Definition
A smart meter measures electrical energy, stores the data, and reports it. It reads voltage, current, power, and energy, and sends them to the cloud platform over power-line carrier, 4G, or WiFi. It is a core device for smart energy, the smart grid, microgrids, demand response, and peak shaving and valley filling. It supports remote meter reading, load forecasting, usage analysis, and AI scheduling.
Functions
The main functions are metering (active and reactive energy, power, and power factor), storage (historical profiles, demand, and event records), communication (reporting over power-line carrier, 4G, WiFi, or LoRa), and control (remote connect and disconnect, and prepayment). In an energy microgrid, smart meters work with distributed PV and energy storage to support peak shaving and valley filling and demand response. Readings are gathered on the cloud platform and analyzed with big data and AI.
Where it is used
In homes and commercial buildings it supports remote reading, tiered tariffs, and prepayment. In industry it monitors electricity for smart manufacturing and Industry 4.0, and it supports peak shaving and valley filling. In an energy microgrid it meters and schedules grid connection for the microgrid, distributed PV, and storage. In demand response it lets sites take part in the electricity market through demand response. An API to the cloud platform can tie the meter into an energy management system and a digital twin.
Architecture and selection
The system is the smart meter, a concentrator or gateway, and the cloud platform. The concentrator reads the meters under it by power-line carrier and sends the data up over 4G or fiber. Selection should cover metering accuracy and certification, the communication method and its coverage, compatibility with power-line carrier, fit with the cloud platform and API, and the standards of the power industry. For smart-energy projects, SimtoLink can supply the meters, the gateway, the cloud platform, and the API together.
Implementation
Deployment starts with coverage and how the meters will be networked, then the meter model and concentrator, then the link to the cloud platform and the energy management system, and the rules for reading and analysis. In an energy microgrid or a demand-response program, the meter has to work with distributed PV and storage. Also check the API into the existing reading system and the cloud platform, and whether load forecasting and AI analysis are available.
Limits of Metering and Communication
A smart meter is still first a metering instrument: it measures active energy, reactive energy, and demand according to its wiring, and stores those quantities over time. Remote reading, tariff control, and remote connect or disconnect are added functions. They cannot redefine it so that "online" means the metering is correct. Frozen energy is energy that has already occurred at a moment. A load profile is the process by period. Using the sum of the profile to change a frozen value breaks the traceability of the metering.
It applies when the transformer ratio, phase sequence, and pulse constant match the parameters inside the meter, and the clock error is inside the allowed range. If the ratio is set wrong, energy is wrong in proportion however stable the communication is. If the clock cannot hold through a power loss, the profile after power returns is attached to the wrong date. A prepaid trip must be recorded separately from metering: the trip is a control event, not energy becoming zero.
Backfill is the step most likely to spoil the timestamp. Data already frozen in the meter during a communication outage should keep the freeze time when uploaded. If the platform overwrites the freeze time with the time the message was received, energy from the day of the outage is recorded on the day communication returned, and both daily energy and demand shift. If a concentrator misses some meters on its poll and fills the gap with a neighbor's data, loss analysis counts the unread user as zero consumption.
A meter does not replace a power quality monitor. It usually cannot provide a complete waveform and sag process. It provides metering and some events. Deciding whether power quality harmed equipment needs a separate waveform record. A voltage compliance rate on the meter alone cannot separate a steady low voltage from one deep sag.
Tariff periods must agree with the clock inside the meter and with the contract. A clock a few minutes slow drops energy from the peak period into the adjacent shoulder period, and the bill structure changes. If remote time setting moves the clock backward, duplicate or blank freeze points can appear. The backfill program must recognize them, not simply overwrite. A prepaid balance should be deducted from frozen energy, not estimated from instantaneous power that is still changing. Otherwise the balance will not match settlement.
Connect and disconnect are control, not metering. If closing fails, consumption does not resume. A platform that looks only at a successful communication acknowledgement thinks the user already has power. Trip reasons should be recorded separately: arrears, overload, a local button, or a master-station command. With the reason blank, it is unknown whether to close on restoration. If a transformer is passed through in the wrong direction, energy accumulates in reverse, and a photovoltaic user and a consuming user are seen with their roles swapped.
Meter address, measurement point, and user in the archive must correspond one to one. If a new meter keeps the old address after replacement and is not cleared, historical energy jumps. If a concentrator is replaced and one meter is left unconfigured, that meter looks like a long outage on the master station. Loss calculation that treats the missing configuration as zero energy assigns the loss to other users. The completeness of meter reading should therefore be calculated against the list that should be read, not against how many messages arrived that day.
Energy on the day a tariff switches must be split before and after the switching moment. The whole day cannot be counted into the new tariff. If demand inside the meter is cleared at the start of the month, the clearing moment should leave an event. Otherwise maximum demand accumulates again from zero and looks like a sudden drop in load. Those events should be kept with the frozen values, so settlement matches the statistical period in the contract.