Meter-to-bill evidence
Scheduled reads, validation and exception workflows can strengthen billing evidence.

RF Mesh AMI for Philippine electric cooperatives
Metro Power supplies the meters, RF Mesh communications, data concentrators, and system support needed to help a cooperative plan an AMI network around real field conditions.
RF MESH AMI · FLAGSHIP ARCHITECTURE
An RF-capable meter records energy at the service. The reading and its time must be usable before a network can add value.
AMI in one clear idea

Advanced metering infrastructure (AMI) connects meters, two-way communications, collection platforms, data management and utility workflows into one operating system.
The practical test: if the information does not improve billing, loss investigation, service or planning, the project has installed technology without creating enough operating value.
Choose a perspective to see how the same field-to-system architecture can improve utility work and the customer relationship. Benefits depend on the deployed functions, integrations and operating procedures.
SMART METERScheduled reads, validation and exception workflows can strengthen billing evidence.
Interval data and events can help teams focus investigation where evidence points.
Supported outage and device events can reach responsible teams sooner.
Authorized functions can reduce selected field visits when controls are in place.

Comcore TA37 single-phase smart meter, 4G model · Product photograph
The data behind a reading
This separate TA37 4G model lets you inspect the register, local reading port and communications module up close. The example below explains the types of records a head-end system can receive. The RF Mesh journey above uses a different, RF-capable meter.
Different meter, same data questionsActual records depend on model and configuration
| Record | What it holds | What it serves |
|---|---|---|
| Energy registers | kWh read for billing. Total active energy imported is the standard register 1-0:1.8.0. | Meter-to-bill evidence |
| Interval profile | Consumption and demand recorded at the configured interval. | The loss balance and planning |
| Events | Outage, voltage and device events the model supports. | Service awareness |
| Tamper conditions | Cover, magnetic field, neutral and reverse energy conditions the offered model supports. | Revenue-protection review, never a verdict |
| Time stamp | The meter's own clock on every read and event. | Aligning supply and consumption in time |
| Relay status | Remote disconnect where specified. Commands return through a controlled, auditable path. | Governed remote service |
Which of these a given meter provides, and how each is reported, is set by its datasheet and its configuration, so each is confirmed in the specification.
A smart meter is only the beginning. Scroll through the path from the customer connection to a utility decision, and the governed command that can travel back.
SCROLL TO ASSEMBLE THE ARCHITECTURE ↓
SMART METERMeasureThe smart meter records energy, demand, interval profiles, supported events and tamper conditions at the point of service.
THE NEXT HANDOFF MUST ALSO SUCCEEDRF Mesh is the primary field network. Nearby meters relay information toward a concentrator; PLC or cellular support areas where field evidence calls for another route.
THE NEXT HANDOFF MUST ALSO SUCCEEDWhere the architecture requires it, the data concentrator unit (DCU) manages local meter traffic and maintains a controlled utility uplink.
THE NEXT HANDOFF MUST ALSO SUCCEEDThe head-end system (HES) manages devices, schedules reads, handles exceptions and governs authorized communication with the field.
THE NEXT HANDOFF MUST ALSO SUCCEEDValidation, estimation and editing (VEE) rules help the meter data management system (MDMS) store and prepare trusted data for billing, analytics and other utility systems.
THE NEXT HANDOFF MUST ALSO SUCCEEDNamed teams use the information for billing, loss investigation, outage awareness, customer service and network planning.
VALUE REQUIRES A NAMED WORKFLOW OWNERUse this short film as orientation, then continue into the system architecture, technical lesson or readiness action.
Follow the complete AMI chain from field measurement to communications, HES, MDMS and utility action.
Choose a cooperative scenario. The ledger follows evidence from the field through platform review to an accountable operating response. It also shows the human ownership required at every step.
FIELD EVENTScheduled interval closesMeter evidence begins the workflowFrom system view to design detail

kWhEnergy registers read for billing
Interval profileDemand record, at the configured interval
EventsOutage recording; other events depend on the model
TamperConditions the offered model supports
01 / Field intelligence
Smart meters capture consumption, interval profiles, events and supported tamper conditions at the consumer connection.
Where it breaksThe meter form, relay or event configuration does not match the service, so the evidence never exists.
What changes inside the utility

Scheduled reads and exception handling support a more accountable meter-to-bill process.
Control pointDefine acceptance, fallback and ownership, not just read frequency.
Interval data, events and network context can help teams narrow where investigation should begin.
Control pointAMI supplies evidence; mapping, analytics and field action create the result.
Supported outage, voltage, tamper and device events can enter defined response workflows.
Control pointEvery event needs a recipient, response rule and closure record.
Where specified, permitted actions may be executed remotely through controlled authority.
Control pointCybersecurity, authorization, audit trails and safety rules are mandatory.
Metro Power's flagship AMI architecture
Meter to meter, hop by hop, to a data concentrator with its own backhaul.
+Benefit: Reaches furthest: coverage grows with every neighbor, whatever state the wiring is in, and no operator is involved.
−Trade off: Every hop adds delay and takes a share of the channel. A meter with no neighbor in range has nobody to relay for it.
Comcore publishes a multi hop coverage radius of 8 km for its LoRa mode and 3 km for its end to end multi hop mode.
The reading rides the low voltage service cable to a concentrator at the transformer.
+Benefit: No antenna, no spectrum, no operator. HPLC adds the rate for frequent interval data.
−Trade off: Bounded by the transformer: every transformer area needs its own concentrator, and noise, poor joints and long branches weaken the signal.
A dual mode meter carries a radio as well and falls back to it where the cable fails.
Each meter links straight to the operator network. No concentrator.
+Benefit: Every meter is individually reachable, wherever it stands, with headroom for frequent interval data and on demand reads.
−Trade off: A data subscription per meter for the life of the asset, and the risk of the operator retiring the network the modem uses.
Budget the subscription over fifteen years, not one.
Mesh needs neighbors and power line needs enough meters per transformer. Scattered meters point to cellular.
Power line depends on the cable. Radio mesh and cellular do not.
Frequent interval data and on demand reads on large customers need cellular or high rate power line.
Dual modeWhere neither path alone is dependable, a meter carrying both power line and radio falls back without a truck roll. Ask for read success by path, not a blended average.
Patchy mobile coverageMesh or power line with a cellular concentrator: only one cellular link per area has to work.
Coverage radii are Comcore's published figures for its own equipment. The coverage that matters is the coverage measured in your pilot area.
In a radio mesh (LoRa based or Wi-SUN, in the sub gigahertz bands) each meter can pass its neighbor's data along, so coverage grows by hops rather than by tower range, and no telecom operator is involved. It does not care what state the low voltage wiring is in. It does care about density: a meter with no neighbor in range has nobody to relay for it, and every hop adds delay and takes a share of the channel. A concentrator or collector with its own backhaul terminates the mesh.
WHERE IT FITSSuits dense areas where the lines are long, tapped, mixed or storm damaged.
Power line carrier (G3-PLC, or the higher rate HPLC variant) sends data along the low voltage conductors from the distribution transformer to the meter. No antenna, no spectrum, no operator. Its boundary is physical: the signal does not pass usefully through a transformer, so every transformer area needs its own concentrator or coupler, and the concentrator count follows the transformer count. It is sensitive to line noise, poor joints and long branch runs. The band plan is a specification item: G3-PLC has regional band plans, and the meter, the concentrator and any filters must all be on the same one.
WHERE IT FITSSuits dense areas where the low voltage network is in good condition.
A cellular meter (NB-IoT, LTE-M or 4G) reaches the head-end system directly through the mobile operator, so no concentrator is needed and each meter is individually reachable. That suits scattered customers, bulk and check metering points and anything that needs frequent interval data. It depends entirely on the operator: coverage where the meter sits, a recurring data charge per meter for the life of the asset, and the operator's technology roadmap. Operators elsewhere in the region have retired older 2G and 3G networks and stranded meters built on them, so the offered module's bands and the operator's commitments belong in the compliance matrix.
WHERE IT FITSSuits scattered customers and high value points. Budget the subscription over fifteen years, not one.
A dual mode meter carries both a power line and a radio module and falls back from one to the other when a link degrades. It costs more per meter and buys resilience in areas where neither path alone is dependable: a poblacion feeder with noisy lines and dense housing, for example. Whether it is worth it is a pilot question, answered by read success on the hard sections.
HOW TO JUDGE ITAsk for read success by path, not a blended average.
The data concentrator
A data concentrator sits at the distribution transformer or on a nearby pole. On its downlink it polls or receives from the meters under it over power line or radio; on its uplink it talks to the head-end system over cellular or Ethernet. In between it buffers and timestamps the data and translates the meter protocol into the head-end system's transport. Because it sits at the transformer it is the natural place to meter the transformer's output, which is what gives loss analysis its supply figure. Four specification lines decide whether it earns its place.
Design noteA meter with its own cellular modem needs no concentrator. A design can mix both under one head-end system.
| Field condition | Pushes toward | Because |
|---|---|---|
| Meters dense, low voltage lines in good condition | Power line carrier | The wires are the network; one concentrator per transformer |
| Meters dense, lines long, tapped, patched or noisy | Radio mesh | Coverage comes from neighbors, not from the wiring |
| Meters scattered, long rural spans, islands | Cellular per meter | No neighbors to relay and too few meters per transformer to justify a concentrator |
| Frequent interval data or on demand reads on large customers | Cellular or high rate power line | Bandwidth and latency headroom |
| Coastal or island barangay with patchy mobile coverage | Mesh or power line with a cellular concentrator | Only one cellular link per area has to work |
| Neither path alone dependable | Dual mode | Fallback without a truck roll |
Comcore publishes its own suitability guidance and coverage figures for each network family. Metro Power treats these as the principal's published figures; the coverage that matters is the coverage measured in your pilot area.
How loss localization works
The method behind AMI loss analysis is an energy balance drawn around each distribution transformer. Meter what the transformer delivered on its low voltage side over a period. For the same period, add up what every customer meter served by that transformer recorded. The difference is the energy that entered the area and was not billed: the area loss. Divide it by the supply and you have the area loss rate.
That remainder contains two kinds of loss. Technical loss is the energy spent heating conductors and connections, and the transformer's own losses if the metering point is on its high side. Non technical loss is everything else: bypassed meters, tampered meters, faulty meters, connections nobody recorded. A steady loss rate is mostly technical. A step change on one date points to a new theft or a meter that has failed. An area running far above similar neighboring areas is the one to send an inspector to first. AMI does not find the theft; it tells the inspector which transformer to start under.
In this example the metering point at the transformer records 100 kWh. The four customer meters it serves record 25, 25, 25 and 5 kWh, 80 kWh in total. The 20 kWh difference, 20 percent of supply, is the unaccounted energy the investigation starts from.
Technical lossEnergy spent heating conductors, joints and connections, and the transformer's own losses when the metering point is on its high side.
Non technical lossBypassed, tampered or faulty meters and connections nobody recorded.
A steady rate is mostly technical. A step change on one date points to a new theft or a meter that has failed. AMI does not find the theft; it tells the inspector which transformer to start under.
What must be true
The balance is only as good as the connectivity model. If a meter is recorded under the wrong transformer, its consumption is subtracted from the wrong area and both areas look wrong. This needs a maintained network model or GIS, reconciled during rollout and kept current after every reconnection.
Supply and consumption must cover the same window. Interval data makes this straightforward; register reads work only if they are synchronized. A two hour offset between the transformer read and the customer reads produces a loss figure that means nothing.
The supply side of the balance is usually measured by the data concentrator at the transformer or by a check meter. That point needs the same accuracy and tamper protection as any revenue meter, because an error there appears as loss in every customer under it.
Water utilities do exactly this with a district metered area: a bulk meter on the inlet, customer meters inside the boundary, and a balance over aligned time windows. The Academy water lesson on measurement and district balance applies almost line for line to electricity, including its warning that a difference is a screening signal, not a conclusion.
Read the water treatment of the same method: Academy, Measurement, intervals and district water balance ↗A switch or optical sensor sits under the terminal cover. Lifting the cover to reach the wiring writes a time stamped event.
A second sensor watches the main cover. Where the housing is ultrasonically welded, it cannot be opened and closed again without visible damage.
A Hall sensor catches a magnet held against the case, the trick that once saturated the current sensors of older designs. An anti magnetic design uses sensors an external field does not affect.

Cutting the neutral used to stop the register while the load ran on through an earth path. Neutral failure metering keeps measuring from the live side sensor alone, so the bypass yields no free energy, and the event is logged.
With a second current sensor on the neutral, the meter compares live and neutral current continuously. A mismatch means current is leaving by another path around the meter. This comparison is what a datasheet calls bypass detection.
Energy registering in the wrong direction, as when the incoming and outgoing wires are swapped, is logged as a reverse energy event.

A three-phase meter watches each phase voltage. A phase lost while load continues on the others is logged against that phase.
Also written to the event logClock changes, relay operations and failed communication attempts, each time stamped and able to raise an alarm.
Each entry is time stamped in the meter's own log and can be pushed as an alarm, over the meter's cellular modem or through a data concentrator.
The head-end system receives the event and hands it to analytics with the meter's interval history.
Did this customer's consumption step down when the event was logged, against their own history and similar customers?
Did the unbilled remainder under this transformer rise on the same date? The balance also shows what no meter reports: a connection made upstream of the meter.
Evidence agrees: a field inspection, a chain of custody for the event logs, then the utility's own disconnection and legal procedure.
Evidence does not: a fault, a wiring error or a real change in use. The record is closed with its reason.
Log entries and curves are illustrative, not records from a customer or a project.
Systems loss above the regulator's cap is not recoverable in rates, which is why loss localization is usually the first business case for AMI. Metro Power does not state a target loss figure for any program; the result depends on the network model, the analytics and the field work that follows the evidence.
Optional AMI capability
Two ways in, one meterIllustrative flow · sample token
Customer paysAt an office, an agent, a kiosk or on a phone. The vending system generates a 20-digit token for this one meter.
Path A · localWorks with no network at all
Receipt or phoneThe token reaches the customer5918 0264 7731
4402 6180
Keyed inOn the meter keypad or on the interface unit5918 0264 7731
4402 6180
Path B · remoteWorks only while the meter is reachable
Head-end system (HES)Receives the token from the vending system
AMI networkPushes it to the meter. No keypad entry.
The same meterValidates the token once only, and only on this meter, then adds the credit.


Specify both paths. The local path is the fallback for outages and dead spots; say which one is primary in each area.
Each setting is a program decision written from the regulator's prepaid service rules and the cooperative's own policy. The meter enforces it; it does not choose it.
When a customer buys credit, the vending system encrypts the amount with a key that belongs to that one meter and stamps it with a time reference. The result is a numeric token. The meter decrypts it, checks it, and adds the credit. The format, the encryption and the life cycle of these tokens are defined by the Standard Transfer Specification, the IEC 62055-41 family, which is what the STS bullet on a meter datasheet refers to.
Control pointA token is money in transit. Treat the vending system as a cash system.
Because the key is derived from the meter's own number, a token bought for one meter is meaningless to any other meter. Because the time reference is part of the encryption, the same token will not load twice on the meter it was sold for. A lost receipt is therefore a customer service problem, not a revenue leak. The same token mechanism also carries management instructions such as a key change, a tariff change or a maximum power limit, so the utility manages the meter with the same tool the customer uses.
Control pointAsk how the bidder proposes to issue, hold and rotate meter keys.
On the local path the customer receives the token on a receipt or a message and keys it into the meter keypad or into a separate interface unit. This works with no network at all, which is why prepayment existed long before AMI. On the remote path the vending system hands the token to the head-end system, which sends it over the AMI network, and the customer does nothing. The remote path only works while the meter is reachable, so a program that depends on it needs the local path as its fallback for outages and dead spots.
Control pointSpecify both paths and say which one is primary in each area.
A customer interface unit is a small display and keypad in the customer's premises that talks to one meter over a short radio, Bluetooth or power line link. It shows remaining credit, warns when credit is low, accepts token entry and shows whether the meter accepted it. It is what makes split metering workable: the measuring unit can be sealed on the pole or in a shared box, and the customer still has a complete view of the account.
Control pointEach unit is paired to its meter at installation. Plan for pairing, batteries and replacement.
A well specified prepaid meter warns in stages before it disconnects, with a visible and an audible signal. Emergency credit lets the customer draw a small negative balance, recovered from the next purchase, so a household is not cut off at night for want of a few pesos. Friendly hours suppress disconnection during defined periods such as night time, Sundays and holidays. In most regulated markets these are consumer protection requirements, not conveniences, so the cooperative's policy should be written from the regulator's prepaid service rules before the meter parameters are set.
Control pointConfirm the current regulatory instrument before the policy is finalized.
Prepayment and time-of-use tariffs can run together, but only if the meter holds the tariff tables, the daily time slots, any seasonal schedules and the holiday calendar, and keeps its clock accurate through outages. Credit then drains faster in peak hours and slower off peak. The vending system, the meter and the billing system must agree on the same tariff version, which is a change control question as much as a technical one.
Control pointWhere a demand cap is wanted, a maximum power limit token can restrict rather than disconnect.
Feature sets vary by model and firmware. The exact prepayment functions, token logs, alarm stages and tariff capacity of any offered meter are confirmed in the project compliance matrix against the principal's published specification.
Metro Power AMPERE Protocol
AMPERE structures evidence, decision gates and accountable next actions before a utility treats AMI as a procurement or scaling exercise.
Explore the complete AMPERE framework ↗Preliminary AMI evidence check
Choose the maturity of the current record for every area. The indication uses evidence quality and critical dependencies, not the number of checked categories. This is preparation guidance, not an AMPERE gate decision, project approval, product selection or certification.
No current information or record can be shown.
The likely owner or source is known, but the record is not controlled yet.
A current record exists, but its responsible source still needs verification.
The current record was checked with the responsible source.
Pilot evidence at each handoff

PILOT CHECK · METERProve the record exists. Confirm which registers, interval profiles and supported events the offered meter can provide, and whether its clock and configuration suit the intended use.
A pilot built to answer decisions
A credible AMI pilot tests representative field conditions and the complete operating chain. Its purpose is to reduce uncertainty before scale, not to create a small demonstration that avoids the difficult conditions.
Agree on the operating problem, stakeholders, baseline and questions the pilot must answer.
Select service types, geography, communications conditions and consumer profiles that reflect rollout risk.
Test meter, network, HES, MDMS, integrations, exceptions, workflows and support, not hardware alone.
Use agreed methods for data delivery, billing acceptance, events, commands, ownership and issue closure.
Scale, redesign, extend or stop based on evidence, unresolved risks and lifecycle implications.
Relevant AMI equipment

01Product photograph
RF Mesh capable AMI applications
02Pictorial drawing
RF and PLC meter aggregation

03Product photograph
ANSI utility applications

04Product photograph
Residential connected metering

05Product photograph
Conventional or connected
Evidence, clearly separated
What makes multi-brand possible
The references above run several makers' meters under one platform because those meters share a data language. DLMS/COSEM, published as the IEC 62056 series, defines both how a meter organizes what it holds (energy registers, load profiles, the clock, the disconnect switch, each identified by a standard OBIS code such as 1-0:1.8.0 for total active energy imported) and how that data is exchanged over any link, from the optical port to power line, radio or cellular. A head-end system that speaks DLMS/COSEM can read a register on any conformant meter without custom work for each maker.
For a cooperative this is a procurement decision, not a technical footnote. A terms of reference that requires DLMS/COSEM conformance, with the certificate scoped to the offered model and firmware, keeps the second and third meter contracts open to competition and lets a pilot run on one brand scale on another. A terms of reference that names only a brand's own protocol chooses that brand for the life of the system. Certificates are issued by the DLMS User Association against a tested subset of the standard; ask for the certificate, and check that the model number on it is the model number being offered.
Quantities are stated in pieces by Comcore and come from Comcore's company profile. They are principal references, not Metro Power project claims.
Questions cooperative leaders usually ask
Clear answers help management and engineering discuss the same project, not different versions of it.
No. Smart meters are one field component. AMI also requires communications, HES, MDMS, integrations, cybersecurity, operating procedures and responsible ownership.
Usually, but only when the pilot has defined technical and operating questions, measurable success criteria and a decision gate for what happens next.
Metro Power starts with RF Mesh as the primary field network. Nearby meters can relay information toward a data concentrator without depending on the condition of the low voltage conductors or requiring a mobile subscription in every meter. The communications study and pilot must still prove meter density, terrain, interference and hard-to-reach locations. PLC or cellular should support only the areas where that evidence shows another route is more responsible. The pilot area should include the hard sections, not only the poblacion.
AMI can improve the evidence used to locate and manage loss. The result depends on data quality, network mapping, analytics and utility processes that act on the information.
More than a meter with a keypad. You need a vending system that sells credit and generates tokens, a way to get each token to the customer (a receipt, a text message, a mobile wallet), a policy for emergency credit and for the hours when the meter must not disconnect, a process for the customer whose token is rejected, and someone who holds the meter keys. The meter can be a keypad meter on the wall or a split arrangement where the measuring unit is sealed out of reach and the customer holds a small interface unit. The Academy lesson on prepayment walks through each of these decisions.
Yes, if the terms of reference require the common meter data language. DLMS/COSEM, published as the IEC 62056 series, defines how a meter organizes its data and how a head-end system reads it, so one platform can manage meters from several makers. Ask each bidder for the conformance certificate that names the exact model and firmware. Without that requirement, the first meter contract also chooses every meter contract after it.
Start with the meter population, service configurations, pilot geography, current reading and billing process, integrations, operating problems and desired outcomes.
Your AMI journey starts with evidence
Metro Power can begin with executive education, an AMPERE readiness workshop, pilot planning or formal TOR review. We will not force the project into the wrong stage.
