In the automated monitoring of large structures such as bridges, dams, and high-rise buildings, we deploy thousands of sensors to detect indicators such as stress, deformation, and vibration. These sensors act as the sensory organs of the structure, but there is an almost unavoidable interference factor: environmental temperature, which can severely affect the accuracy of the signals transmitted by these sensors.
1. The Necessity of Temperature Compensation
To understand this issue, a classic example of a spring scale is very intuitive. For instance, if we want to use a spring scale to weigh a heavy object, it simulates the sensor measuring force or strain.
Ideal Situation: In a temperature-controlled laboratory, a 1 kilogram weight stretches the spring, and the pointer stabilizes at 1.0 kilograms. The measurement is accurate.
Field Situation:
On a summer afternoon, the surface temperature of the bridge can reach 60℃. The sensor monitoring the strain of the bridge has its internal “spring” (which may be a metal strain gauge or the fiber itself) soften due to heat. At this point, even if the actual load on the bridge remains unchanged, this softened “spring” will stretch a bit more, causing the system to mistakenly believe that the “weight has increased”, reading a false signal of 1.2 kilograms.
On a winter night, the temperature drops below 0℃. The “spring” becomes hard when cold. Under the same load, the spring does not stretch enough, and the system will mistakenly believe that the “weight has decreased”, reading 0.8 kilograms.
In structural monitoring, this “weight” represents strain or stress, which are the measured quantities. If we do not correct for temperature effects, we cannot determine whether changes in readings indicate a real problem with the structure (such as crack development leading to stress redistribution) or are merely due to temperature fluctuations, which would render the entire monitoring system ineffective for early warning.
2. Temperature Compensation Strategies
For large projects that need to operate continuously for decades, engineers adopt the following stable and reliable compensation strategies.
Strategy 1: Reference Sensor Method, setting up a “zero stress” observer: This is the most commonly used and reliable method in structural monitoring, especially suitable for fiber Bragg grating (FBG) and resistive strain gauges.
Specific Implementation:
1. Install a working sensor at the location where strain needs to be measured (e.g., on the surface of a steel beam).
2. Install an identical compensation sensor as close as possible next to it.
3.Key Difference: The working sensor is firmly bonded to the structure and deforms together with it; while the compensation sensor is installed in a special way (e.g., pre-loosened on a small bracket) to ensure it can sense the same temperature changes but does not bear mechanical stress from the structure.
Working Principle:
Working sensor reading = real strain signal + temperature interference signal
Compensation sensor reading = 0 + temperature interference signal
System Automatically Calculates:Working sensor reading – Compensation sensor reading = real strain signal
This is akin to placing a similar spring next to the spring scale, but this spring does not have a weight attached. When it is hot, both springs soften; when it is cold, both harden. We only need to focus on the difference in length between the two springs, which is only related to the weight and not to the temperature.
Strategy 2: Software Algorithm Compensation Method, directly obtaining the temperature-response corresponding scale to eliminate: For smart sensor systems with embedded processing units, this is a core method.
Specific Implementation:
1.Learning Phase (Calibration): At the initial state of structural completion or system installation (when the internal forces of the structure are known or relatively stable), the system will automatically record sensor readings over a whole year under different seasons and temperatures. Through long-term data accumulation, the system will learn and establish an accurate mathematical model: real response = original sensor reading – temperature influence function.
2.Working Phase (Compensation): Thereafter, the system collects readings from each sensor and the ambient temperature in real-time, automatically invoking this model for correction, directly outputting the real strain values.
This is like giving the monitoring system a “dictionary”, which knows that when it sees a current temperature of 35℃, it needs to subtract a specific value from the original reading to obtain the real answer.
Strategy 3: Material and Structural Design, making sensors “inherently” stable: This is a fundamental solution taken at the sensor manufacturing level.
Specific Implementation:
Self-compensating strain gauges: Use special alloy materials (such as Kovar alloy) to make strain gauges, ensuring their thermal expansion coefficient matches that of the monitored steel or concrete structures. When the temperature changes, the strain gauge and the structure expand and contract synchronously, fundamentally reducing the false strain caused by thermal output.
Optimized Installation Process: By improving adhesives and bonding processes, ensure that strain can be efficiently and losslessly transmitted from the structure to the sensor while reducing the influence of the adhesive layer’s own temperature drift.
In structural health monitoring, temperature compensation is a necessary consideration and a prerequisite for ensuring data validity. Without reliable temperature compensation, the monitoring system will be misled by changes in environmental temperature. Therefore, when selecting sensors and setting up monitoring systems, the concept of temperature compensation must be consistently integrated throughout.