RESISTIVE THERMAL FLOW METER WITH COMPENSATED DRIFT AND METHOD FOR MANUFACTURING SUCH A FLOW METER

Description

Technical Field

[0001] The invention relates to a resistive heat flux sensor, or heat flux meter, and more particularly to a compensation for the measurement drift of a resistive heat flux meter. Previous technique

[0002] Along with pressure and temperature, heat flow is one of the major parameters to consider in the design of an engine and for improving the aerodynamics of a vehicle.

[0003] Measuring heat flux is particularly important in directly determining the cooling requirements of hot areas of blades or vanes.

[0004] Furthermore, if the surface temperature and gas temperature are known, measuring a heat flux can provide a value for the heat transfer coefficient that can be compared with the theoretical value provided.

[0005] A heat flux sensor, or heat flux meter, is a transducer that produces a signal proportional to the local heat flux. This heat flux can have different origins; in principle, heat fluxes by convection, radiation, and conduction can be measured. Heat flux meters consist of small thermocouples connected in series. Some instruments are actually specific heat flux sensors, such as pyranometers (for measuring solar radiation). Boelter-Schmidt gauges (for measuring the heat flux of fire) are also used. Other types include Gardon gauge heat flux meters, plug gauges, thin-film thermocouple arrays, and thin-film Wheatstone bridges.

[0006] The use of thin-film, thermocouple or Wheatstone bridge flow meters offers several advantages over other sensors.

[0007] Thin-film fluxmeters do not require a specific conformation of the component on which they are mounted. Furthermore, with a thickness typically less than 10 microns, they are thinner than other fluxmeters.

[0008] Thin-film fluxmeters are therefore less bulky and disruptive to the operating environment, and have a minimal impact on the physical and thermal characteristics of the support structure.

[0009] In general, the operation of a heat flux meter depends on Fourier's laws of thermal conduction. The temperature difference ΔT across a given thickness Δx of a material whose thermal conductivity k is also known is measured to then determine the heat flux Q. Q = Δ T Δ x

[0010] In a thin-film Wheatstone bridge heat flow meter, the temperature difference across the thickness Δx is measured by the variation in resistance in one or more arms of the Wheatstone bridge.

[0011] In a double-sided Wheatstone bridge heat flow meter configuration, two arms of the bridge are located on one surface of the substrate, while the other two arms are located on the opposite face. In a single-sided Wheatstone bridge heat flow meter configuration, all four arms are located on the same surface of the substrate, with two of the arms having a different insulation thickness than the other two.

[0012] If we consider a heat flux meter with a thickness I and a thermal conductivity k1, to measure the heat flux in a substrate of thickness L and thermal conductivity k2, the measured signal will satisfy the following equation: V sig = VβQ 1 k 1 2 + βQ 1 k 1 + 2 L k 2

[0013] With V the bias voltage of the bridge, β the temperature coefficient of resistance of the bridge material, and Q the heat flux.

[0014] Typically, in a double-sided configuration, the heat flux meter thickness I would be 1.016 mm, the substrate thickness L to be measured would be approximately 2.54 cm, the heat flux meter thermal conductivity k1 would be 36 W / m / K for alumina, and the substrate thermal conductivity k2 would be 15 W / m / K for stainless steel, and the thermal resistance coefficient of the bridge material would be on the order of 3.98 x 10⁻³ K⁻¹. With a bias voltage V on the order of 1 V and a heat flux on the order of 1 W / cm², a signal Vsig on the order of 528 (µV / V) / (W / cm²) would be obtained.

[0015] A Wheatstone bridge thin-film heat fluxmeter operates as follows. With no heat flux applied to the heat fluxmeter, all elements of the bridge, i.e., the four resistive arms, are at an initial temperature T₀ and have an initial resistance R₀. When heat flux is applied to the heat fluxmeter, the two resistive arms of the bridge that are not covered by an insulating layer are at a surface temperature TS, and the two resistive arms of the bridge covered by an insulating layer are at a bottom temperature TF. The resistance of the arms is then equal to R₀[1 + β(TS - T₀)] for the uninsulated arms and R₀[1 + β(TF - T₀)] for the insulated arms, respectively, where β is the linear temperature coefficient of the resistances. With a bias voltage V, the output signal of an isolated arm satisfies the following equation: V 1 = V R 0 1 + β T F − T 0 R 0 1 + β T S − T 0 + R 0 1 + β T F − T 0

[0016] And the output signal of a non-isolated arm satisfies the following equation: V 2 = V R 0 1 + β T S − T 0 R 0 1 + β T S − T 0 + R 0 1 + β T F − T 0

[0017] The output signal of the fluxmeter then satisfies the following equation: V sig = V 2 − V 1 = V β T S − T F 2 + β T F − T 0 + T S − T 0

[0018] The main known fluxmeters are based on a Wheatstone bridge deposited by a thin-film deposition process in two layers separated by a silica layer deposited by a thin-film deposition process. All the layers are deposited on an insulating substrate. One such fluxmeter based on a Wheatstone bridge is known from CN 106 225 959 A.

[0019] Unfortunately, it has been observed that heating of the heat flux meter leads to measurement drift. This drift distorts the measurement when the sensor is subjected to heat flux. Furthermore, in the absence of heat flux, the measurement does not return to zero, thus compromising subsequent measurements.

[0020] This drift stems from the overall temperature rise on the sensor's surface. The deposited bridge is not perfect; that is, the resistances of each of the four arms, or branches, of the Wheatstone bridge are not identical. In other words, the resistance values ​​at a given temperature are different, and their temperature changes vary. Thus, each resistance evolves differently with temperature, creating a parasitic imbalance in the Wheatstone bridge. This imbalance contributes to the observed drift and is responsible for the measurement not returning to zero in the absence of flux. Description of the invention

[0021] The main purpose of the present invention is therefore to provide a solution to compensate for the drift of the fluxmeter, that is to say to compensate for the imperfection of the Wheatstone bridge.

[0022] In a first object of the invention, a method for manufacturing a drift-compensated resistive heat flow meter is proposed, the heat flow meter comprising a substrate on which a Wheatstone bridge is deposited by a thin-film deposition process, having two first resistive branches formed in a first layer and two second resistive branches formed in a second layer, the first and second layers being separated by a silica layer deposited by a thin-film deposition process, the manufacturing process comprising: at least two sets of measurements at at least two different temperatures, each set of measurements comprising at least one measurement of the resistance of each first and second resistive branch of the Wheatstone bridge, one measurement of a fluxmeter supply voltage, one measurement of the fluxmeter output voltage, and one measurement of the fluxmeter temperature, for each first and second resistive branch of the Wheatstone bridge, a determination, from the measurements made, of a set of compensation resistors, an equivalent resistance of a resistive branch corresponding to the resistance resulting from a parallel coupling of a resistive branch of the Wheatstone bridge with the associated compensation resistor, a coupling of the compensation resistors to all or part of the resistive branches of the Wheatstone bridge.

[0023] In one aspect of the compensation process, the Wheatstone bridge can be made by a platinum-based thin-film deposition process, and the temperatures at which resistance measurements are taken can be within a temperature range three times smaller than the operating temperature range of the Wheatstone bridge.

[0024] When the temperature coefficient of variation of the Wheatstone bridge resistances is relatively constant, it is not necessary to characterize the resistances over the entire operating temperature range of the Wheatstone bridge. This allows the temperature range for measurements to be reduced to a more limited range.

[0025] In a second aspect of the compensation process, the determination of a compensation resistance value for each first and second resistive branch may include a selection of a resistance value from a given set of resistances.

[0026] The set of resistors provided preferably includes commercially available resistors.

[0027] In a second object of the invention, a resistive thermal fluxmeter is proposed comprising a substrate on which is deposited a Wheatstone bridge deposited by a thin film deposition process, the Wheatstone bridge having two first resistive branches formed in a first layer and two second resistive branches formed in a second layer, the first and second layers being separated by a layer of silica deposited by a thin film deposition process.

[0028] According to a general feature of the resistive heat flow meter according to the invention, the heat flow meter further comprises four compensating resistors, each compensating resistor being coupled in parallel to a specific resistive branch of the Wheatstone bridge, an equivalent resistance of each resistive branch having a coefficient of temperature variation equivalent to that of the other resistive branches up to one standard deviation, the equivalent resistance of a resistive branch corresponding to the resistance resulting from a parallel coupling of a resistive branch of the Wheatstone bridge with the associated compensating resistor.

[0029] Adding compensation resistors thus makes it possible to greatly reduce or even eliminate the non-return to zero and the drift observed during the measurement.

[0030] According to one aspect of the resistive heat fluxmeter, each compensation resistance preferably has a fixed value.

[0031] According to a second aspect of the resistive thermal fluxmeter, said substrate can be alumina.

[0032] According to a third aspect of the resistive heat fluxmeter, the first and second resistive branches of the Wheatstone bridge can be made of platinum.

[0033] According to a fourth aspect of the resistive heat fluxmeter, the compensation resistors can be mounted on a circuit remote from the substrate on which the resistive branches of the Wheatstone bridge are deposited. Brief description of the drawings

[0034] Other features and advantages of the present invention will become apparent from the description given below, with reference to the attached drawings which illustrate an example of an embodiment without any limiting character. [ Fig. 1 ] There figure 1schematically represents a top view of a drift-compensated resistive heat flow meter according to one embodiment of the invention. Fig. 2 ] There figure 2 schematically represents a cross-sectional view of the drift-compensated resistive heat fluxmeter of the figure 1 . [ Fig. 3 ] There figure 3 is a graphical representation of the time evolution of the output voltage of the resistive heat fluxmeter of the figure 1 compared to that of a resistive heat flux according to the state of the art. Fig. 4 ] There figure 4 schematically presents a flowchart of a manufacturing process for a resistive thermal fluxmeter. figure 1 according to a method of implementing the invention. Description of the implementation methods

[0035] On the figure 1 is schematically represented a top view of a resistive thermal fluxmeter 1 with drift compensation according to an embodiment of the invention, and on the figure 2A schematic cross-sectional view of the fluxmeter is shown. figure 1 .

[0036] As illustrated on the figures 1 and 2 The fluxmeter 1 comprises a substrate 2 on which a Wheatstone bridge 3 is deposited by a thin-film deposition process and includes four resistive arms 4 and 5. The Wheatstone bridge 2 has two first resistive arms 4 formed in a first layer 6 and two second resistive arms 5 formed in a second layer 7. The first layer 6 and the second layer 7 are separated by a silica layer 8 deposited by a thin-film deposition process. This results in a stack of layers comprising successively the substrate 2, the first layer 6, the silica layer 8, and the second layer 7, with the silica layer 8 thus isolating the first layer 6 from the second layer 7.

[0037] Each first resistive arm 4 is coupled between the two second resistive arms 5, and each second resistive arm 5 is coupled between the first two resistive arms 4, thus forming four successive nodes A, B, C, and D. Each resistive arm forms a resistance R1 to R4. The first resistance R1 is located on a second resistive arm 5 between nodes A and B, the second resistance R2 is located on a first resistive arm 4 between nodes B and C, the third resistance R3 is located on another second resistive arm 5 between nodes C and D, and the fourth resistance R2 is located on another first resistive arm 4 between nodes A and D.

[0038] The fluxmeter 1 further includes four compensation resistors 9. Each compensation resistor 9 is coupled in parallel to a separate resistive arm 4 or 5 of the Wheatstone bridge 3. A first compensation resistor 9 of value r1 is coupled in parallel to the resistive arm having the first resistance R1, a second compensation resistor 9 of value r2 is coupled in parallel to the resistive arm having the second resistance R2, a third compensation resistor 9 of value r3 is coupled in parallel to the resistive arm having the third resistance R3, a fourth compensation resistor 9 of value r4 is coupled in parallel to the resistive arm having the fourth resistance R4.

[0039] Thus, between each pair of nodes between which a resistive arm is located, we obtain the same equivalent resistance, that is to say an equivalent resistance equal to that of the others up to a standard deviation, the equivalent resistance between two nodes corresponding to the equivalent resistance of the parallel coupling of the resistance of the resistive arm with the associated compensation resistance.

[0040] The Wheatstone 3 bridge is supplied with an applied supply voltage V between nodes A and C, and the output voltage Vmes of the Wheatstone 3 bridge is measured between nodes B and D.

[0041] On the figure 3 The graph represents the time evolution in seconds of the output voltage V measured in millivolts of the resistive heat fluxmeter. figure 1 solid line compared to that of a resistive thermal flux according to the state of the art in dotted line.

[0042] We can see on this figure 3On the one hand, without compensation the output voltage never completely drops back to 0, unlike the measurement with the fluxmeter of the figure 1 , and, on the other hand, that the voltage measured by the state-of-the-art fluxmeter is never stable and constantly increases, unlike the measurement with the fluxmeter of the figure 1 is stabilized at the same voltage.

[0043] On the figure 4 A flowchart of a manufacturing process for the drift-compensated resistive heat fluxmeter 1 is shown. figure 1 .

[0044] In the first step of process 400, the flow meter is produced. figure 1 without the compensating resistors 9.

[0045] In a second step of procedure 410, at least two sets of measurements are performed at at least two distinct temperatures. For example, the fluxmeter without compensating resistance is placed in an environment at a first temperature between 0°C and 150°C, for example, 20°C. Once the fluxmeter without compensating resistance has stabilized at the first temperature, a first set of measurements is performed. Then, the fluxmeter without compensating resistance is placed in an environment at a second temperature between 0°C and 150°C, for example, 10°C. Once the fluxmeter without compensating resistance has stabilized at the first temperature, a second set of measurements is performed. Additional sets of measurements can be performed if necessary.

[0046] Each set of measurements includes a measurement of the resistance R1 to R4 of each first and second resistive arm 4 and 5 of the Wheatstone bridge 3, a measurement of the supply voltage Valim, a measurement of the output voltage Vmes, and a measurement of the temperature.

[0047] In a third step 420, for each first and second resistive arm 4 and 5 of the Wheatstone bridge 3, compensation resistances 9, r1 to r4 are determined from the measurements made, so that the equivalent resistance of each resistive arm 4 and 5 has a coefficient of temperature variation equal to or close to that of the other resistive arms within one standard deviation.

[0048] The equivalent resistance of a resistive arm corresponds to the resistance for a parallel coupling of a resistive arm 4 or 5 of the Wheatstone bridge 3 with the associated compensation resistance 9.

[0049] The compensation resistors (9) are determined from the impedance values ​​measured for each resistive branch and the bridge imbalance under a 10V DC supply, at various test temperatures. Measurements taken at different temperatures allow for the temperature characterization of the resistances of each branch before compensation. Then, for each resistive branch and for each pre-selected "compensation" resistor value, the temperature evolution of the bridge imbalance (over a predetermined temperature range based on the application and requirements) is calculated for identical or different supply conditions (voltage supply with a value other than 10V or DC supply). This then allows for the calculation of the root mean square deviation of the offset relative to its mean over the considered temperature range.The choice then falls on the set of compensating resistors 9 which minimizes this quadratic error. This is the set of resistors that will be installed. in fine on the product. After compensation, the imbalance of the bridge is consequently modified compared to the state before compensation.

[0050] In this third step, the value of the compensation resistors is chosen from a set of commercial resistors so that the production of fluxmeter 1 is simplified and its manufacturing cost is reduced.

[0051] In a fourth step 430, the compensation resistors 9 are coupled to each of the resistive arms 4 and 5 of the Wheatstone bridge 3.

[0052] The fluxmeter thus proposed by the invention makes it possible to compensate for the drift due to heat and thus to improve the accuracy of the fluxmeter and the consistency of its measurements.

Claims

1. A method of manufacturing a resistive heat flow meter with compensated thermal drift, the flow meter comprising a substrate on which is deposited, by a thin layer deposition method, a Wheatstone bridge comprising two first resistive branches formed in a first layer and two second resistive branches formed in a second layer, the first and second layers being separated by a layer of silica deposited by a thin layer deposition method, the manufacturing method comprising: - at least two series of measurements (410) at least at two different temperatures, each series of measurements comprising at least one measurement of the resistance of each first and second resistive branch of the Wheatstone bridge, a measurement of a supply voltage of the flow meter, a measurement of the output voltage of the flow meter, and a measurement of the temperature of the flow meter, - for each first and second resistive branch of the Wheatstone bridge, determining (420), from the measurements done, a set of compensation resistors, an equivalent resistance of each resistive branch having a coefficient of temperature variation equivalent to that of the other resistive branches to within one standard deviation, the equivalent resistance of a resistive branch corresponding to the resistance resulting from a parallel coupling of a resistive branch of the Wheatstone bridge with the associated compensation resistor, - coupling (430) the compensation resistors to all or part of the resistors of the Wheatstone bridge.

2. The method according to claim 1, wherein the Wheatstone bridge is made by a platinum-based thin layer deposition method, and the temperatures at which the resistance measurements are done are within a range three times smaller than the operating temperature range of the Wheatstone bridge.

3. The method according to one of claims 1 or 2, wherein the determination of a compensation resistance value for each first and second resistive branch comprises selecting a resistance value from among a given set of resistors.

4. The method according to claim 3, wherein the set of resistors comprises only commercial resistors.

5. A resistive heat flowmeter with compensated thermal drift (1) comprising a substrate (2) on which is deposited a Wheatstone bridge (3) deposited by a thin layer deposition method, the Wheatstone bridge (3) comprising two first resistive branches (4) formed in a first layer (6) and two second resistive branches (5) formed in a second layer (7), the first and second layers (6, 7) being separated by a layer of silica (8) deposited by a thin layer deposition method, the flow meter further comprising four compensation resistors (9), each compensation resistor (9) being coupled in parallel with a resistive branch (4, 5) specific to the Wheatstone bridge (3), an equivalent resistance of each resistive branch having a temperature variation coefficient equivalent to that of the other resistive branches (4, 5) to within one standard deviation, the equivalent resistance of a resistive branch (4, 5) corresponding to the resistance resulting from a parallel coupling of a resistive branch (4, 5) of the Wheatstone bridge (3) with the associated compensation resistor (9).

6. The flow meter (1) according to claim 5, wherein each compensation resistor has a fixed value.

7. The flow meter (1) according to one of claims 5 or 6, wherein said substrate (2) is made of alumina.

8. The flow meter (1) according to one of claims 5 to 7, wherein the first and second resistive branches (4, 5) of the Wheatstone bridge (3) are made of platinum.

9. The flow meter (1) according to one of claims 5 to 8, wherein the compensation resistors (9) are mounted on a remote circuit with respect to the substrate (2) on which the resistive branches (4, 5) of the Wheatstone bridge (3) are deposited.