Home / Solutions / Energy Anomaly and Waste Identification

Energy Anomaly and Waste Identification

Energy anomaly identification requires unified metering boundaries, sampling periods and operating-condition labels first, then aggregation through an edge gateway, and platform-side baselines, rules and manual review to form anomaly records. The system can detect deviations and indicate investigation directions, but it cannot judge equipment faults from electrical parameters alone, nor commit to a specific energy saving effect without a reference period and operating-condition data.

Energy DataStatus Visualization

Solution overview

The solution targets parks and plants that already have meters or an energy platform but lack anomaly explanation. The focus is not adding more meters, but connecting sub-metered data, shift or line labels, threshold rules, trend changes and response records, so that an anomaly becomes a traceable energy management task rather than a single alarm.

Applicable industries

  • Industrial parks
  • Manufacturing plants

Primary user roles

  • Energy managers
  • Electrical O&M
  • Business management

Problems addressed

  • Real-time meters exist, but anomalies and their causes are not explained
  • Operating-condition labels such as equipment, shift and production line are incomplete, which makes real anomalies hard to separate from normal fluctuation
  • Once an anomaly is found, there are no records of review, assignment and closure

Recommended product combination

CW

CW Edge Computing Gateway

Aggregates field meter data and performs protocol conversion, buffering and upload to the platform

FEXLINK

FEXLINK IoT Platform

Builds sub-metering models, baselines, rules, anomaly lists, response records and reports

SFA

SFA Multi-Parameter Electrical Controller

Serves as an edge acquisition terminal on circuits that need several electrical parameters, residual current or temperature in one point; capabilities are confirmed by model

System architecture and data flow

Energy Anomaly and Waste Identification System TopologyFour-layer data flow across perception, edge control, platform and application for Energy Anomaly and Waste Identification.01Perception layer
Electrical parameter terminals are deployed on the main incomer, workshop, production line or key equipment…
02Edge / control layer
CW handles field device connection, point mapping, necessary buffering and upload;where the business require…
03Platform layer
FEXLINK builds the data model by organization, area, device, energy type and shift, and maintains baselines,…
04Application layer
Provides sub-metering comparison, anomaly trends and review lists for energy managers, circuit location and…

Perception layer

Electrical parameter terminals are deployed on the main incomer, workshop, production line or key equipment circuits; where power quality analysis is required, add the appropriate monitoring device and define the measurement object

Edge / control layer

CW handles field device connection, point mapping, necessary buffering and upload; where the business requires local alarms, threshold evaluation can be configured at the edge

Platform layer

FEXLINK builds the data model by organization, area, device, energy type and shift, and maintains baselines, anomaly rules and the response process

Application layer

Provides sub-metering comparison, anomaly trends and review lists for energy managers, circuit location and response records for O&M staff, and periodic summaries for management roles

Data flow

  1. Terminals upload voltage, current, power, energy and device status at the agreed interval
  2. CW maps devices to points, buffers data during network outages according to configuration, and backfills on recovery
  3. The platform calculates baselines, year-on-year and month-on-month values and anomaly deviations by metering boundary and operating-condition label
  4. Anomalies enter the pending review list, where O&M staff add the cause, response action and closure status
  5. The platform produces anomaly statistics, key circuit rankings and exportable audit records

Integration and interfaces

  • Field meters preferably use the project-confirmed RS485 or Modbus-RTU interface, with addresses, baud rate and register tables verified
  • Gateway upload can use project-confirmed MQTT or HTTP API; network and platform deployment are determined on site
  • To correlate output, shift or device status, MES, PLC or manually tagged data can be connected; the data owner and update cycle must be confirmed first
  • Power quality, harmonic and accuracy capabilities must be tied to a specific model and datasheet and cannot be described by series name alone

Applicable standards

  • Applicable project standards for electrical metering, consumption monitoring and energy management; the final edition is confirmed against project location, industry and project documents
  • Field communication uses the project-confirmed Modbus-RTU, Modbus-TCP, MQTT or HTTP API

Acceptance and delivery boundaries

Suggested acceptance dimensions

  • Verify metering boundaries, meter multipliers, point addresses and operating-condition labels, and sample-check the explanation of differences between the main meter and sub-metered totals
  • Run against the anomaly rules and reference period agreed by both parties, and check that anomaly records contain time, object, metric, deviation description and evidence data
  • Validate network-loss buffering, backfill, duplicate data handling and time consistency
  • Sample-check the complete record from detection through confirmation and assignment to closure, and verify that reports can be exported

Limitations and exclusions

  • With only the main circuit or a few sub-meters, anomalies of specific terminal devices or individual production machines cannot be judged accurately
  • Without a reference period, output or shift data, the platform can indicate deviations but cannot commit to a saving rate or attribution conclusion
  • An electrical parameter anomaly is not equivalent to a mechanical equipment fault; equipment diagnosis requires process, vibration, temperature or maintenance information
  • Analysis models, rule counts and prediction capability are confirmed against the actual FEXLINK version, configuration and acceptance method

Operational value

  • Turns scattered consumption data into an energy view comparable by area, device or shift
  • Helps O&M prioritize persistent deviations and correlated anomalies instead of reacting to single-point alarms
  • Retains consistent definitions and process records for before-and-after efficiency retrofit comparison

Selection guidance

  • Decide first whether the target is total deviation, sub-metering waste, equipment idling or power quality anomalies, then choose between new meters and reusing existing data
  • Where harmonic analysis is needed, select a power quality terminal with the corresponding capability; for sub-metered energy only, not every circuit needs a high-spec meter
  • On unstable networks, prefer a gateway with local buffering and backfill, and confirm buffer capacity and data retention policy
  • Without usable operating-condition labels, include label acquisition in the first phase; otherwise anomaly models are hard to interpret

Typical implementation patterns

  • Typical implementation pattern: a park selects several key workshops and equipment circuits to establish sub-metering and shift labels, aggregates the data through an edge gateway to the platform, and tests rule usability with persistent deviations, anomalous periods and manual review records. This is an anonymized application pattern and does not represent a specific customer project or verified savings.

Before retrofit → With the solution

Before retrofit

  • Real-time meters exist, but anomalies and their causes are not explained
  • Operating-condition labels such as equipment, shift and production line are incomplete, which makes real anomalies hard to separate from normal fluctuation
  • Once an anomaly is found, there are no records of review, assignment and closure

With the solution

  • Turns scattered consumption data into an energy view comparable by area, device or shift
  • Helps O&M prioritize persistent deviations and correlated anomalies instead of reacting to single-point alarms
  • Retains consistent definitions and process records for before-and-after efficiency retrofit comparison

How to evaluate the effect

01

Measurable signal

Electrical parameter terminals are deployed on the main incomer, workshop, production li…

02

Data source

ESA Smart Meter (40+ Electrical Parameters) · ESE Power Quality Monitor (Harmonics & THD)

03

Review method

Verify metering boundaries, meter multipliers, point addresses and operating-condition labels, and sample-check the explanation of differences between the main meter and sub-metered totals

04

Acceptance boundary

With only the main circuit or a few sub-meters, anomalies of specific terminal devices or individual production machines cannot be judged accurately

Frequently asked questions

Can energy anomalies be identified with only a main meter?

Total consumption trends and period comparisons are possible, but location capability is limited. Separating workshops, lines or equipment requires additional sub-metering and a matching boundary register.

How do energy anomalies differ from fault alarms?

An energy anomaly means the data deviates from a baseline or expectation; a fault alarm usually comes from device status or protection signals. The former needs operating-condition review and cannot replace equipment fault diagnosis.

Can the platform find all waste automatically?

No. It can detect deviations by rule, rank them and provide evidence, but production changes, equipment start/stop and manual operations still require review. Capability follows the actual algorithm and configuration.

Why are operating-condition labels needed?

The same consumption means different things during shutdown, trial production and full-load operation. Shift, output and equipment status labels help judge whether a deviation is reasonable and improve the credibility of the explanation.

Is data lost during a network outage?

Buffering and backfill depend on the gateway model and configuration. Confirm local storage, buffer duration, clock source and backfill strategy during selection, and validate them with an outage test.

Request a matched product combination and project review

Share the application, circuit conditions and existing system interfaces so engineers can confirm selection and implementation boundaries.

Contact FEXLINK