Monitoring of industrial emissions using the Gauss method

Every year, the country's enterprises emit tens of thousands of tons of toxic substances into the atmosphere: co₂, noₓ, and ch₄ are the main components that affect human and ecosystem health.

Effects: the growth of respiratory diseases, acid rains, and the acceleration of climate change.

Studies show that when Co₂ concentrations exceed 1000 ppm, a person's cognitive function decreases by 15-20% (source: Harvard T.H. Chan School of Public Health).

Project conceptA team of schoolchildren has developed a monitoring system based on a CubeSat 1U (10×10×10 cm) nanosatellite. Key functions:

Gas detection: sensors MQ-135 (NOₓ, nh₃), MQ-9 (CO, ch₄), MQ-131 (O₃).

Forecasting the spread using ML models.

Comparison of data with ground stations.

Technical implementationHardware part:

PETG body with gyroid structure (strength up to 18G).

Power system: 2×Li-ion 18650 (5.5 hours of operation).

Thermal insulation: 4 mm foam + aluminum screen.

Software:

Gaussian model for calculating emission dispersion:

pythondef gaussian_model(Q, u, σy, σz, H):return (Q / (2 * np.pi * u * σy * σz)) * np.exp(-(H*2)/(2 * σz*2))Neural networks:

Motion forecast (92% accuracy, Random Forest + LSTM ensemble).

Long-term forecast of air quality (gradient boosting, 87% accuracy).

Methodology and resultsData collection:

NASA's open datasets and local measurements were used.

Video of the sensor calibration process: YouTube.

Validation:

Sensor error: ±5% for Co₂, ±7% for O₃.

Trajectory prediction error: <1.5 km/24 h.

3D satellite modelCubeSat designAnimation of the assembly: link.

Potential and developmentApplication: monitoring of oil fields, megacities, emergency zones.

Plans:

Stratospheric tests with Al-Farabi Kazakh National University.

Integration with emergency notification systems.

The project is open for collaboration: code and documentation on GitHub.

ConclusionThis project demonstrates how even school teams can offer innovative solutions to global environmental problems. The combination of hardware developments and ML algorithms opens up new possibilities for environmental monitoring.

Authors: Nurbakyt Karazhakov, Adilzhan Galimzhanov, Arkat Khasanov, Aslan Amangaliev (school No. 9, Kazakhstan). Participants of the NASA Space App Challenge, FIRST Tech Challenge.

VIDEO:https://youtu.be/JTQfSfqd8lM