PROJECT PAGE

Pump House Water Monitoring & Leak Detection

A field-deployed Raspberry Pi system that monitors flow (reed switch pulses) and main line pressure (100 PSI sensor), uploads time-series data to a dashboard, and is already set up for two final hardware installs: a 5 PSI rainwater tank sensor (level → gallons) and a cutoff valve for a nightly leak test.

Live: flow + main pressure MCP3008 (SPI ADC) Dashboard: bryangrissom.com/dash To-Do: 5 PSI tank sensor To-Do: cutoff valve leak test Outbound-only telemetry

Enclosure / Build Video

The monitoring box is installed and operating. The only remaining work is physical installation of the 5 PSI tank sensor and the cutoff valve — the enclosure, ADC channels, and wiring plan are already prepared for both.

Direct link: youtube.com/shorts/yTRCEp5bhfQ

Operational posture: monitoring runs continuously. Control features remain disabled until hardware is installed and explicitly enabled.

Project Snapshot

This system exists to catch failures early: leaks, continuous water usage, and unstable pressure behavior. It produces clean, timestamped telemetry and makes it visible through a single dashboard.

Primary signal
Flow (reed switch)
Pulses are bucketed (typically 60 seconds) to compute gallons used and to flag continuous-flow conditions.
Primary signal
Main line pressure (100 PSI)
Pressure trends reveal switch/regulator behavior and help detect abnormal cycling or drops.
Final upgrades
Tank gallons + leak test
Install the 5 PSI tank sensor to estimate rainwater gallons, and install the cutoff valve to enable nightly pressure-hold leak testing.

Current Status

The system is deployed and running in monitoring mode. Flow and main pressure data are uploading to the dashboard. The only remaining hardware installs are the 5 PSI rainwater sensor and the cutoff valve — both already planned and staged in the enclosure.

How It Works

The Pi samples sensors on a fixed interval (commonly 60 seconds), converts raw readings into real units, then posts authenticated JSON to the website. MySQL stores the time series and the dashboard displays trends and flags anomalies.

Inputs

  • Reed switch flow: pulse counting → gallons per interval/day; used for continuous-flow detection.
  • 100 PSI pressure sensor: read via ADC → PSI for main line stability and cycle analysis.
  • 5 PSI tank sensor (after install): read via ADC → PSI → calibrated to tank level → gallons.

Conversions

  • Flow: pulses × calibration factor → gallons.
  • Pressure: ADC counts → volts → PSI (sensor transfer function).
  • Tank gallons: PSI → water column height → gallons (lookup table or formula).

Outputs

  • Dashboard charts + anomaly flags at bryangrissom.com/dash
  • Optional morning email summary (system heartbeat)
  • (After valve install) nightly leak test result logged as an event
Design rule: monitoring never depends on control. Control is an opt-in layer enabled only after installation + validation.

To-Do (Final Installs)

The enclosure and software architecture are already staged for these. After installation, enabling them is primarily configuration and software updates.

1
Install 5 PSI rainwater tank sensor
Wire to the reserved ADC channel, then calibrate PSI → level → gallons for a stable tank-volume estimate on the dashboard.
2
Install cutoff valve
Wire through the relay output, verify manual bypass/override, then enable the nightly leak test logic.
After
Enable in software
Turn on tank gallons reporting + leak-test schedule + alerting thresholds + dashboard widgets.

Nightly Leak Test (5-minute shutoff)

After the cutoff valve is installed, the system can run a controlled overnight leak check: close the valve for ~5 minutes and verify whether main line pressure holds.

Algorithm

  1. Record baseline main line pressure.
  2. Close valve for 5 minutes.
  3. Measure pressure drop over the window.
  4. If drop exceeds threshold: flag probable leak/downstream demand.
  5. Reopen valve and log the result.
Guardrails: armed mode + cooldowns + automatic reopen on error. Monitoring remains active regardless.

Morning Status Email

A daily email summary provides a heartbeat report without needing to log in.

Security Model

The field node is treated like a sensor appliance. It posts data outward; it does not accept inbound control. This keeps the attack surface small while still enabling reliable remote monitoring.

Dashboard: https://bryangrissom.com/dash