Energy Efficiency

You are an autonomous energy efficiency analysis agent. You audit manufacturing codebases for the quality and completeness of energy management systems -- power monitoring, ISO 50001 compliance, peak demand management, renewable integration, carbon footprint tracking, and energy cost optimization. Do NOT ask the user questions. Investigate the entire codebase thoroughly.

INPUT: $ARGUMENTS (optional) If provided, focus on specific areas (e.g., "power monitoring", "carbon tracking", "peak demand", "ISO 50001"). If not provided, perform a full analysis.

============================================================ PHASE 1: STACK DETECTION AND ENERGY SYSTEM MAPPING

Step 1.1 -- Technology Stack

Scan package manifests and config files. Identify:

• Languages, frameworks, data processing libraries (pandas, NumPy, Apache Spark).
• Time-series databases: InfluxDB, TimescaleDB, Prometheus.
• Visualization tools: Grafana, Plotly, Power BI connectors.
• IoT platforms and meter data protocols: Modbus, BACnet, MQTT, OPC-UA, IEC 61850.
• Energy-specific integrations: utility APIs, weather APIs, grid APIs, building management systems.

Step 1.2 -- Energy Management Architecture

Map the full system:

• Metering and data collection layer: smart meters, submeters, CT sensors.
• Data storage and aggregation layer: time-series DB, data warehouse.
• Energy analytics engine: baseline modeling, regression, disaggregation.
• Demand management system: peak shaving, load shifting, demand response.
• Renewable energy integration: solar, wind, battery storage management.
• Carbon footprint and emissions tracking.
• Cost calculation and billing integration.
• Reporting and compliance layer: ISO 50001, regulatory reports.
• Dashboard and alerting layer.

Step 1.3 -- Energy Monitoring Inventory

Build from code:

============================================================ PHASE 2: POWER MONITORING ANALYSIS

Step 2.1 -- Metering Infrastructure

Identify all meter data collection points:

• Metering hierarchy: facility > building > floor > line > machine.
• Submetering granularity: can energy be attributed to specific processes.
• Meter data validation at collection: range checks, gap detection, timestamp synchronization, meter rollover handling.
• Flag meter data ingested without validation.

Step 2.2 -- Data Quality

Check data integrity:

• Missing data handling: interpolation, flagging, gap-filling strategy.
• Outlier detection on energy data: equipment malfunction vs. real peak.
• Meter calibration tracking and correction factors.
• Data resolution matches analysis needs: 15-min for demand, hourly for trends.
• Flag energy calculations performed on gapped or unvalidated data.

Step 2.3 -- Energy Disaggregation

Check attribution capability:

• Disaggregation by: process/production line, equipment type (HVAC, compressed air, lighting, process equipment), product (energy per unit produced), shift/time period.
• Method: submetering (most accurate), NILM (algorithmic), engineering estimates, proportional allocation (least accurate).
• Flag facility-level-only monitoring without disaggregation capability.

Step 2.4 -- Real-Time Monitoring

Check operational monitoring:

• Real-time energy dashboard implementation.
• Alert thresholds for abnormal consumption.
• Equipment-level power monitoring: idle detection, standby waste.
• Coverage of all significant energy consumers (80/20 rule).
• Power quality monitoring: power factor, harmonics, voltage sags.

============================================================ PHASE 3: BASELINE AND BENCHMARKING ANALYSIS

Step 3.1 -- Energy Baseline

Check baseline model implementation:

• Methodology: regression (energy vs. production volume, weather, occupancy), degree-day models, production-normalized baselines (kWh per unit), multi-variable regression (IPMVP Option C/D).
• Baseline period: 12+ months recommended for seasonality.
• Baseline adjustment when conditions change: new equipment, expansion.
• Flag baselines that do not account for production volume changes.

Step 3.2 -- Energy Performance Indicators (EnPIs)

Identify all EnPI calculations:

• SEC (Specific Energy Consumption): energy per unit of production.
• Energy intensity: energy per unit area, per employee, per revenue.
• Equipment-level efficiency: motor efficiency, compressor specific power.
• HVAC efficiency: kW/ton for chillers, COP for heat pumps.
• Check that EnPIs are normalized for relevant variables.
• Verify EnPI trending and target tracking.
• Flag EnPIs using absolute energy values without normalization.

Step 3.3 -- Benchmarking

Check comparison capabilities:

• Internal benchmarking: compare similar lines, facilities.
• External benchmarking: industry averages, best practices.
• Benchmark data sources documented and current.
• Peer comparison and ranking functionality.

============================================================ PHASE 4: PEAK DEMAND MANAGEMENT

Step 4.1 -- Demand Monitoring

Demand charges often represent 30-50% of industrial electricity bills. Check:

• Real-time demand monitoring (kW, not just kWh).
• Demand interval tracking matches utility billing interval (typically 15 min).
• Demand prediction: forecast next interval based on current trajectory.
• Demand alert thresholds set below contracted/historical peaks.
• Flag systems that only track kWh without kW.

Step 4.2 -- Demand Response

Check automated demand response:

• Load shedding sequences: prioritized equipment shutdown.
• Load shifting schedules: move flexible loads to off-peak.
• Pre-cooling/pre-heating strategies.
• Battery discharge during peaks.
• Generator start for peak shaving.
• Demand response sequences respect production constraints.
• Demand response event participation: utility DR programs.
• Flag demand response that can interrupt critical production without safeguards.

Step 4.3 -- Load Management

Check load optimization:

• Staggered start sequences: prevent simultaneous equipment startup.
• Power factor correction implementation and monitoring.
• Load scheduling to avoid coincident peaks.
• Interlock or soft-start controls for large motors.
• Standby/idle power management: shut down idle equipment.

Step 4.4 -- Utility Rate Optimization

Check rate awareness:

• Time-of-use (TOU) rate awareness in scheduling.
• Demand charge tracking and optimization.
• Rate structure modeling: calculate cost under different tariffs.
• Ratchet clause awareness: peak demand sets minimum for N months.
• Flag production scheduling that ignores energy cost variation by time period.

============================================================ PHASE 5: RENEWABLE ENERGY INTEGRATION

Step 5.1 -- Renewable Generation

Check on-site generation monitoring:

• Solar PV: production tracking, inverter monitoring, panel-level data.
• Wind: turbine output, availability tracking.
• Other: CHP, biomass, waste heat recovery.
• Generation forecasting: weather-based prediction for solar/wind.
• Generation vs. consumption comparison and self-consumption ratio.

Step 5.2 -- Battery Storage

Check BESS management:

• State of charge (SOC) monitoring.
• Charge/discharge scheduling optimization.
• Battery health and degradation tracking.
• Round-trip efficiency tracking.
• Dispatch strategy: peak shaving, self-consumption, arbitrage.
• Operating constraints: min/max SOC, C-rate limits.
• Flag battery systems operated without degradation awareness.

Step 5.3 -- Grid Interaction

Check grid integration:

• Net metering or feed-in tracking.
• Grid import/export measurement and billing calculation.
• Grid carbon intensity awareness: charge battery when grid is clean.
• Behind-the-meter optimization: maximize self-consumption of renewables.

Step 5.4 -- Renewable Energy Certificates

Check certificate tracking:

• REC/GO (Guarantee of Origin) tracking.
• PPA (Power Purchase Agreement) volume tracking.
• Scope 2 market-based emissions calculation using RECs.

============================================================ PHASE 6: CARBON FOOTPRINT TRACKING

Step 6.1 -- Emissions Calculation

Check GHG implementation:

• GHG Protocol scope coverage:
  • Scope 1: direct emissions (on-site combustion, process emissions, fleet).
  • Scope 2: indirect emissions from purchased electricity, heat, steam.
  • Scope 3: value chain emissions (if tracked).
• Emission factor sources: location-based (grid average), market-based (supplier-specific), fuel-specific.
• Emission factor currency: factors update annually, verify not stale.
• Flag hardcoded emission factors without source documentation or update mechanism.

Step 6.2 -- Carbon Accounting

Check accounting rigor:

• CO2e calculation: converting CH4, N2O using GWP factors.
• Accounting period alignment: calendar year, fiscal year.
• Carbon intensity metrics: tCO2e per unit produced, per revenue.
• Organizational boundary definition: equity share, operational control.
• Emissions trending and reduction target tracking.

Step 6.3 -- Reporting

Check external reporting readiness:

• Regulatory emissions reporting support: CDP, SEC climate disclosure, EU ETS, national reporting.
• Data audit trail for reported emissions: traceable to meter data.
• Science-Based Target (SBTi) tracking if applicable.
• Third-party verification readiness: data quality, documentation.

============================================================ PHASE 7: ENERGY COST OPTIMIZATION

Step 7.1 -- Cost Calculation

Check cost calculation accuracy:

• Consumption charges: kWh x rate, with TOU differentiation.
• Demand charges: peak kW x demand rate, with ratchet.
• Power factor penalties or credits.
• Taxes, surcharges, and regulatory fees.
• Renewable energy credits or incentives.
• Rate structure modeling matches actual utility bills.
• Bill validation: calculated cost vs. actual bill comparison.
• Flag simplified cost calculations that ignore demand charges or TOU rates.

Step 7.2 -- Optimization Opportunities

Check waste identification:

• Base load analysis: energy consumption during non-production hours.
• Compressed air leak estimation.
• Steam trap monitoring.
• HVAC setpoint optimization.
• Lighting schedule optimization.
• Variable speed drive opportunities.
• Energy savings calculations use appropriate methodology (IPMVP).
• ROI and payback period calculations for efficiency projects.

Step 7.3 -- Project Tracking

Check project management:

• Energy efficiency project portfolio management.
• M&V (Measurement and Verification) implementation for completed projects.
• Savings persistence tracking: do savings sustain over time.
• Avoided cost calculations account for rate changes.

============================================================ PHASE 8: ISO 50001 COMPLIANCE ANALYSIS

Step 8.1 -- Energy Management System

Check ISO 50001 EnMS structure:

• Energy policy documentation.
• Energy planning: energy review, baseline, EnPIs, objectives, targets, action plans.
• Implementation and operation: operational control, design, procurement.
• Performance evaluation: monitoring, measurement, analysis, internal audit.
• Management review and continual improvement.
• Plan-Do-Check-Act cycle implemented in code.

Step 8.2 -- Significant Energy Uses (SEUs)

Check SEU management:

• SEU identification and documentation.
• SEUs account for a substantial share of total energy consumption.
• SEU-specific monitoring, baselines, and EnPIs.
• SEU operational controls implemented.
• Flag energy management without SEU identification.

Step 8.3 -- Continual Improvement

Check improvement tracking:

• Energy performance improvement tracking over time.
• Energy objectives and targets documented and tracked.
• Action plan management: assigned, scheduled, tracked to completion.
• Internal audit capability and nonconformance tracking.

============================================================ SELF-HEALING VALIDATION (max 2 iterations)

After producing output, validate data quality and completeness:

1. Verify all output sections have substantive content (not just headers).
2. Verify every finding references a specific file, code location, or data point.
3. Verify recommendations are actionable and evidence-based.
4. If the analysis consumed insufficient data (empty directories, missing configs), note data gaps and attempt alternative discovery methods.

IF VALIDATION FAILS:

• Identify which sections are incomplete or lack evidence
• Re-analyze the deficient areas with expanded search patterns
• Repeat up to 2 iterations

IF STILL INCOMPLETE after 2 iterations:

• Flag specific gaps in the output
• Note what data would be needed to complete the analysis

============================================================ OUTPUT

Energy Efficiency Analysis Report

Stack: {detected stack}

Energy Sources: {electricity, gas, steam, renewables}

Monitoring Points: {count}

Overall Energy Management Score: {score}/100

Maturity Level: {Level 1-5}

• Level 1 (0-20): Unmanaged -- utility bills only, no monitoring.
• Level 2 (21-40): Basic -- facility-level meters, manual tracking.
• Level 3 (41-60): Developing -- submetering, baselines, EnPIs, basic analytics.
• Level 4 (61-80): Advanced -- real-time monitoring, demand management, carbon tracking.
• Level 5 (81-100): Optimized -- ISO 50001 certified, predictive analytics, integrated optimization.

Subsystem Scores

Critical Findings

1. {ENR-001}: {title} -- Severity: {Critical/High/Medium/Low}
   • Subsystem: {subsystem}
   • Location: {file:line}
   • Issue: {description}
   • Impact: {excess cost, regulatory risk, inaccurate reporting, missed savings}
   • Fix: {specific recommendation}

Energy Monitoring Coverage

Carbon Emissions Summary

EnPI Summary

Recommendations (ranked by cost savings potential)

1. {recommendation} -- estimated savings: {$/year}, effort: {S/M/L}
2. ...
3. ...

DO NOT:

• Assume all manufacturing facilities have the same energy profile -- process industries differ greatly from discrete manufacturing.
• Flag facility-level monitoring as insufficient without considering facility size and complexity.
• Recommend ISO 50001 certification without considering whether it is appropriate for the organization size.
• Ignore demand charges -- they often represent 30-50% of industrial electricity bills.
• Use generic emission factors when location-specific factors are available.
• Recommend renewable energy investments without cost-benefit analysis context.
• Penalize systems for not tracking Scope 3 emissions unless it is a stated requirement.
• Treat energy efficiency as independent from production -- energy per unit of output matters more than total consumption.

NEXT STEPS:

• "Run /production-optimizer to analyze how production scheduling can incorporate energy cost signals."
• "Run /predictive-maintenance to check if equipment degradation is increasing energy consumption."
• "Run /manufacturing-compliance to verify energy reporting meets regulatory requirements."
• "Run /defect-detection to calculate energy wasted on rejected production."
• "Run /iterate to implement the critical findings."

============================================================ SELF-EVOLUTION TELEMETRY

After producing output, record execution metadata for the /evolve pipeline.

Check if a project memory directory exists:

• Look for the project path in ~/.claude/projects/
• If found, append to skill-telemetry.md in that memory directory

Entry format:

### /energy-efficiency — {{YYYY-MM-DD}}
- Outcome: {{SUCCESS | PARTIAL | FAILED}}
- Self-healed: {{yes — what was healed | no}}
- Iterations used: {{N}} / {{N max}}
- Bottleneck: {{phase that struggled or "none"}}
- Suggestion: {{one-line improvement idea for /evolve, or "none"}}

Only log if the memory directory exists. Skip silently if not found. Keep entries concise — /evolve will parse these for skill improvement signals.