Multi-point distributed thermocouple array and real-time temperature field monitoring method

Abstract

The invention discloses a multipoint distributed thermocouple array and a real-time temperature field monitoring method, and relates to the technical field of temperature measurement. According to the array, micro thermojunctions are integrated on a substrate with the density larger than 1000 point cm, BiTe / Ag, Ni-Cr-Si / Cu-Ni and W-26% Re / Pt-13% Rh are adopted in a partitioned mode, and temperature measurement at the temperature ranging from-200 DEG C to 2000 DEG C is achieved. And the row and column gating-FPGA pipeline architecture completes full array scanning and IDW interpolation within 1ms, and outputs a high-resolution temperature field. An optical fiber or 1.6 GHz backscattering link guarantee gt; reliable data transmission under 900 DEG C and EMI environment is realized; a visual interface lt; and in 10s, an alarm is triggered or closed-loop cooling is executed. The device is suitable for aero-engines, high-temperature industry and biomedical precise thermal therapy.

Description

Technical Field

[0001] The present invention relates to the technical field of temperature measurement, in particular to a multi-point distributed thermocouple array and a real-time temperature field monitoring method. Background Art

[0002] Existing multi-point temperature measurement relies primarily on three approaches: single-material wire thermocouples, thin-film thermocouple arrays, and fiber-optic distributed temperature sensing. Although single-material thermocouples are resistant to high temperatures, they can only be deployed in small numbers, require complex wiring, and exhibit decreased sensitivity below -200°C. Thin-film arrays are limited to flexible electronics ≤100°C or precious metal sections ≤950°C, have a small number of points, slow scanning, and often use neural networks to compensate for crosstalk, resulting in significant inference delays. While fiber-optic DTS can measure temperature continuously over long distances, it is limited by pulse width and scattered signal intensity, resulting in a typical spatial resolution of ≥1m, which cannot meet the submillimeter requirements of turbine blades or tissue ablation. Furthermore, existing solutions often require expensive heat-resistant wiring or shielding in environments with high temperatures ≥900°C and strong electromagnetic interference, resulting in high overall system costs and extensive maintenance. Summary of the Invention

[0003] In order to overcome the defects of the above-mentioned prior art, the present invention provides the following technical solutions: a multi-point distributed thermocouple array, the array comprising: a multi-material gradient thermocouple array unit, which forms a thermoelectric junction matrix of more than 1000 points cm⁻² on a substrate, and the partitions respectively use a first thermoelectric material pair covering the temperature range of -200°C to 0°C, a second thermoelectric material pair covering the temperature range of 0°C to 900°C, and a third thermoelectric material pair covering the temperature range of 900°C to 2000°C, to achieve full temperature range measurement from -200°C to 2000°C; a cold end constant temperature reference zone, whose temperature is stable at a set reference value and is electrically connected to all cold ends of the thermocouple array unit through a thermal conductive line; a high A fast row and column gating signal acquisition module, which includes a row and column gating switch matrix for thermocouple array units, a low-noise amplifier and a high-resolution analog-to-digital converter, and is configured to complete the potential reading of all nodes of the thermoelectric junction matrix in ≤1ms; a hardware deterministic temperature field reconstruction module, which includes a table lookup linearization circuit and an inverse distance weighted interpolation circuit, and is configured to output a rasterized temperature field matrix immediately after receiving the read data; an optical and / or wireless signal transmission link, which is configured to transmit the read data to a safe area via optical or radio frequency signals in an environment with a temperature greater than 900°C or in the presence of electromagnetic interference; and a visualization and alarm interface, which is configured to output an alarm signal when the temperature exceeds the limit and display the temperature field matrix in real time.

[0004] Preferably, the first thermoelectric material pair is Bi2Te3 / Ag, which is suitable for -200°C to 0°C; the second thermoelectric material pair is Ni-Cr-Si / Cu-Ni, which is suitable for 0°C to 900°C; and the third thermoelectric material pair is tungsten-26% rhenium / platinum-13% rhodium, which is suitable for 900°C to 2000°C.

[0005] Preferably, the thermocouple array unit is prepared by MEMS-sputtering-3D direct writing composite process, and the thermoelectric junction unit area is .

[0006] Preferably, the row and column gating switch matrix is a GaN-HEMT board-level array, with a driving clock ≥40 MHz and a crosstalk suppression ratio ≥80 dB.

[0007] Preferably, the interpolation circuit of the hardware deterministic temperature field reconstruction module adopts an FPGA parallel pipeline to achieve a frame rate of ≥1kHz and a same-frame delay of ≤50µs.

[0008] Preferably, the optical signal transmission link is completed through 1550nm wavelength division multiplexing optical fiber, and the data bandwidth is ≥10Gbps; the wireless link adopts 1.6GHz zbackscatter, and the transmission distance is ≤50cm.

[0009] Preferably, the visualization and alarm interface outputs a TTL alarm with a delay of <10µs and synchronously sends a Modbus-TCP control word when the temperature exceeds a set threshold.

[0010] A real-time temperature field monitoring method for a multi-point distributed thermocouple array is characterized by comprising the following steps: S1, deploying a multi-material gradient thermocouple array unit on the surface or inside of an object to be measured and connecting it to a cold-end constant temperature reference zone; S2, starting a high-speed row and column selection signal acquisition module to read the output potential of all thermoelectric junctions with a period of ≤1ms; S3, performing linear conversion and interpolation processing on the read potential in a hardware deterministic temperature field reconstruction module, and outputting a temperature field matrix; S4, sending the temperature field matrix to an external visualization and alarm interface via an optical and / or wireless transmission link; S5, outputting an alarm or control signal when the temperature of any point in the temperature field matrix exceeds a preset threshold.

[0011] Preferably, the linearization conversion in step S3 adopts a fifth-order polynomial lookup table method, and the single-point conversion time is ≤100ns.

[0012] Preferably, the interpolation process in step S3 adopts an inverse distance weighted interpolation algorithm, the weight index is 2, and the interpolation range does not exceed 5×5 adjacent temperature measurement pixels.

[0013] Compared with the prior art, the present invention has the following advantages: (1) The present invention achieves continuous temperature measurement coverage from -200°C to 2000°C through the first, second and third thermoelectric material pairs in the same array, eliminating the complex process of traditional multi-probe segmented layout and switching calibration; maintains a reading error of ±0.5% in an ultra-wide temperature range, so that extremely low temperature and extremely high temperature conditions can be presented simultaneously on a "temperature map", greatly improving the efficiency of integrated measurement and control; (2) The array of the present invention forms more than 1000 points on the substrate Thermoelectric junction matrix, single junction area , the spatial resolution is improved to the submillimeter level; compared with traditional single-point or centimeter-level infrared temperature measurement, the fineness of thermal gradient imaging is significantly improved, providing a data basis for fine thermal management; (3) The row and column gating scheme of the present invention adopts GaN-HEMT switch matrix and parallel pipeline FPGA, with a complete frame reading time of 51.2µs and a whole machine frame rate of 1kHz. Combined with the intrinsic µs thermal inertia brought by the micro-nano heat capacity structure, it can analyze the transient temperature rise during shock wave, explosion or radio frequency ablation in less than milliseconds, breaking through the limitation of tens of milliseconds lag of existing thermocouples; (4) The fiber wavelength division multiplexing link and 1.6GHz backscattering link of the present invention respectively realize metal-free lead transmission for high temperature areas >900°C and strong electromagnetic interference areas, eliminating the risks of high temperature oxidation, contact resistance drift and EMI noise coupling. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0015] The technical solution of the present invention will be further described below with reference to the accompanying drawings and through specific implementation methods.

[0016] The real-time temperature field monitoring mechanism provided by this invention utilizes the Seebeck effect of thermocouples, enabling real-time, high-precision temperature field reconstruction. Direct thermoelectric conversion and differential measurement are employed: each microsensing unit comprises a thermoelectric junction made of two different materials, and the temperature difference directly generates a millivolt-level potential signal. A common reference node or busbar is designed into the array: for example, the cold ends of all thermocouples are connected to a constant-temperature reference area (or a reference junction built into the material). This allows the output voltage of each sensing point to effectively represent the temperature difference between that point and the reference. By integrating micro-thermostatic baths / reference blocks at the edges of the array or at specific locations, absolute temperature measurement is achieved, eliminating uncertainty in relative values at each point. High-frequency multi-channel scanning: To acquire data from all sensing points, a high-speed electronic switch matrix or integrated multi-channel readout circuit is employed. In hardware, the array's row and column conductors and thermoelectric junctions form a passive addressable matrix architecture. High-speed row-by-column scanning enables point-by-point polling and reading of thousands of sensors. Sub-millisecond overall scanning frequencies are achieved using a switch array chip or FPGA. For example, if there are 1000 temperature measurement points and each point takes 0.1ms to sample, then Full array scanning can be completed; through multi-channel parallel scanning and hardware pipelining, the total refresh cycle can be further compressed to milliseconds or even lower. For more demanding requirements, parallel readout can be performed by dividing the array into multiple blocks for simultaneous acquisition, achieving true real-time temperature field updates. Low-noise, high-speed amplification: Since the output voltage of a single microthermocouple is small (in the μV / °C range), signal conditioning circuitry is crucial. Each channel is equipped with a nanometer-scale low-noise amplifier and a high-speed analog-to-digital converter to ensure sufficient signal-to-noise ratio and resolution even with sub-millisecond sampling. Instead of traditional point-by-point manual cold junction compensation, an automatic cold junction compensation bridge is integrated into the circuit to correct for ambient temperature drift in real time. Furthermore, to minimize crosstalk, diode isolation or microelectromechanical switches are used at matrix intersections to ensure that each thermoelectric junction is disconnected from the readout circuitry when not in use, eliminating signal aliasing issues caused by dense wiring. Optical / wireless readout options: Optical or wireless signal transmission solutions are available for specialized environments, such as those with strong electromagnetic interference or extremely high temperatures where metal leads cannot be used. For example, the electrical signal from a microthermocouple can be converted into an optical signal. Using optical fiber as the transmission medium, the voltage change at each sensing node modulates a low-power laser beam (intensity or interferometric modulation). All optical signals are then multiplexed within the fiber at different wavelengths and time divisions to be transmitted out of the high-temperature zone. Alternatively, in the medium-temperature range (<300°C), a micro-radio frequency (RF) tag can be integrated. Powered by the thermocouple, it transmits a temperature-encoded signal, enabling contactless reading. This ensures reliable and real-time data acquisition, enabling uninterrupted monitoring even in extreme environments. Ultrafast response: The micro- and nanoscale sensing membrane and junctions offer extremely low intrinsic thermal inertia, enabling the capture of submillisecond or even nanosecond temperature changes. Thin-film thermocouple sensors can detect a sudden temperature rise of 200°C in less than 10 nanoseconds. The thin-film / microwire structure enables the sensor to respond to temperature changes almost instantaneously. Furthermore, the tiny size and extremely low thermal capacity of the sensing points in the array, combined with direct electrical signal conversion, avoid thermal hysteresis and complex signal processing delays. Combining these measures, rapid temperature changes can be captured in real time, such as combustion transients, explosion shocks, or temperature field dynamics during radiofrequency ablation surgery of human organs, providing temperature data with a time resolution of less than milliseconds for these transient processes.

[0017] Temperature field reconstruction and monitoring are directly achieved through physical models and electronic circuits. Real-time data calibration and linearization: The collected thermovoltage at each point first passes through an embedded calibration circuit, converting the voltage value into a temperature value based on pre-calibrated material thermoelectric characteristic curves. Because thermocouples of various materials are used, multiple corresponding calibration curves are stored internally and automatically switched according to sensor type. All conversion processes are performed in real time using hardware lookup tables or polynomial approximation circuits, eliminating the need for complex calculations. For a wide temperature range, a piecewise linearization strategy is employed to ensure excellent measurement accuracy from low to high temperatures. Temperature field reconstruction algorithm: After obtaining discrete point temperatures, the system must reconstruct them into a continuous temperature field map. Traditional solutions rely on machine learning for interpolation and fitting, while the present invention utilizes deterministic algorithms and hardware implementation. For example, inverse distance weighted interpolation (IDW) or spline interpolation is used to rapidly infer the temperature at any spatial location from the values of multiple adjacent sensor points. This algorithm can be implemented in parallel using an FPGA or signal processing ASIC. Interpolation operations for hundreds of points can be completed in microseconds, outputting a gridded temperature field matrix signal. In addition, physical models can be incorporated to aid specific applications. For example, when measuring temperature within solid materials, interpolation and extrapolation can be performed based on analytical solutions to the heat conduction equation. For fluid temperature field monitoring, fluid dynamics models can be incorporated to improve reconstruction accuracy. All of these computational processes are implemented in hardware logic and firmware, ensuring interpretable and reliable results. Data flow and visualization: The data processing chain is highly optimized, employing a pipelined architecture from sensor to temperature field image. Massive data from the sensor matrix is fed into the FPGA via a high-speed ADC, where calibration, filtering, and interpolation are performed in parallel to instantly generate a temperature field matrix. The temperature field data is then transmitted to a host computer or display module for real-time visualization, such as pseudo-color isothermal maps and 3D temperature distribution plots. Without lengthy algorithm iterations, each refresh achieves true "real-time" performance (with frame rates reaching hundreds of Hz). For critical monitoring, threshold monitoring can be configured in hardware: once the temperature in a specific area exceeds a set threshold, the system immediately triggers an alarm or initiates control (closed-loop cooling / interlock control: triggering spray cooling, power regulation, mechanical switching). Response time is limited only by electronic circuit latency (in the microsecond range).

[0018] The multi-point distributed thermocouple array adopts a five-stage composite process of "MEMS microstructure pre-processing - multi-target magnetron sputtering / electron beam evaporation - sol-gel 3D direct writing - laser localized cladding - rapid annealing + ALD surface protection". The specific steps and inspection standards are as follows: MEMS microstructure prefabrication: a. Substrate: 25µm thick polyimide-aluminum oxide composite film; b. Process: Deep reactive ion etching (DRIE) to form 100–500nm deep and shallow grooves and isolation islands; Photolithography minimum line width 5µm, sidewall angle ≥87°; c. Quality control: Etch depth uniformity <±5%, sidewall roughness <10nm (AFM), substrate thermal warpage <50µm.

[0019] Multi-target magnetron sputtering and electron beam evaporation zone deposition a. Chamber base pressure , substrate temperature 150–250°C; b. Low temperature deposition Film 600nm, sputtering rate c. Deposition of a 200nm Ni-Si seed layer in the medium-temperature zone; d. Co-deposition of 3µm thick W-26%Re and Pt-13%Rh films in the high-temperature zone, biased at −50V; e. Quality control: film stress <±100MPa (bending test), composition uniformity <±2at% (EDS), roughness Ra <5nm.

[0020] Sol-gel 3D direct writing wiring a. Nozzle diameter 2.5µm, printing speed b. Printing line width 8µm±0.5µm, thickness 5–10µm; c. Post-sintering at 400°C, ; d. Quality control: resistivity , perforation filling rate ≥98% (X-CT).

[0021] Laser localized cladding to construct thermoelectric junctions. a. Laser wavelength 355nm, pulse width 10ns, power 25mW; b. Scanning rate , fusing the bi-material film in a 30µm×30µm area; c. Quality control: thermoelectric junction resistance <10µΩ, alloy diffusion bandwidth <5µm (SEM-EDS).

[0022] Rapid annealing and ALD surface protection a. W-26% Re / Pt-13% Rh area: 1800°C, 15 min, argon rapid annealing; b. Bi2Te3 area: 350°C, 30 min, nitrogen annealing; c. ALD coverage of 100nm Y2O3 + 200nm SiO2 composite layer; d. Quality control: film stress after annealing < ±50MPa, ALD leakage current .

[0023] High-speed row and column gated acquisition: Matrix addressing: 250µm wide copper-nickel traces are set on each row and column of the array, dividing 2048×2048 temperature measurement pixels; the rows and columns are driven by N=2048 3-terminal GaN-HEMT switch arrays, polled at a 40MHz clock, and the complete frame read time is 51.2µs. Signal conditioning: LNA is arranged on the local amplifier board of each 64×64 pixel subarray (input noise ) + 18-bit 10MSPS SAR ADC; then, the data is sent in parallel to the FPGA via an LVDS link. Linearization and compensation: The FPGA uses a lookup table to segmentally fit the thermoelectric potential-temperature curves of different materials (fifth-order polynomials), with an error of ≤±0.1°C. Real-time interpolation: A hardware pipeline implements inverse distance weighted (IDW) interpolation with a weighting exponent α=2. For every 8×8 temperature measurement pixels, a 1× equivalent high-resolution grid point is generated, with a full-frame update rate of 1kHz.

[0024] Optical / wireless signal link: Optical readout: High-temperature voltage signal is converted to an analog voltage-controlled microring resonant modulator, with a 1550nm optical carrier transmitted via a single-mode fiber. The demodulated error is <±0.2°C. Wireless readout: Medium-temperature RF-tag operates at 1.6GHz, and the thermoelectric voltage is modulated into the carrier phase using a delta-sigma ADC. Off-state power consumption is <2µW.

[0025] Application Examples: Engine turbine blades: The outer surface of the blade was sprayed with heat-resistant oxidation ceramic and then coated with an array. Vacuum ground ignition tests showed a maximum temperature of 1540°C, a hotspot radius of 0.8mm, and a response time of 0.4ms. Radiofrequency ablation needles: A 32×32 embedded array was implanted on the outer wall of a 1.2mm diameter nickel-titanium tube. In an in vitro pig liver experiment, the diffusion trajectory of the 55°C isothermal surface as a function of power was fully characterized within 30 seconds.

Claims

1. A multi-point distributed thermocouple array, characterized in that: The array comprises: Multi-material gradient thermocouple array unit, which forms >1000 points on the substrate The thermoelectric junction matrix is constructed, and the partitions use a first thermoelectric material pair covering the temperature range of -200°C to 0°C, a second thermoelectric material pair covering the temperature range of 0°C to 900°C, and a third thermoelectric material pair covering the temperature range of 900°C to 2000°C, respectively, to achieve full temperature range measurement from -200°C to 2000°C; A cold-end constant-temperature reference zone, the temperature of which is stabilized at a set reference value and electrically connected to all cold ends of the thermocouple array units via a heat-conducting wire; A high-speed row and column gating signal acquisition module, which includes a row and column gating switch matrix for the thermocouple array elements, a low-noise amplifier, and a high-resolution analog-to-digital converter, configured to complete the potential reading of all nodes in the thermoelectric junction matrix in ≤1ms; A hardware deterministic temperature field reconstruction module, comprising a table lookup linearization circuit and an inverse distance weighted interpolation circuit, configured to output a rasterized temperature field matrix immediately after receiving read data; an optical and / or wireless signal transmission link configured to transmit the read data to a safe area via optical or radio frequency signals in an environment with a temperature greater than 900° C. or in the presence of electromagnetic interference; The visualization and alarm interface is configured to output an alarm signal when the temperature exceeds the limit and display the temperature field matrix in real time.

2. A multi-point distributed thermocouple array according to claim 1, characterized in that: The first thermoelectric material pair is Bi2Te3 / Ag, which is suitable for -200°C to 0°C; the second thermoelectric material pair is Ni-Cr-Si / Cu-Ni, which is suitable for 0°C to 900°C; the third thermoelectric material pair is tungsten-26% rhenium / platinum-13% rhodium, which is suitable for 900°C to 2000°C.

3. The multi-point distributed thermocouple array according to claim 1, wherein: The thermocouple array unit is prepared by MEMS-sputtering-3D direct writing composite process, and the thermoelectric junction unit area is .

4. The multi-point distributed thermocouple array according to claim 1, wherein: The row and column gating switch matrix is a GaN-HEMT board-level array, with a driving clock of ≥40 MHz and a crosstalk suppression ratio of ≥80 dB.

5. The multi-point distributed thermocouple array according to claim 1, wherein: The interpolation circuit of the hardware deterministic temperature field reconstruction module adopts an FPGA parallel pipeline to achieve a frame rate of ≥1kHz and a same-frame delay of ≤50µs.

6. The multi-point distributed thermocouple array according to claim 1, characterized in that: The optical signal transmission link is completed through 1550nm wavelength division multiplexing optical fiber, with a data bandwidth of ≥10Gbps; the wireless link adopts 1.6GHz backscatter, with a transmission distance of ≤50cm.

7. The multi-point distributed thermocouple array according to claim 1, characterized in that: The visualization and alarm interface outputs a TTL alarm with a delay of <10µs and simultaneously sends a Modbus-TCP control word when the temperature exceeds a set threshold.

8. A real-time temperature field monitoring method for a multi-point distributed thermocouple array according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1. Deploy a multi-material gradient thermocouple array unit on the surface or inside the object to be measured and connect it to the cold-end constant temperature reference area; S2, start the high-speed row and column selection signal acquisition module to read the output potential of all thermoelectric junctions with a period of ≤1ms; S3. Performing linear transformation and interpolation processing on the read potential in the hardware deterministic temperature field reconstruction module, and outputting the temperature field matrix; S4. Sending the temperature field matrix to an external visualization and alarm interface via an optical and / or wireless transmission link; S5. When the temperature of any point in the temperature field matrix exceeds the preset threshold, an alarm or control signal is output.

9. The real-time temperature field monitoring method according to claim 8, characterized in that: The linearization conversion in step S3 adopts a fifth-order polynomial lookup table method, and the single-point conversion time is ≤100ns.

10. The real-time temperature field monitoring method according to claim 9, characterized in that: The interpolation process in step S3 adopts an inverse distance weighted interpolation algorithm with a weight index of 2 and an interpolation range not exceeding 5×5 adjacent temperature measurement pixels.

Images