, or distortion may require event-capable meters and finer time resolution.

What changed operationally?

Equipment starts, process state, maintenance, and utility events must be aligned before suggesting that one caused another.

Why modernize: Give engineering and operations a shared, time-aligned record so they can target electrical investigation, avoid unnecessary equipment replacement, and verify any approved change.

Illustrative operating example

A packaging line trips when several motors start together.

This packaging-line situation is illustrative. It is not a claim about a completed customer installation, an electrical diagnosis, or measured energy savings.

Imagine a packaging line with intermittent drive faults during morning startup. The plant also sees a high monthly demand charge, so the two are assumed to share a cause. An electrician-approved temporary power-quality study shows a short voltage dip near the affected line during some starts, while utility interval data places the month's billing-demand maximum on a different afternoon. The plant now has two questions: why the drive trips during a short event, and why a separate sustained load combination raises billed demand. A single energy graph would have blurred them together.

Map the electrical path.Qualified personnel identify service, panel, feeder, affected drive, meter locations, and any utility-side information needed.
Capture synchronized evidence.Record event voltage and current at suitable locations, line starts, drive fault codes, production timing, and the utility's actual demand interval.
Separate the questions.Review the trip with event-level electrical evidence and the bill with interval demand. Do not infer one from the other.
Review candidate causes.An electrical engineer tests supply, conductor, protection, drive, and load hypotheses; operations checks whether starts can be sequenced without harming production.
Verify an approved change.After qualified work or a supervised scheduling trial, compare similar startup and demand conditions, reliability, throughput, and quality.

What the reviewer would see

Line P-2: two different electrical questions

Short event evidence and billing demand remain separate.

Trip event: Timestamp, meter point, voltage and current event, drive fault, and machine state.
Demand peak: Utility interval, metered kW, concurrent loads, tariff definition, and bill period.
Investigation: Electrical topology, qualified findings, alternative explanations, and uncertainty.
Decision: Approved remedy or operating trial with measured after-condition.

Other high-priority situations

Different patterns need different instruments.

Transformer or feeder loading drifts

Trend phase currents and operating context to see whether a growing load or imbalance merits engineering review. A dashboard does not establish conductor or equipment adequacy.

Useful report: peak by phase, duration, job context, meter validity, and qualified assessment.

Nonlinear loads affect sensitive equipment

Where a power-quality study is justified, measure appropriate waveform and harmonic indicators at defined points. Compare events with equipment symptoms rather than assigning fault from distortion alone.

Useful report: meter location, harmonic profile, event timing, and confirmed impact.

What a modernization would connect

Choose meter placement and resolution before choosing analytics.

Electrical data is valuable only when its source and measurement limits are known. The installation and any work inside energized equipment belong to qualified personnel under site safety procedures. A read-only data service should not become a path for remote switching or protection changes.

Required operating evidence

Single-line diagram, service, panel, feeder, meter point, and asset mapping
Voltage, current, kW, kWh, power factor, and phase information as needed
Event-capable waveform or quality record only for a defined question
Sampling interval, clock quality, meter accuracy, and data completeness
Utility tariff, demand interval, bill and interval-data definitions
Equipment state, start time, trip code, production and maintenance context

Meter placement and billing calculations are documented; a submeter is not assumed to reproduce the utility bill exactly.

Connected technical layers

Approved permanent or temporary meters installed by qualified staff
Protected read-only collection and time synchronization
Utility interval and billing data under the correct tariff
Equipment-event and production records with asset identity
Review workspace for engineering findings and approved actions

Protection relays, switchgear, and machine controls remain outside the analysis system's authority.

Qualified meteringEquipment contextSeparated event and interval viewsEngineering reviewVerified remedy

A meter outage or clock mismatch is a data-quality finding. Do not join events and propose a cause when the time relationship is unreliable.

Where automation and AI fit

Automate comparisons, not electrical decisions.

Deterministic automation

Calculate interval demand and energy under explicit definitions, align meter and equipment events, detect gaps, and flag conditions that cross engineering-approved review limits.

Analytical methods

Compare start sequences, line states, and electrical events to surface repeatable associations. A model may rank investigation candidates but cannot establish electrical causality alone.

Evidence-linked AI assistance

Summarize event windows, meter locations, tariff assumptions, and alternate explanations for an engineer. It must not recommend unreviewed switching or protection settings.

Human operating authority

Qualified electricians and engineers install and interpret meters, approve electrical work, and decide corrective measures. Operations approves any process-scheduling trial.

How we examine a proposed solution

AI advisors challenge the electrical story before it is told.

These Adaptiv Stratum AI advisory roles can support bounded research and design review. They are not licensed electrical designers or protection-system operators.

Argus

Checks meter identity, clock alignment, and raw-event lineage.

Hephaestus

Reviews motor and drive operating context and measurement placement.

Pythia

Tests whether pattern models improve on explicit event comparison.

Aegis

Challenges false causality, missing phases, and unsafe recommendations.

Clio

Defines comparable before-and-after operating evidence.

Themis

Checks electrical work authority and change approval boundaries.

What becomes visible

Keep event quality, demand, and energy on their own clocks.

The report should make the electrical point of measurement and time resolution impossible to miss. Engineers need event evidence; energy managers need interval and billing context; operations needs the connection to actual production.

Event investigation

Meter point, phase, timestamp, waveform or event summary
Equipment state, start, trip, and fault sequence
Clock quality and missing-data flags
Upstream and downstream observations when available
Qualified finding and approved corrective action

Load and cost review

kWh by period and submeter boundary
Interval kW peaks under the actual tariff definition
Concurrent process loads and production context
Power factor or quality trends where justified
Meter coverage and reconciliation to utility data
Reliability, throughput, and energy after an approved trial

Do not claim that shifting a load reduces a demand charge until the tariff, demand window, and operational effects have been checked.

Operating safeguards

Metering must not become a hidden control project.

Qualified review of installation and applicable electrical safety procedures
Documented meter points, accuracy, and time resolution
Read-only collection with segregated access
Independent protection and machine-control functions
Visible gaps, stale data, and clock uncertainty
Tariff-aware calculations rather than invented savings
Engineering review before any electrical modification
Production and quality checks for load-scheduling trials

A practical first phase

Start with one recurring trip and one metering boundary.

Separate the questions.Define whether the first objective is a short trip, a demand peak, or both, and what evidence each needs.
Review the one-line.Electrical staff select safe meter points and confirm utility interval and tariff data.
Collect a representative window.Capture events and operating states through several normal starts and fault instances with synchronized clocks.
Test a qualified hypothesis.Review alternate causes before an electrical intervention or supervised load-sequencing trial.
Verify the outcome.Compare repeat events, demand intervals, output, and quality under similar conditions before broadening the program.

Begin with a measurable question

Is the problem a short electrical event or a sustained load peak?

That distinction determines the instrument, the report, and the right people to involve.

Discuss an electrical question →

Evidence and further reading

Match the meter to the electrical question.

These DOE sources inform the method. The packaging-line example is illustrative, not an Adaptiv Stratum customer case study.

DOE: Operations and Maintenance Best Practices Guide describes metering levels and when circuit-level or end-use measurements support power-quality study.
DOE: Evaluating Your Utility Rate Options explains interval data, load shape, demand, and tariff-specific charges.
DOE: Energy Management Information System Capabilities describes using interval data to understand demand profiles and manage identified issues.