Physical data generation in Great Britain in May 2026: 37 million SMETS2 meters, substation SCADA, and the 87-type taxonomy across 13 categories

Where physical data generation stands in May 2026

Three changes in the last twelve months have moved the physical-data picture enough to warrant a short orientation before the mechanics. The first is the Market-wide Half Hourly Settlement migration. Migration began on 22 October 2025 under BSC Modification P408, reached the M10 to M13 cluster of programme milestones in the first quarter of 2026, and recorded around ten million Meter Point Administration Number initiations by mid-quarter; the programme cuts over at Milestone M16 in July 2027.^{9 10} After cutover, every electricity meter point in GB settles half-hourly rather than the Time of Use minority that settles half-hourly today. The data-volume implication is large: roughly 33 million Meter Point Administration Numbers each producing 48 half-hourly observations per day is around 1.6 billion observations per day or close to 500 billion observations per year flowing through the settlement chain.

The second is the smart-meter rules decision of 30 January 2026. Ofgem strengthened the operational Licence Condition on suppliers, introduced a ninety-day in-smart-mode fix obligation, and rolled out three Guaranteed Standards of Performance from 23 February 2026 (the six-week installation-appointment offer, the in-smart-mode fix duty, and the five-working-day smart-mode-fault assessment). At the end of 2025 the rollout had reached around 41 million smart and advanced meters installed, with more than 37 million operating in smart mode, around 71 percent of the GB meter base. Around 2.8 million domestic meters were installed in calendar 2025. About 11.3 percent of domestic smart meters operate in prepayment mode. The DCC carries the resulting message traffic across the Smart Meter Communications Licence pipe, with Schedule 16 of the Data (Use and Access) Act 2025 holding the smart-meter communications duty as the statutory parent.¹¹

The third is the Long Term Development Statement publishing pattern. Stage 2 publishes on 29 May 2026 under the third Ofgem derogation letter of 13 May 2026, which reshaped Stage 2 contents while holding the publication date; Stage 3 publishes on 30 November 2026.¹ The LTDS Stage 2 model is the asset-register substrate against which network telemetry is interpreted: a voltage reading at substation X resolves to a stable Master Resource Identifier in the CGMES 3.0 Equipment profile, so a downstream analyst can compare the May 2026 reading against the same substation in a later Stage 3 or Stage 2 publication. The voltage-cascade section below ties the LTDS Common Information Model labels back to the physical asset population that generates the data.

The Energy Smart Data and Privacy Framework sits over UK GDPR for the data the physical estate produces.¹² Purpose limitation by default, granular consent for finer-than-monthly resolution, and a Consumer Consent Solution operated by RECCo with a Minimum Viable Product launch in late 2026 set the consumer side of the picture. The physical-data picture below is the engineering side; the consumer side appears in the rules-and-governance sub-domain of the parent energy data layer page.

The four physical sources of bytes in the GB energy data layer

The physical-data picture resolves into four stacked sources rather than one undifferentiated stream. Each source has a distinct measurement device, a distinct sample rate, a distinct owner, a distinct published feed, and a distinct latency budget for downstream use. None is purely a software construct; every byte begins as a voltage, a current, a frequency or a phase angle sensed by a piece of equipment that a regulated party owns. Reading the four sources end-to-end is the only way to keep a particular data publication straight in the head: a 2025 fault count on a 33 kilovolt feeder lives in the DNO's SCADA-EMS, surfaces in the Embedded Capacity Register and the LTDS, and if it spilled into a settlement event, also appears as an imbalance volume in the Balancing Mechanism Reporting Service. Pulling one feed without knowing the other exists is the most common cause of misreading GB energy data.

Source one is the smart-meter fleet. SMETS2 electricity and gas meters at the premise, with the dual-band Communications Hub, the Home Area Network running Zigbee Smart Energy Profile, and the In-Home Display, the Pre-Payment Meter Interface Device and any Consumer Access Device on the same HAN. Around 37 million meters operating in smart mode in May 2026, each producing 48 half-hourly observations per day for electricity and a daily wake-up for gas, on a 13-month rolling local store before the DCC carries the data outward.⁹

Source two is the transmission and distribution SCADA estate. Around 1,500 transmission substations (operated by National Grid Electricity Transmission, SP Transmission and SSEN Transmission) and roughly 350,000 distribution substations across the 14 DNO licence areas, each instrumented to a degree determined by the voltage tier and the criticality of the asset. A grid supply point at 400 kilovolts has multiple Remote Terminal Units, dedicated Intelligent Electronic Devices and (at the more recent sites) full IEC 61850 instrumentation with Generic Object Oriented Substation Events messaging. A secondary substation at 11/0.4 kilovolts may have nothing more than fault-passage indicators. The aggregate output is voltage, current, real and reactive power, switch state, transformer tap position, temperature and fault waveform data flowing on a one-second to ten-second cadence into the SCADA-EMS at each operator.⁸

Source three is the Phasor Measurement Unit estate. Around 80 PMU installations across the GB transmission network, GPS-locked to the global timing signal, producing 50 phasor samples per second of voltage and current magnitude and phase angle, transmitted over IEEE C37.118.2 frame format on dedicated wide-area communications. The PMU layer is operated jointly by NESO and the three transmission owners under the wide-area monitoring service; it is the operational substrate for system stability monitoring and oscillation detection, surfacing into the Stability Pathfinder reports and the Electricity National Control Centre dashboards.

Source four is the published market and policy-data flow. The Elexon Balancing Mechanism Reporting Service publishes physical notifications, final physical notifications, bid-offer acceptances, system buy and sell prices, system frequency, system imbalance volumes and the half-hourly settlement runs at the operational cadence under the Balancing and Settlement Code. The Insights Solution endpoints replaced the legacy BMRS on 31 May 2024; the IRIS push API delivers the latency-sensitive feeds in AsyncAPI 2.6 over MQTT. NESO publishes the Carbon Intensity API, the demand outturn, the wind generation telemetry, the interconnector flows and the Future Energy Scenarios under the NESO Open Licence. DESNZ publishes the Digest of UK Energy Statistics, the Energy and Emissions Projections and the Sub-national Energy Consumption statistics on the policy cadence.² The Common Information Model substrate that holds the LTDS Stage 2 publication of 29 May 2026 is the data-model layer that ties source four back to sources one to three through stable Master Resource Identifiers.^{3 4}

The first three sources above the published-data tier are the operational layer. They feed real-time decisions: dispatch, constraint management, fault clearance, voltage control, balancing actions. The fourth source is the published layer; it feeds half-hourly reconciliation through Elexon and the central settlement systems, plus every market-facing analysis that runs against open data. The same physical network carries both layers but the data lives in different systems with different governance, different latency budgets, and different commercial frameworks. The boundary is where most analyst confusion sits in 2026, because the Energy Digitalisation Framework treats both layers as part of the same publishing system while the operational layer remains held inside each licensee under their own SCADA and EMS arrangements.

SMETS2 smart meters: 37 million sensors at 30-minute cadence

The SMETS2 cohort is the largest single source of bytes in the GB energy data layer by observation count. Around 41 million smart and advanced meters were installed in domestic and small non-domestic premises across Great Britain at the end of 2025, of which around 37 million were operating in smart mode (a meter is in smart mode when the DCC central system has it on a known communications path and is exchanging service requests with it; the remainder sit in dumb mode for reasons that range from DCC connectivity gaps to supplier configuration to a SMETS1 meter awaiting Mid-Operational Cohort migration). Around 33 million of the operating cohort are SMETS2 (the second-generation specification under the Smart Energy Code); around 11 million SMETS1 meters have migrated to DCC central operation under the Initial Operating Capability and Mid-Operational Cohort programmes since the Smart Energy Code Section H pre-enrolment regime opened the migration path in 2018.⁹

An electricity SMETS2 meter (Electricity Smart Metering Equipment, ESME) takes an active import reading at the close of each half-hour, time-stamped to the meter's local clock (synchronised through the DCC central system to within plus or minus one second of GB Mean Time). The reading is stored locally for 13 months in non-volatile memory, signed under the Smart Metering Key Infrastructure, and held until the next service request from the registered Supplier or authorised party reads it. The default service request from the supplier is a daily pull of the previous day's 48 half-hourly readings under the DCC User Interface Specification's "Read Active Import Register" command; a Time of Use tariff supplier pulls the readings closer to real time through a more frequent service request. The Market-wide Half Hourly Settlement programme generalises the Time of Use pull pattern to every meter point after cutover at Milestone M16 in July 2027.

A gas SMETS2 meter (Gas Smart Metering Equipment, GSME) operates differently because it runs on internal battery power (typical battery life around ten years). The GSME wakes daily, exchanges its reading and any commands with the Communications Hub, and returns to a low-power state. The half-hourly read cadence is therefore not the gas pattern; gas settlement remains on a daily-read cadence under the Uniform Network Code, and gas pricing rolls up to the daily allocation rather than the half-hourly allocation that electricity uses.

The In-Home Display, the Pre-Payment Meter Interface Device and any authorised Consumer Access Device sit on the same Home Area Network as the ESME and the GSME. The HAN runs Zigbee Smart Energy Profile at 2.4 gigahertz primary (for the IHD, the PPMID, the CAD) plus 868 megahertz secondary (for the gas meter, distant electricity meters in basements, the Home Compatible Auxiliary Load Control Switch and the Auxiliary Proportional Controller). The dual-band Communications Hub is the architectural reason GB smart meters can talk to a basement gas meter through brick; the primary 2.4 gigahertz HAN carries the higher-bandwidth in-room interfaces, the secondary 868 megahertz HAN extends reach through walls and floors. HAN coverage at premise has lifted from around 70 percent under single-band hardware to around 96.5 percent under the dual-band design.

Three operational specifics shape what a SMETS2 reading is and is not. First, the reading is signed under the SMKI on each per-message exchange between the meter and the DCC. The signature chains back to the Root Certificate Authority that the Trusted Service Provider operates under the Smart Energy Code. Certificate lifetime has been extended to fourteen years under the PKI-E Programme to align with the Root CA lifetime. Second, the reading is not encrypted to the DCC; the end-to-end encryption sits between the meter and the supplier head-end (or any other authorised party), and the DCC sits in the middle as a pipe by design. The Data Service Provider has the routing rules and the DCC Key Infrastructure signature responsibility but does not have read access to the payload. Third, the reading carries the meter's local clock timestamp at half-hour close, not the DCC's central clock; the BSC settlement run uses the meter timestamp for the half-hour allocation, with a tolerance for clock drift that the SMETS2 specification bounds.

The Schedule 16 of the Data (Use and Access) Act 2025 is the statutory parent for the smart-meter communications regime, holding the duty for the Secretary of State to grant smart-meter communications licences in line with the Energy Act 2008 section 91A.¹¹ The Smart Meter Communications Licence (SMCL) is the operational instrument; Smart DCC Ltd holds it; the licence expires in September 2027, with the successor licensee tender running through 2025 and the operational handover due in the 12 to 15 months that follow. The Smart Energy Code is the multilateral code that binds anyone who reads or writes data through the DCC pipe, including SMETS2 device specifications, the DCC User Interface Specification, the security regime, the operational performance regime, the threshold anomaly detection (TAD) provisions for prepayment top-up integrity, the consumer access provisions for CAD authorisation, and the dispute resolution machinery.

The Energy Smart Data and Privacy Framework is the privacy overlay for the readings the SMETS2 cohort produces.¹² The default consumer-data resolution for a domestic supplier is monthly; an opt-in lifts the resolution to daily; a finer opt-in lifts it to half-hourly. A microbusiness can opt out of any granularity finer than monthly. The MHHS programme cutover at Milestone M16 in July 2027 changes the default for settlement (every meter point flows half-hourly through the central settlement service) but it does not change the consumer-data resolution default; the consumer keeps the right to restrict supplier-side access to monthly even after the settlement engine processes the half-hourly observation. The LTDS Stage 2 publication of 29 May 2026 is the asset-register substrate that lets a downstream analyst tie a particular meter's MPAN to the secondary substation that feeds it and the primary substation upstream of that, under stable Master Resource Identifiers in the Common Information Model.¹

The DCC pipe, the Meter Point Administration Number, and the SMKI signature

The DCC is the licensed pipe between the 41 million meter cohort and every authorised user of the data. The architecture has four working parts. The Data Service Provider is the central message broker that authenticates the supplier, applies routing rules from the Smart Energy Code, signs the outbound message under the DCC Key Infrastructure, and hands the encrypted payload to the regional Communications Service Provider. Two CSPs split GB: Arqiva covers the North (Scotland and northern England, around 30 percent of premises) on a long-range mesh radio at 412 megahertz licensed spectrum; Virgin Media O2 covers Central and South (around 70 percent of premises) on cellular networks transitioning to 4G as the 2G and 3G sunsets progress through 2026 to 2033. The DCC coverage obligation is no less than 99.25 percent of GB premises. The Communications Hub at the premise translates from the WAN side to the HAN side, running Zigbee Smart Energy Profile on the HAN and cellular or 412 megahertz mesh on the WAN. The DCC User Interface Specification (DUIS) is the field protocol; the on-the-wire payload between the supplier and the meter is ASN.1 over the Great Britain Companion Specification (GBCS) running on top of DLMS/COSEM application messages.

A Meter Point Administration Number (MPAN) is the 21-digit identifier that uniquely identifies an electricity supply point in GB. Gas supply points use the Meter Point Reference Number (MPRN) on the same architectural pattern. Every reading the DCC carries resolves to an MPAN (for electricity) or an MPRN (for gas) through the Central Switching Service held by RECCo. The CSS holds the MPAN and MPRN master registry, the current registered Supplier of record, the customer-of-record reference, and the connection metadata needed to route the next service request to the right meter through the right CSP. The switching completion target under the REC is next working day; the CSS has processed around 37 million switches since go-live in July 2022 on the DCC reporting. A switching event under the REC triggers a SEC-governed re-pairing of the device population with the new Supplier; the previous Supplier's authorisations have to be revoked; the DCC has to update the routing tables so that the next service request from the new Supplier reaches the right meter. The Guaranteed Standards of Performance from 23 February 2026 guarantee the smoothness of that handoff to domestic consumers.

The Smart Metering Key Infrastructure is the end-to-end signature layer between the meter and the authorised party (typically the supplier, but also any authorised CAD operator under the consent layer). Every command the supplier sends and every response the meter returns is signed under the SMKI; the signature chains back to the Root Certificate Authority that the Trusted Service Provider operates. Certificate lifetime has been extended to fourteen years under the PKI-E Programme to align with the Root CA lifetime; the operational handover from the legacy infrastructure to the modernised arrangement was a 2025 workstream. The DCC Key Infrastructure (DCCKI) is the parallel signature layer between the DCC and the supplier on the supplier-facing interface; TLS plus organisation certificates wrap the supplier head-end's API connection to the DSP. The two crypto layers together hold the data trustworthy across the pipe.

The Smart Energy Code defines the routing rules. A SEC Modification is the mechanism by which any change to the metering data flows is approved. The SEC Modifications Panel holds regular working groups; SECCo (the Smart Energy Code Company) holds the code itself; SECAS (the SEC Administrator and Secretariat) administers the modification process. The Energy Act 2023 changed the parent regime for the SEC: Ofgem now selects and licenses code managers under the Code Manager Selection Regulations 2024, with SECCo invited to be the code-manager candidate. The standard licence conditions for the SEC code-manager role were consulted on in May and June 2025. The SEC binds the device population (SMETS2 ESME and GSME, plus IHD, PPMID and CAD on the HAN), the message protocol (GBCS over DLMS/COSEM), the security regime (SMKI, DCCKI, the National Cyber Security Centre's Commercial Product Assurance scheme for device security), the operational performance regime, the threshold anomaly detection provisions for prepayment top-up integrity, the consumer access provisions for CAD authorisation, and the dispute resolution machinery.

The data the DCC carries sits inside the Energy Smart Data and Privacy Framework for purpose limitation and consent.¹² The framework grades consumer consent: monthly granularity is the default for a domestic supplier; an opt-in lifts the resolution to daily; a further opt-in lifts it to half-hourly. The "Data Guide for Smart Meters" (Energy UK and Citizens Advice, 2013, updated since) is the consumer-facing Privacy Charter document. The Data (Use and Access) Act 2025 (Royal Assent 19 June 2025) widens the legal base for third-party access to meter data; Section 138 and the majority of Part 5's data-protection provisions came into force in early February 2026 under Commencement Orders 5 and 6.¹¹ The Consumer Consent Solution (CCS) that RECCo is delivering under the REC is the consent layer for third-party reads of the data the DCC carries; the Minimum Viable Product launches in late 2026, with the API technical specification draft scheduled for summer 2026 and early adopters in Q2 and Q3 2026.

A worked walk through the pipe is the easiest way to keep the picture straight. A supplier head-end originates a service request at its data centre (a daily pull of the previous day's 48 half-hourly readings, say). The DSP receives the request over the supplier-facing API, authenticates it under DCCKI, looks up the MPAN in the CSS registry to confirm the supplier of record, applies the SEC routing rules to identify the CSP region, signs the message under DCCKI, and hands the message to the regional CSP. The CSP carries the message over the WAN (Arqiva long-range mesh in the North, Virgin Media O2 cellular in the Central and South) to the Communications Hub at the premise. The Communications Hub translates from WAN to HAN and delivers the message over the Zigbee SEP HAN to the meter. The meter executes the request, returns the 48 half-hourly readings, signs the response under SMKI, and the response transits the same path in reverse. The supplier head-end ingests the readings, applies the active tariff, and feeds the half-hourly aggregated meter data to Elexon for the half-hourly settlement run under the Balancing and Settlement Code. The whole walk runs typically in under 20 seconds for a small batch, longer for a bulk request.

Network telemetry: SCADA, Remote Terminal Units, and Phasor Measurement Units

Network telemetry is the second-largest source of bytes in the GB energy data layer by volume per asset, although the aggregate observation count is far smaller than the smart-meter cohort because the asset population is far smaller. The transmission and distribution Supervisory Control and Data Acquisition systems hold a complete picture of every substation at every moment. The SCADA polling cadence is typically one second to ten seconds at transmission, two to ten seconds at distribution; the dataset per poll is voltage on each bus, current on each circuit, real and reactive power on each transformer and circuit, switch state on each breaker and disconnector, transformer tap position, transformer winding and oil temperature, and any fault-related event flags. The data flows into the SCADA-EMS at the licensee, where the operational engineers read it on screen and the EMS runs the state estimator and the contingency analysis against it.

A Remote Terminal Unit is the field device at the substation that polls the protective relays, the bus voltage measurements, the current transformers and the transformer auxiliaries, and transmits the data back to the SCADA control centre over a dedicated wide-area communications link. The protocol stack at transmission is typically IEC 60870-5-104 (a TCP/IP-based protocol that replaces the older serial 60870-5-101); at distribution it is more often DNP3 (a widely deployed North American protocol that has become the de facto distribution SCADA standard in GB). Sub-second peer-to-peer messaging between Intelligent Electronic Devices in IEC 61850-conformant substations runs under the IEC 61850 Generic Object Oriented Substation Events (GOOSE) profile; GOOSE messaging is used for protection-grade trip signals where the 60870-5-104 polling cadence is too slow.

The Phasor Measurement Unit (PMU) layer sits above SCADA. A PMU samples the voltage and the current waveform at 50 hertz (one sample per 20 milliseconds, matching the GB grid frequency, with a sub-cycle resolution achieved through interpolation), computes the phasor (the complex-valued representation of the waveform), and timestamps the result against the GPS time signal. The output is transmitted in the IEEE C37.118.2 frame format at 50 reports per second on dedicated wide-area communications. Around 80 PMU installations across the GB transmission network produce the wide-area monitoring data that NESO and the three TOs use for system stability monitoring, oscillation detection, and post-event forensic analysis. The aggregate output across the PMU estate is around 240 megabytes per second of raw phasor data when every PMU is reporting; downsampling and event-triggered retention reduce the storage footprint to a fraction of that for long-term analysis.

The telemetry data does not leave the licensee in normal operation. It is the operational substrate that NESO and the TO or DNO use to manage the system in real time, and it feeds into the Energy Balancing Service Adjustment Data that publishes after each settlement period through the Balancing Mechanism Reporting Service. NESO's Carbon Intensity API publishes a 30-minute aggregated generation-mix dataset derived in part from the transmission SCADA stream; the dataset is licensed under the NESO Open Licence (OGL v3.0-derived).⁸ The demand outturn at the national level updates every 5 minutes; the wind generation telemetry at the transmission level updates every minute. None of these public feeds publishes the underlying SCADA stream itself; they publish aggregations and derived statistics computed against the SCADA stream inside the operator.

The LTDS Stage 2 publication of 29 May 2026 is the asset-register against which the telemetry stream is interpreted. Each substation in the SCADA estate resolves to a stable Master Resource Identifier (mRID) in the CGMES 3.0 Equipment profile, with the asset attributes (rated voltage, rated current, transformer tap range, fault level) named in IEC 61968-13 for the distribution case and IEC 61970-301 for the transmission case.^{3 4} An analyst reading the SCADA voltage at a substation can resolve the asset attributes from the LTDS publication and the operating context from the SSEP or the DFES; the analyst's tooling needs the three pieces together to interpret the reading correctly. Without the LTDS substrate, the SCADA reading is a voltage at a node that the analyst cannot tie back to the asset population in any stable way. The CIM substrate is the data-model layer that makes the SCADA and PMU stream readable across licensees.

Dynamic line ratings, the operational answer to a constrained system

Dynamic Line Rating (DLR) is the operational technique that lets a transmission or distribution operator carry more power on an existing overhead line by computing the conductor's real-time thermal capacity from local weather data rather than from the static seasonal rating that the planning standard uses. A standard rating treats the conductor as fixed against a winter or summer ambient temperature, a fixed wind speed (typically 0.6 metres per second across the conductor at 90 degrees, the IEC 61597 reference condition), and a fixed solar radiation; the rating is the current the conductor can carry under those reference conditions without exceeding its thermal limit. A dynamic rating measures the actual wind speed (typically faster than 0.6 metres per second), the actual ambient temperature (often cooler than the winter reference), the actual solar radiation (often lower than the summer reference) and the actual conductor temperature directly, and computes the current the conductor can carry right now without exceeding its thermal limit. The dynamic rating is typically 10 to 30 percent above the static rating, occasionally more during cold and windy weather.

The data flow for DLR is a downstream extension of the SCADA telemetry. Weather sensors on the conductor or on the supporting tower (wind speed, wind direction, ambient temperature, solar radiation) plus conductor temperature sensors (typically distributed temperature sensing along the conductor, or a thermal model fed by the weather measurements) feed a DLR computation engine. The engine outputs the current rating as a real-time signal back into the SCADA-EMS, where the operator's contingency analysis uses it for the dispatch decision. The cadence is one minute to ten minutes; the data leaves the licensee aggregated into the constraint-management signals that publish through the Balancing Mechanism, not as raw weather telemetry. The Met Office numerical weather prediction tiles (in NetCDF and JSON formats) are the upstream weather dataset that NESO and the TOs use for short-term wind and solar forecast inputs into the dispatch decision; the DLR computation uses the same upstream weather data but at the line-segment level rather than the regional aggregate.⁸

The operational case for DLR has lifted with the Connections Reform Gate 2 outcomes of April 2026, which progressed 283 gigawatts of generation and storage and 99 gigawatts of demand to firm offers under Phase 1 to 2030 and Phase 2 to 2035. The new connection load on the existing network exceeds the static rating headroom in many areas of the network, and the planning solution (build new line) takes 5 to 10 years to complete, sometimes longer. DLR is the operational measure that lifts the existing line's effective capacity by 10 to 30 percent in the interim, at the cost of an additional telemetry layer and a more complex dispatch decision under the Grid Code and the Distribution Code operational planning provisions.⁶ Several TO and DNO sites have DLR in production by mid-2026; the data architecture that supports it is a downstream extension of the SCADA pattern rather than a separate data layer.

The voltage cascade with CIM labels, from 400 kilovolts to 230 volts

The physical-data picture cannot be read without holding the voltage cascade in the head at the same time. The cascade has five tiers from the 400 kilovolt and 275 kilovolt transmission super-grid through 132 kilovolts (the boundary tier owned by the DNOs in England and Wales but by SP Transmission and SSEN Transmission in Scotland), 33 kilovolts, 11 kilovolts and the 230-volt single-phase or 400-volt three-phase low-voltage service that arrives at the meter. Each tier has its own measurement device population, its own SCADA cadence, its own CIM data-model representation and its own publishing arrangement under the LTDS. The voltage cascade page at voltage covers the engineering side; the table below holds the data-side correspondence so a reader can resolve a voltage tier to the asset population that generates the data and to the CIM class that represents the asset in the LTDS Stage 2 publication of 29 May 2026.

The IEC base for the CIM data-model labels is IEC 61970-301 Edition 7.0 with Amendment 1:2022, the Common Information Model base for the energy management system application program interface.⁴ The distribution profile is IEC 61968-13 Edition 2.0 (BS EN IEC 61968-13:2021), the Common Distribution Power System Model profiles for distribution-network exchange.³ The CGMES 3.0 profiles (Equipment, Topology, State Variables, Steady State Hypothesis) that the LTDS Stage 2 publication uses are layered on the 61970-301 base. Each substation, each transformer, each line section and each connection point at every voltage tier resolves to a stable Master Resource Identifier (mRID) in the published model; the mRID is a UUID assigned at the first publication and retained across every subsequent publication that includes that element, so a downstream analyst can compare two successive LTDS publications and identify which elements have changed.¹

The reading order between the voltage cascade and the physical-data picture matters because the data resolution and the data ownership change at each tier. At 400 kilovolts the licensee is one of three TOs (National Grid Electricity Transmission, SP Transmission, SSEN Transmission); the SCADA cadence is one to two seconds; the PMU cadence on instrumented circuits is 50 samples per second; the data flows into the NESO Electricity National Control Centre and into the TO control room. At 132 kilovolts in England and Wales the licensee is one of the 14 DNOs; in Scotland it is SP Transmission or SSEN Transmission; the SCADA cadence is two to ten seconds. At 33 kilovolts and 11 kilovolts the licensee is the DNO everywhere; the SCADA cadence is two to ten seconds at the primary substation, slower (or absent) at the secondary substation. At the 230-volt service the licensee is the DNO to the meter, the meter is the smart-meter cohort with the supplier owning the data and the DCC carrying it. The data resolution falls by orders of magnitude from the PMU layer at the top to the daily-read gas meter at the bottom; the data-ownership pattern shifts from the operator at the top through the operator and the supplier at the middle to the supplier at the bottom.

The voltage cascade is the structural axis of the physical-data picture. A data flow at the operational tier (NESO managing the transmission system in real time, a DNO managing a 11 kilovolt feeder during a fault) reads top-down: PMU at 50 samples per second, transmission SCADA at one second, distribution SCADA at two to ten seconds, smart-meter readings at 30 minutes. A data flow at the planning tier (a connection assessment under the LTDS, a Distribution Future Energy Scenarios update) reads bottom-up: the LTDS Stage 2 publication holds the asset-register at each voltage tier, the DFES holds the projected demand and embedded generation at each substation, the SSEP holds the strategic spatial allocation that the DFES projects against. The voltage cascade page at voltage ties the engineering parameters to the statutory limits at each tier; this section ties the data-model labels to the asset population at each tier so the planner can hold both pictures together.^{5 6}

The 30-minute settlement data flow under BSC P408

The half-hour is the canonical settlement period in GB. Every electricity trade settles in half-hourly intervals under the Balancing and Settlement Code (BSC). The lead diagram above walks the half-hourly observation from a SMETS2 meter through the DCC and the MPAN registry to the settlement run and the supplier head-end; this section sets out the eight stages of the settlement flow in the prose detail that the diagram cannot hold and ties each stage to the BSC modification log that governs it.

Stage one is measurement. A SMETS2 ESME takes an active import reading at the close of each half-hour and stores it in non-volatile memory. The reading is signed under the SMKI and held against the next service request from the supplier or other authorised party. Stage two is local validation. The meter checks the reading against its own clock (synchronised through the SMETS2 communications hub) and produces a signed observation. Stage three is transport. The DCC carries the observation over the SMETS2 communications hub from the meter to the registered Supplier, the Distribution Network Operator that owns the local network where required, and any other authorised party (Time of Use tariff providers, in-home display routers, third-party services authorised under the Energy Smart Data scheme). The DCC pipe is described in detail above; the transport stage is the part that takes the longest in absolute clock time, because the service request from the supplier is typically batched (a daily pull of the previous day's 48 readings is the default, not a real-time push of each individual half-hour close).

Stage four is supplier ingestion. The supplier ingests the half-hourly read into its billing engine and applies the active tariff to compute the consumer-side cost of the half-hour. Where the consumer is on a fixed monthly direct debit, the supplier reconciles the billed amount against the actual half-hourly cost at the end of the billing period. Where the consumer is on a Time of Use tariff (an Economy 7, an Economy 10, a half-hourly EV tariff like the Octopus Agile pricing pattern), the supplier applies the time-stamped tariff rate to the half-hour and accumulates the cost. Stage five is settlement aggregation. The supplier submits the half-hourly aggregated meter data to Elexon under the BSC. Stage six is the settlement run. Elexon runs the half-hourly settlement under BSC Section S; the run computes the settlement-period imbalance volume for each balance responsible party (typically a supplier or a generator) and publishes the result. BSC P408 was the modification approved in 2021 that defined the post-MHHS settlement run; the algorithm runs over the half-hourly reads from the central settlement system rather than the profiled non-half-hourly reads that were the pre-MHHS norm.¹⁰

Stage seven is the reconciliation. The Initial Settlement Run (II, Initial Information) publishes one day after the settlement day, followed by R1 (Initial Settlement) at 16 working days, R2 (Reconciliation) at 75 working days, R3 (Reconciliation) at 145 working days, RF (Final) at around 14 months and DF (Dispute Final) at around 20 months after the settlement period. Each reconciliation run corrects for late-arriving meter data and disputed readings, with the final DF run being the binding settlement. Stage eight is the consumer bill. The supplier issues the bill to the consumer (monthly, quarterly or annual depending on the tariff) showing the energy charge (computed from the tariff and the metered consumption), the standing charge, network charges (passed through from the Distribution Use of System and Transmission Network Use of System charges published by the DNO and the TO), the Renewables Obligation pass-through, the Climate Change Levy on non-domestic supplies, and Value Added Tax.

Two reforms reshape this flow in the period the workspace covers. The first is MHHS itself. Migration began on 22 October 2025 at M10; the M10 to M13 cluster of milestones reached around 10 million MPAN initiations by mid-Q1 2026; the M14 readiness gate around October 2026 has approximately 80 percent of MPANs migrated, with the Insights Solution publishing aggregated MHHS data from that point; the M15 full-implementation date is in May 2027; the M16 cutover is in July 2027. After cutover, every electricity meter point settles half-hourly through the central settlement service. The data-volume change is large: the central settlement service processes around 1.6 billion observations per day at steady state rather than the few hundred million it processes today (around 33 million MPANs * 48 half-hourly observations per day, with the Time of Use minority that currently settles half-hourly accounting for the lower current load).⁹

The second reform is the Insights Solution endpoints that replaced the legacy Balancing Mechanism Reporting Service on 31 May 2024. The new endpoints publish OpenAPI 3.1 specifications and use JSON over HTTPS, which lowered the integration cost for downstream analytics tools. The IRIS push API is the latency-sensitive companion to the Insights Solution; it uses an AsyncAPI 2.6 specification over MQTT for the feeds where polling against REST is too slow (Bid-Offer Acceptances, final physical notifications, BMU stacks, curtailment events). The market-side flow of the half-hourly settlement run publishes through these endpoints; a downstream analyst pulls the half-hourly imbalance price, the system buy and sell prices, the physical notifications and the bid-offer acceptances from the Insights Solution at the operational tier and ties them back to the LTDS Stage 2 asset register at the planning tier.²

The MHHS data flow lifts the settlement chain off the legacy Data Collection and Aggregation regime that the pre-MHHS Non-Half-Hourly classes 1 to 4 used. Half-hourly readings flow from the meter through the HAN to the Communications Hub, over the WAN through the relevant CSP, through the DCC Data Service Provider, and out to the supplier or to a Smart Data Service appointed under the new MHHS arrangements. The supplier or Smart Data Service feeds the readings to the settlement engine through Elexon's data services. The new arrangements introduce three Smart Data Service roles: the Lead Supplier role, the Half Hourly Data Service role, and the Unmetered Half Hourly Data Service role. Each role is licensed under the BSC; each role's interface specification is published on the MHHS programme portal. The cross-link to the digital infrastructure page at digital infrastructure holds the MHHS migration timetable and the M14, M15, M16 milestones in the operational reading.

The 87 named data types across 13 Data Best Practice categories

The 87 named data types form the catalogue that converts a casual question about "the data on X" into a named row with an owner, a cadence, a format and a citation. Every row meets four tests: an obligation under primary or secondary legislation, an industry code or a Standard Licence Condition; a named owner that publishes or holds the data on behalf of GB; a regular rather than ad hoc cadence; and a machine-readable format. The Energy Data Taskforce reported in July 2019 that GB lacked a single registry of data assets, and the recommendations led through Data Best Practice (Ofgem, 2021), Modernising Energy Data Access and the Energy Digitalisation Framework to the Digital Spine Coordinator role appointed at NESO from 1 April 2025. The 87 types reconstruct that registry from the bottom up: every published dataset across the five-pillar architecture (DCC, SEC, REC and CSS, MHHS, NCSC security) plus the planning tier published by NESO and the DNOs.

The 87 types fall into 13 Data Best Practice categories. The category names are: meter readings (T01 to T11); telemetry (T12 to T24); balancing and ancillary services (O25 to O27, O34 to O43, O60, O61); settlement (O28 to O33, O58, O62); capacity (O36 to O38); switching (O44 to O48); embedded capacity and outages (O49 to O54); operational performance and modifications (O55 to O57); future energy scenarios and ETYS (P63 to P67); the Long Term Development Statement (P68 to P76); CIM and CGMES profiles (P77 to P80); investment and regulatory returns (P81, P82, P83, P86); and strategic planning (P84, P85, P87). The categories follow the Energy Digitalisation Framework's "Data Best Practice" tiering rather than the older three-domain split (telemetry, operational, planning), because the Data Best Practice categories track the publishing obligation rather than the cadence; an analyst reading the catalogue needs the publishing obligation to know who to ask for the data, not the cadence which is a derived parameter.

Three category counts are worth holding in the head as orientation. The meter and telemetry categories (T01 to T24) hold 24 types: 11 meter-reading types (SMETS2 electricity and gas, SMETS1 post-migration, derived feeds) plus 13 telemetry types (SCADA, PMU, IEC 61850 GOOSE, frequency, voltage, MVAR, Carbon Intensity, demand outturn, wind generation, BMU notifications, interconnector flows, storage state of charge, EV charge point status, weather, Carbon Intensity forecast). The balancing, settlement, capacity, switching, embedded capacity, operational and modification categories (O25 to O62) hold 38 types: the settlement runs from II to RF, the bid-offer pairs, the balancing actions, the capacity market T-4 and T-1 auctions, the ancillary services products (Dynamic Containment, Dynamic Moderation, Dynamic Regulation, STOR, Black Start), the REC switching events, the MPxN registry, the theft and damage reports, the Embedded Capacity Register, the outage Notice of Resource Restriction, the Customer Interruption and Customer Minutes Lost reporting, the DNO connection offers, the DCC service performance reports, the SEC and BSC modifications, the NIS incident notifications, and the post-MHHS settlement readings. The planning categories (P63 to P87) hold 25 types: the Future Energy Scenarios, the Distribution Future Energy Scenarios, the Electricity Ten Year Statement, the LTDS Tables 1 to 9, the CGMES profiles (EQ, TP, SV, SSH), the RIIO ED2 Regulatory Information Returns, the Network Options Assessment, the Cost Benefit Analysis cost stacks, the Strategic Spatial Energy Plan, the Holistic Network Design, the RIIO ET3 business plans, and the Connections Reform queue data.

The arithmetic is 11 + 13 + 16 + 8 + 3 + 5 + 6 + 3 + 5 + 9 + 4 + 4 + 3 = 90 named types if every row stands alone, with three boundary types double-counted across two adjacent categories (the Embedded Capacity Register sits between operational and planning depending on the use; the system imbalance prices sit between balancing and settlement; the post-MHHS settlement readings sit between meter readings and settlement). De-duplicating the boundary cases brings the total to the 87 named types in the catalogue. The catalogue page at data types holds the full table with the canonical IDs, the lead owner, the cadence, the dominant format and the source register entry on each row.

Format conventions consolidate around a small number of dominant encodings driven by tooling, standards-body inertia and procurement cycles. Telemetry is largely binary (DNP3 and IEC 60870-5-104 at SCADA, IEEE C37.118 at PMU, DLMS/COSEM through DUIS at the smart meter). Operational data is JSON-led (the Insights Solution endpoints publish JSON over REST with OpenAPI 3.1 specifications; the IRIS push API uses AsyncAPI 2.6 over MQTT; the BSC settlement files use CSV with the BSC P-format definitions; the REC switching messages use XML against the Schedule 8 schemas). Planning data consolidates around CIM and CGMES (RDF/XML for the network topology, with the LTDS Stage 2 publication of 29 May 2026 holding the first major GB instance of regulator-mandated CIM-based publication), Excel workbooks for the LTDS Form of Statement, and GeoJSON for the spatial overlays in DFES and SSEP.^{3 2}

The Data Best Practice Guidance is the obligation under which the catalogue publishes. Version 3.5 of June 2025 updated the terminology from "Licensee/s" to "Obligated Party/ies" so the guidance can extend beyond network licensees to code bodies via licence modification and consequential code changes. The defining principle is Principle 11: presumed open. Every Data Asset, its Metadata and the Software Scripts used to process it are presumed open by default; the data custodian must provide objective justification if openness is restricted.⁷ The Data Triage Playbook, refreshed in 2024 by the ENA's Data and Digitalisation Steering Group, operationalises Principle 11 with five categories of legitimate barrier to openness: privacy, security, negative consumer impact, commercial sensitivity, and legislation or regulation. The Energy Smart Data and Privacy Framework sits on top of the Data Best Practice Guidance at the consumer-data layer for purpose limitation and consent.¹²

The 87-type catalogue is the working inventory that converts the physical-data picture into a publishing picture. A reader who arrives wanting to know what data exists for a particular question (the half-hourly demand at a particular substation, the bid-offer pairs for a particular generator, the LTDS table holding the planned investment for a particular DNO area) consults the catalogue, identifies the named type, and reads back to the owner, the cadence and the format. The catalogue does not tell the reader what the data says; it tells the reader where the data lives, who publishes it, on what cadence and in what format, so the reader can go to the source. The cross-link to the data-types catalogue page at data types holds the row-level treatment. The cross-link to the lifecycle and settlement sub-domain of the parent energy data page holds the half-hourly observation walking through the data chain from the meter on the wall to the settlement run at Elexon to the line item on a consumer bill.