CureCheck

A cheap tag cast into concrete that logs temperature and resistivity and signs every reading, so small builders can verify strength instead of trusting a few swappable test cubes.

Comment

Inspiration

Concrete strength is usually judged by a few small cubes cast on pour day. Those cubes may not represent the structure, and they can be swapped or faked. A 2020 study of Indian bridge failures from 1977 to 2017 found that 2,130 bridges failed, and about 10.1% of cases involved deterioration of materials. Natural disasters caused about 80%, so bad concrete is not the biggest cause, but it is a preventable one. Cases like a new Gujarat bridge that collapsed before opening (the contractor was blacklisted for substandard concrete) made us ask: why can't an ordinary builder cheaply check what is really inside the concrete?

What it does

CureCheck is a low-cost tag cast into concrete. It logs two signals from the moment of pouring:

  • Temperature, which shows how fast the concrete is gaining strength (the standard maturity method).
  • Electrical resistivity, which reflects water content and pore structure, so it can flag extra water added on site.

A phone reads the tag over NFC or Bluetooth. Each reading is timestamped, geotagged and signed, so records can't be swapped like test cubes. A model learns how sensor curves relate to crushed-cube results, and shows a strength estimate and an extra-water flag.

The maturity method estimates strength from the temperature history:

$$M = \sum (T_a - T_0)\,\Delta t$$

where ( T_a ) is the average concrete temperature in each interval, ( T_0 ) is a datum temperature, and ( \Delta t ) is the interval length.

How we built it

For this ideathon we designed the concept and the test plan. We did not build a working prototype. The plan:

  1. Build a tag from an ESP32, a thermistor and two stainless electrodes.
  2. Cast 20 to 30 cubes with deliberately different water content.
  3. Log temperature and resistivity from pour onward.
  4. Crush cubes at 3, 7 and 28 days and compare with the sensor's prediction.

We set the pass/fail rule before testing: the prediction must land inside an error limit we choose in advance (for example, 15%).

What we learned

  • Temperature-based strength sensors already exist (for example Giatec SmartRock), mostly on large projects.
  • Research and patents already combine temperature and resistivity. Imec's CoMModoRE project built an embedded sensor with AI strength estimation.
  • So the opportunity is not the sensing idea. It is making something cheap, tamper-proof and usable by small builders.

Challenges

  • Resistivity is a proxy for strength, not a direct measurement, so accuracy has to be proven by testing.
  • Calibration changes with each concrete mix, so a model needs data from many pours.
  • Getting engineers and inspectors to accept a new check takes time and validation.
  • Tamper-proofing means the signed record has to be trustworthy end to end.

What's next

Run the cube experiment, refine the model with more pours, and pilot with a small contractor or testing lab.

Built With

Share this project:

Updates