Smart Flow Sensor
The work
Rameez led the design across mechanical, electrical and firmware, starting with the trade-off between sensing technologies and then working through ultrasonic frequency selection and calculation, DOEs on the chosen design and successive iterations until the readings were accurate and repeatable to within 2%.
Getting it into the field took more than the sensor. He created a software calibration tool for the manufacturing line, wrote the manufacturing instructions and operating manuals, and worked with firefighters and subject-matter experts to keep the design aligned with NFPA and FCC requirements. He also had a customer-facing app developed, so users can watch flow and pressure in real time, download records for audits, and build the data history predictive analytics needs later.
Ultrasonic flow and pressure sensor with LTE, GPS and telemetry.
Stream IQ intelligent flow meter
Customer-facing app showing real-time flow and pressure curves
Calibration tool
- Compares values of Stream IQ with another known reference measurement
- Automatically defines stability regions across multiple flow regimes and calculates a k-factor for each
- Averages those k-factors and provides an accuracy measurement with PASS/FAIL criteria for production
Calibration study
- Initially the sensor was calibrated against a known calibrated turbine flow sensor; however, HEN’s design was ultrasonic and needed its own methodology to understand the design limitations
- Performed against known smooth-bore machined nozzles: 5 different types with 10 different flow vs. pressure profiles
- Added temperature compensation, since the density of water changes with temperature
- Tested in the test lab and in the field to understand the role of bubbles and debris in the system
- Integrated software filters to smooth the data within spec
Ultrasonic flow sensor
- Custom high-frequency transducer configuration
- Field-calibrated against a pitot tube
- The measurement system has a fixed resolution, so when delta_t is tiny even 1 ns of error has a big impact, whereas when delta_t is large 1 ns has a tiny impact
- Flow is derived from a timing signal. A small timing error causes a large relative error in flow at the low end, which is why ultrasonic sensors are more accurate at high flows
- Analogy: it’s like measuring distances with a defined ruler