Operator Brief / Joule by Azlan Data
Joule is an advisory layer for tower operators. It polls each site's telemetry every five minutes and recommends the optimal mix of solar, grid, battery and diesel — adapting to live load, weather and tariff conditions. Typical reduction: 15–30% of energy OpEx, with zero hardware change to the underlying site.
01 — What it is
Joule is a decision-intelligence platform for tower energy, built by Azlan Data. It sits alongside existing tower infrastructure as an advisory agent — polling each site's RMS device every five minutes for 27 telemetry points, fetching weather forecasts hourly, and recommending the optimal power-source mix to the operator's existing NOC or SCADA system.
The agent does not actuate switches itself. It emits a signed webhook with a recommendation, a human-readable reason, and a confidence score. The NOC keeps the steering wheel; Joule reads the map.
02 — Use cases
Each use case runs against the same advisory architecture — telemetry in, decision logic, webhook out — and is independently configurable per site. Operators typically adopt the source-optimisation use case first, then layer anomaly detection and fleet intelligence as confidence in the platform builds.
The shape of the engagement is the same in all three.
01
A five-case priority cascade evaluates solar, battery, grid and diesel every five minutes against live demand, tariffs and battery state-of-charge. Anti-flap protection enforces minimum hold times (15 min) and diesel cool-down. Savings thresholds and hysteresis bands prevent micro-switching on partly cloudy days. Every recommendation is deterministic, audit-trailed, and accompanied by a human-readable reason.
02
Joule learns a per-site, per-hour behavioural baseline over 30 days. Readings that deviate by more than three standard deviations inside a single five-minute poll cycle are flagged with a probable cause: unauthorised power draw, equipment theft, breaker trip, panel damage, inverter failure, battery cell degradation, or grid tampering. No manual threshold setting required.
03
A single agent instance manages 50+ sites with shared InfluxDB storage and a geographic weather cache (≈11 km radius). At a 10,000-site portfolio, the architecture reduces weather-API fetches to roughly 300 per hour. Nineteen event types cover operational info, site-level alerts and cross-site patterns — regional diesel shortage, grid outage cluster, fleet battery aging, solar underperformance.
03 — What changes
The numbers below are the typical operating envelope reported by Joule deployments. Specific results vary with site mix (solar-heavy versus grid-dependent), tariff structure, and the maturity of the operator's existing energy-management practice.
Reduction in tower energy OpEx
Solar utilisation against available irradiance
Diesel runtime as share of total operating hours
Grid consumption at peak tariff bands
Agent uptime under live operating conditions
Time between decisions, per site
Worked at portfolio scale: a 20% reduction against energy OpEx — energy being 30–40% of total tower operating expenditure — yields a material, recurring contribution to EBITDA from day one of full rollout. The size of that contribution scales linearly with portfolio size and site cost base; the operator's CFO is best placed to size it for their own footprint.
04 — How it fits
Joule is an advisory layer. It assumes the existence of an RMS device with a REST API, a working NOC or SCADA system, and a tariff schedule per site. The combinations below cover the situations most operators present.
Static rules — switch at 06:00, switch back at 18:00 — are blind to today's irradiance, today's tariff window, today's battery state, and today's load curve. Joule replaces the timer logic with a per-site, per-five-minute decision. Existing rules can be retained as safety guardrails; Joule's recommendation is constrained to honour them.
Where a NOC operator manually decides power-source switches across a portfolio, Joule provides a recommendation per site per cycle, with reasoning. The operator may auto-approve, batch-approve, or selectively intervene. Decision throughput moves from operator-bound to platform-bound; the operator's attention shifts to exceptions and trend management.
Joule does not replace SCADA. It consumes SCADA telemetry where available, emits recommendations to the same NOC, and adds a decision layer above the existing platform. Operators preserve their installed infrastructure, vendor relationships, and the audit history embedded in the SCADA system of record.
Joule needs no new hardware at the site. Where a retrofit is being considered (additional storage, larger solar, battery chemistry change), Joule can be deployed first as the diagnostic: the savings achievable from intelligence alone, before any capital is committed to new physical assets.
05 — How it works
Joule is structured as four layers — ingestion, intelligence, output, and platform. Each layer is independently versioned and the boundaries are deliberately thin, so operators can audit any decision back to its inputs.
Joule polls the site's RMS device over REST every five minutes for 27 telemetry points across solar, grid, diesel, battery and load. A weather API is fetched hourly, with results cached geographically — neighbouring sites within ≈11 km share a single fetch. At a 10,000-site portfolio, total weather-API load is approximately 300 calls per hour.
A priority cascade evaluates the five power-source combinations in order: solar only, solar plus battery, solar plus grid, grid only, diesel. ML forecasting layers on top — a 1–24 hour solar generation outlook combining weather API data with the site's historical efficiency curve; a 30-day load-history model accounting for hour-of-day and seasonal variance; a RAG-enriched battery dispatch schedule using thermal, stress and grid-reliability context.
Recommendations are emitted as signed webhooks (HMAC-SHA256) to the operator's NOC or SCADA endpoint. Each recommendation carries the recommended source mix, a human-readable reason, a confidence score, the contributing inputs, and the dollar impact of the decision against the do-nothing baseline. Anti-flap protection enforces minimum hold times (15 min) and diesel cool-down periods.
A single agent instance manages 50+ sites concurrently with independent poll loops, sharing InfluxDB storage and connection pools. A 14-endpoint REST API surface exposes fleet summary, site management, hot-reload configuration, and historical query. Sites can be added or removed in seconds without restarting the agent. Nineteen event types cover operational info, site-level alerts and fleet-wide patterns.
Every recommendation is deterministic and audit-trailed. The system is not a black box: the inputs, the rules they were evaluated against, and the resulting decision are all retrievable for any cycle, any site.
06 — Pilot path
Joule deployments follow a single shape — a 90-day pilot across a representative subset of sites, with baseline measurement, ROI validation, and a go/no-go decision before portfolio rollout. The framing below is the Azlan reference timeline; Mayfair21's commercial-representation framework wraps the engagement and underwrites the outcome.
Select 5–10 representative sites — urban and rural, solar-heavy and grid-dependent. Establish baseline energy cost per site over the preceding 30 days. Connect the RMS API, the weather feed, and the operator's NOC webhook endpoint. ROI dashboard active by week 2.
The ML layer learns each site's solar curve, load profile, and tariff response. Daily reports correlate weather patterns with energy anomalies. Anti-flap parameters and savings thresholds are tuned per site against early operational data.
Daily reports quantify savings against the week-0 baseline. AC optimisation comes online. The operator's NOC integrates Joule recommendations into its incident workflow. Anomaly alerts begin firing against the trained baseline. Fuel runtime, peak grid usage and solar capture are measured weekly.
Go / no-go against the agreed savings target. If go: full portfolio rollout strategy and financial projection, with per-site configuration migrated from the pilot subset. If no-go: the pilot fee is structured against the result, in line with Mayfair21's 3:1 guarantee on the analysis phase.
07 — About
Joule is built by Azlan Data, an engineering firm specialising in real-time energy optimisation for asset-heavy industries. The platform has been developed against live tower-operator data and is deployed under operating conditions, not as a research prototype.
Mayfair21 represents Joule under its commercial-representation framework. Engagements are opened at executive level at the operator, conducted on a paid analysis basis, and underwritten by Mayfair21's commitment to deliver at least three times the analysis fee in measurable return — or the analysis is free.
Procurement still runs procurement. The engagement enters that process with executive sponsorship already attached, the technical groundwork visible, and the commercial structure already agreed. It does not remove the competition. It does mean the conversation starts at a different altitude.
08 — Contact
All enquiries on Joule, the pilot programme, or the broader commercial-representation framework are treated with discretion.