Telemetry-Driven Analysis
Ingest, validate, and normalize operational signals from PI historians, SCADA, and metering systems.
TowerWatch
Economic Supervisory Intelligence for Institutional Buildings
Operational intelligence that connects metering to decisions.
TowerWatch ingests real operational telemetry from building and campus energy systems, structures it into analysis-ready features, and evaluates tariff-aware operating strategies. It helps institutional operators understand performance, costs, and flexibility — grounded in measured data, not assumptions.
Current Focus
RMI Brewery, Winery, and Food Pilot Facility — a UC Davis institutional microgrid with behind-the-meter PV, complex process loads, and SMUD tariff exposure. The current study evaluates degradation-aware economic dispatch for a simulated second-life battery system using 6 years of high-resolution PI telemetry.
What TowerWatch Can Do
Each capability is implemented and tested in the current codebase.
Feature Store
Generate analysis-ready features: net load, gross load, PV generation, import/export, ramp rate, rolling peak demand.
Demand Charge & Bill Analysis
Full SMUD CI-TOD bill calculation with TOU period classification, seasonal rates, and demand charge tracking.
Dispatch Strategy Comparison
Compare baseline, peak shaving, TOU dispatch, and forecast-assisted strategies on the same load profile.
Degradation-Aware Battery Modeling
Project capacity retention over 20 years with throughput-based cycling and calendar aging.
Sizing & NPV Analysis
Sweep battery sizes, compute NPV with replacement costs, and identify optimal configurations.
RMI Pilot Findings
Base case: 200 kWh simulated second-life battery, SMUD CITS-1 proxy tariff, 52,000 hourly data points.
Peak Shaving $6,690/yr savings (1.6%). 11.5% peak demand reduction. Battery lasts 18 years at 4.7 EFC/yr.
TOU Dispatch Highest NPV ($109k at 50 kWh) but battery life of only 2 years at 614 EFC/yr. Economically attractive; operationally destructive.
Optimal Sizing 50 kWh batteries outperform 200–400 kWh on NPV across all strategies. Diminishing returns from larger capacity.
Key Tradeoff Demand charge reduction (peak shaving) is more durable than energy arbitrage (TOU) for second-life batteries in institutional settings.
Battery behavior is simulated. PV generation is reconstructed from net meter deltas. Tariff is a proxy schedule. See methodology for provenance details.
How TowerWatch Works
1. Telemetry Meters, environmental sensors, and building management systems via PI Web API.
2. Validation Signal classification, semantic mapping, coverage assessment, and quality checks.
3. Features Net load, gross load, PV generation, import/export, ramp rate, rolling peak demand.
4. Cost Analysis Interval-level TOU energy charges, billing-month demand charges, export credits.
5. Strategy Comparison Baseline replay, peak shaving, TOU dispatch, forecast-assisted — all on the same data.
6. Degradation & Economics Battery capacity fade, replacement timing, NPV, payback, sensitivity analysis.
TowerWatch + SolMod
SolMod: Deploy
SolMod supports deployment of retrofit and distributed energy interventions — solar arrays, storage systems, efficiency upgrades, and building envelope improvements.
TowerWatch: Verify
TowerWatch provides the measurement, verification, and operational intelligence layer — answering whether interventions perform as expected and where adjustments improve outcomes.
Together: Learn
Deployment data from SolMod feeds TowerWatch analytics. Performance insights from TowerWatch guide SolMod decisions. Continuous validation replaces one-time commissioning.
Design Philosophy
Clarity over novelty. Operators need answers, not features.
Operational relevance. Every metric ties to a cost, a decision, or a constraint.
Data provenance. Measured, derived, reconstructed, and simulated are always distinguished.
Facility-manager legibility. If the facilities team can't read it, it doesn't ship.
Next Phase
The current pipeline establishes the economic baseline. The next phase adds forecasting and operator guidance: load and PV forecasting models, tariff-aware dispatch optimization driven by real forecasts, and cost-outcome-linked operator recommendations with confidence bounds. Grounded in institutional decision support — not generic sustainability nudges.