Abstract
To improve the accuracy and reliability of temperature measurements, this paper presents a high-precision non-contact temperature measuring system utilizing a quartz temperature sensor array and microprocessor-based data fusion. The system employs a hybrid sensor consisting of a quartz tuning fork and a temperature-sensitive polymer line. By arranging these sensors in a diamond grid and applying Bayesian estimation for data fusion, the system achieves a measurement precision of 0.02°C and a sensitivity of $71.5 \times 10^{-6} / ^\circ C$ within a range of 0°C to 100°C.
Introduction
In modern industrial processes, such as metallurgy and pharmaceutical manufacturing, precise temperature control is vital. Traditional contact-based sensors, like thermocouples, often suffer from slow response times because they must reach thermal equilibrium with the target object. While non-contact infrared sensors offer speed, achieving high precision remains difficult. This research addresses these limitations by developing a low-cost, high-precision, non-contact sensing system that leverages the stability of quartz resonators and advanced data fusion techniques.
Methodology
The Hybrid Quartz-Polymer Sensor
The core of the system is a modified 32.768 kHz quartz tuning fork. A polymer line made of N-isopropyl acrylamide (NIPA) is attached to the fork's arms. NIPA is highly sensitive to temperature; changes in ambient heat cause the polymer to absorb or release water, altering its mass and elasticity. This physical change induces a shift in the resonant frequency of the quartz fork.
As shown in the figure above, the addition of the polymer line creates a distinct frequency offset compared to a standard quartz tuning fork, providing a highly accurate proxy for temperature measurement.
Diamond Grid Array Configuration
To cover a measurement area effectively, the sensors are deployed in a "diamond grid" layout rather than a standard rectangular array. This geometric arrangement ensures seamless coverage while minimizing interference between adjacent sensors.
The diamond layout reduces the "confidence distance" between sensors, meaning the data collected across the array is more consistent and compatible for the subsequent fusion process.
Bayesian Data Fusion
The system processes raw frequency data through a multi-stage algorithm to ensure reliability:
- Histogram Pruning: Outliers and invalid data points are identified and removed using a histogram method.
- Bayesian Estimation: The remaining data is fused using Bayesian math. By treating temperature as a random variable and incorporating prior information, the algorithm calculates the most probable "true" temperature, effectively smoothing out individual sensor errors and noise.
Results and Discussion
Experimental results demonstrate that the system is highly effective across a range of 0°C to 100°C. Key performance metrics include:
- Precision: 0.02°C measurement precision.
- Sensitivity: $71.5 \times 10^{-6} / ^\circ C$.
- Stability: The Bayesian fusion curve proved significantly more stable than raw data from individual sensors, demonstrating the robustness of the array approach.
The use of mass-produced quartz tuning forks keeps the hardware cost low, while the software-based fusion ensures high-end performance.
Conclusion
The proposed non-contact temperature sensing system offers a compelling combination of high precision, stability, and cost-effectiveness. By integrating physical sensor innovation with optimized spatial layouts and statistical data fusion, the researchers have created a reliable tool for industrial thermal monitoring. Future work may focus on the long-term environmental stability of the NIPA polymer and extending the operational range for high-temperature industrial applications.
Summary updated