Power & Electricity

India's smart meter rollout (under RDSS) is one of the world's largest IoT deployments. Architecture: 1) Meter Hardware: Smart meters with bidirectional communication — RF mesh (for dense urban), cellular NB-IoT/4G (for rural/sparse). Each meter: measures kWh (import/export for solar), records 15-min interval data, stores 45 days locally, supports remote connect/disconnect, tamper detection (magnetic, bypass, reverse current). 2) Communication: Tiered architecture — meters talk to Data Concentrator Units (DCUs) via RF mesh (~500 meters per DCU). DCUs aggregate and forward to Head-End System (HES) via cellular/fiber. For cellular meters: direct to HES via NB-IoT. Protocol: DLMS/COSEM (international smart metering standard). 3) Head-End System: Receives data from all meters — 250M × 96 readings/day = 24 billion readings/day. Must handle: scheduled reads (every 15 min), on-demand reads (billing verification), events (tamper, outage, power quality). Architecture: distributed message queue (Kafka) → data validation → time-series DB (TimescaleDB cluster). Partitioned by DISCOM/zone. 4) Meter Data Management (MDM): Validates raw readings: gap filling (missing readings estimated), outlier detection (impossible consumption), theft detection (consumption pattern anomalies). Validated data → billing system. VEE process: Validation, Estimation, Editing — standard utility practice. 5) Scale Challenges: Storage: 24B readings/day × 365 days × 5 years retention = petabytes. Solution: hot (30 days in TimescaleDB), warm (1 year in columnar storage), cold (archive in S3/Glacier). Compute: billing run for 5 crore consumers in 4-hour window — parallel processing across partitioned data. Network: 250M concurrent IoT connections — carrier-grade infrastructure.

SCADA (Supervisory Control and Data Acquisition) is the nerve center of power grid operations. Architecture: 1) Field Devices: Remote Terminal Units (RTUs) at substations — interface with: circuit breakers (open/close commands), transformers (tap changer control), measurement devices (voltage, current, power flow). Modern: Intelligent Electronic Devices (IEDs) with IEC 61850 protocol. Data types: analog (voltage 220kV, current 500A, frequency 50.02Hz) and digital (breaker open/closed, alarm triggered). 2) Communication: Redundant paths — primary (fiber optic, 99.99% availability) + backup (microwave radio). Protocol: IEC 60870-5-104 (TCP/IP based) or DNP3. Latency requirement: < 2 seconds for monitoring, < 100ms for protection signals. 3) SCADA Server: Receives data from 1,000+ RTUs, 100,000+ data points, updated every 2-4 seconds. Functions: data acquisition, alarm processing (priority-based, flood suppression), event logging, trending. Redundancy: dual servers in hot-standby. If primary fails, standby takes over in < 2 seconds. 4) EMS Functions (on top of SCADA): State Estimation: compute the 'true' state of the grid from noisy measurements (weighted least squares). AGC (Automatic Generation Control): maintain frequency at 50Hz by adjusting generator output every 4 seconds. Economic Dispatch: minimize generation cost while meeting demand. Contingency Analysis: simulate 'what-if' — if Line X trips, will any other equipment overload? 5) Design Considerations: Availability: 99.99%+ (< 53 min downtime/year). Security: air-gapped networks (SCADA not connected to internet), defense-in-depth, IEC 62351 security. Determinism: real-time OS, guaranteed response times. Cybersecurity is critical — Stuxnet showed SCADA vulnerabilities.

India's electricity market has evolved from a state-monopoly model to a competitive market. Structure: 1) Players: Generators (NTPC, private IPPs, solar/wind developers). Transmission: PGCIL (central), state transmission utilities. Distribution: state DISCOMs (MSEDCL, BSES, etc.). Traders and exchanges: IEX, PXIL. Regulators: CERC (central), SERCs (state). 2) Market Segments: a) Long-term: Power Purchase Agreements (PPAs) — 10-25 year contracts between generator and DISCOM. 80% of India's power sold this way. Tariff set by CERC/SERC or discovered through competitive bidding. b) Medium-term: Bilateral contracts (1 month to 5 years). c) Short-term (IEX/PXIL): Day-Ahead Market (DAM): auction-based, trades for next day's 96 time-blocks. Real-Time Market (RTM): introduced 2020, trades every 15 minutes for immediate delivery. Green DAM: separate market for renewable energy. d) Ancillary services: frequency regulation, spinning reserves. 3) IEX Trading Process: Participants deposit margin money (bank guarantee). Bids submitted by 10 AM for next day. Market clearing: uniform price auction (all matched trades at single price per block). Congestion management: if transmission constrained, market splits into zones. Settlement: T+1 (next business day). Average price: ₹3-8/unit depending on time and season. 4) Price Dynamics: Peak (6 PM-10 PM): highest prices — ₹8-15/unit (cooling demand + lighting). Off-peak (2 AM-6 AM): lowest — ₹2-3/unit. Solar hours (10 AM-3 PM): increasingly cheap due to solar supply — 'duck curve' emerging. Monsoon: hydro abundant → lower prices. Summer: cooling demand → higher prices.

AT&C losses are the power sector's biggest challenge — India loses ~15-20% of electricity between generation and revenue collection. Breakdown: 1) Technical Losses (8-10%): Inherent in transmission and distribution — resistance losses in conductors, transformer core losses. Unavoidable but reducible. Fix: upgrade conductors (ACSR to HTLS), install capacitor banks (power factor correction), reduce transformation stages, use higher voltage distribution (11kV instead of LT). 2) Commercial Losses (10-15%): Theft and billing inefficiency. Types: a) Direct theft: hooking (illegal connection before meter), meter tampering (slowing/stopping meter). b) Billing errors: incorrect meter reading, wrong tariff category, unmetered supply. c) Collection losses: billed but not collected (defaults, political interference). 3) Smart Grid Solution: AMI deployment is the primary tool. Smart meters detect: magnetic tamper (magnet placed near meter to slow it), bypass (wire around meter), reverse current (meter runs backward). Remote disconnect for non-payment. Interval data reveals: consumption during meter-tamper events, unexplained consumption drops. 4) Analytics Approach: Compare: energy input to a feeder (measured at transformer) vs. sum of all consumer meters on that feeder. Difference = losses on that feeder. Drill down: which section? which consumer? ML models flag: consumers whose consumption pattern suddenly dropped, areas with high loss %, transformers with input-output mismatch. 5) Financial Impact: India's DISCOMs lose ₹90,000+ crore annually to AT&C losses. RDSS targets: reduce from ~20% to 12-15%. 1% loss reduction = ₹4,500 crore saved nationally. ROI on smart meters: investment recovered in 3-5 years through loss reduction.

Renewable energy integration is the power sector's most important technology challenge. India targets 500 GW renewable by 2030. Challenges: 1) Intermittency: Solar: available only 6-8 hours/day, varies with cloud cover. A cloud over a 500 MW solar farm can cause 200 MW drop in 5 minutes. Wind: varies seasonally (monsoon = high wind) and hourly. Grid must balance supply-demand every second — frequency must stay at 50Hz ±0.05Hz. 2) Duck Curve: As solar capacity increases, net demand (total demand minus solar) creates a 'duck curve': low during solar hours (10 AM-3 PM), rapid ramp up at sunset (3 PM-7 PM). California saw this first; India now experiencing it. Challenge: need flexible generation (gas turbines, battery, pumped hydro) that can ramp fast to cover the 'neck of the duck'. 3) Forecasting: Must forecast solar/wind generation 1 day ahead for market scheduling. Inputs: satellite imagery (cloud cover), weather models (wind speed, temperature), historical patterns. Accuracy: best models achieve 90-95% for day-ahead. Error has financial penalty (deviation charges). ML models: LSTM networks on historical + weather data. 4) Grid Stability: Solar/wind are inverter-based (no rotating mass) — don't provide inertia like thermal plants. Low inertia → frequency more sensitive to supply-demand imbalance. Solutions: synthetic inertia from inverters (grid-forming inverters), synchronous condensers, battery energy storage (BESS). India mandating grid-forming standards for new renewable plants. 5) Curtailment: Sometimes grid can't absorb all renewable generation (transmission congestion, low demand). Solar/wind curtailed (asked to reduce output) — wasted clean energy. Solution: better forecasting, storage, demand response (shift EV charging to solar hours), green hydrogen production during surplus.