Fault detection in a thermal sensor device

Abstract

A thermal sensor device is configured to determine a fluid parameter of a fluid based on the heat transfer behavior of the fluid. The sensor device comprises one or more heaters and means for determining a response of the sensor device to heater power being supplied to the heaters. For detecting sensor faults, the sensor device is operated in two different modes of operation. First and second values (c.sub.static, c.sub.dynamic) of the same fluid parameter are determined in the two modes. A fault indicator value (F) is derived by comparing the first and second values. The first mode of operation may be a steady-state mode, the first value (c.sub.static) being based on a steady-state response of the sensor device to heater power being supplied to the heaters, and the second mode of operation may be a dynamic mode, the second value (c.sub.static) being based on a transient response.

Claims

1. A thermal sensor device for determining a fluid parameter associated with a fluid in thermal contact with the thermal sensor device based on a heat transfer behavior of the fluid, the thermal sensor device comprising: one or more heaters; means for determining a response of the sensor device to heater power being supplied to the one or more heaters; and processing circuitry for supplying the heater power and for processing the response of the sensor device to determine at least one value of the fluid parameter based on said response, the processing circuitry being configured to carry out a method for detecting faults of the thermal sensor device comprising the steps of: a) operating the thermal sensor device in a first mode of operation to determine a first value of the fluid parameter; b) operating the thermal sensor device in a second mode of operation to determine a second value of the fluid parameter; and c) deriving a fault indicator value based on a comparison of the first and second values of the fluid parameter.

2. The thermal sensor device of claim 1, wherein the means for determining a response of the sensor device to heater power comprise one or more temperature sensors, and/or wherein the means for determining a response of the sensor device to heater power comprise circuitry for measuring a resistance of the one or more heaters.

3. The thermal sensor device of claim 1, wherein the fault indicator value correlates with a difference or ratio of the first and second values of the fluid parameter.

4. The thermal sensor device of claim 1, wherein the processing circuitry is configured to carry out the steps of: repeating steps a) to c) at a plurality of different times; and based on the fault indicator values at the different times, extrapolating a predicted fault indicator value at a later time or determining a predicted time interval until the fault indicator value reaches a threshold.

5. The thermal sensor device of claim 1, wherein the one or more heaters and at least a portion of the processing circuitry are integrated on a common silicon chip.

6. The thermal sensor device of claim 1, wherein the fluid parameter is a material parameter that correlates with at least one of thermal conductivity, specific heat capacity and thermal diffusivity of the fluid.

7. The thermal sensor device of claim 6, wherein the fluid is a mixture of at least two constituents, and wherein the fluid parameter is a mixing ratio of the mixture or a concentration of one of the constituents in the mixture.

8. The thermal sensor device of claim 1, wherein the processing circuitry is configured to obtain at least one auxiliary parameter from at least one auxiliary sensor element, and wherein the processing circuitry is configured to take the auxiliary parameter into account in the determination of the first and second values of the fluid parameter, the first value having a different correlation with the auxiliary parameter than the second value.

9. The thermal sensor device of claim 8, wherein the at least one auxiliary parameter is ambient temperature, pressure and/or humidity of the fluid.

10. The thermal sensor device of claim 1, wherein the processing circuitry is configured to output the fault indicator value or a parameter derived therefrom.

11. The thermal sensor device of claim 10, comprising a dedicated contact for outputting the fault indicator value or the parameter derived therefrom.

12. The thermal sensor device of claim 10, wherein the processing circuitry is configured to derive a Boolean alarm indicator value from the fault indicator value and to output the Boolean alarm indicator value.

13. The thermal sensor device of claim 1, wherein the first mode of operation is a steady-state mode comprising: supplying heater power to at least one of the one or more heaters; measuring a steady-state response of the sensor device to the heater power using the means for determining a response of the sensor device to heater power; and determining the first value of the fluid parameter based on the measured steady-state response, and wherein the second mode of operation is a dynamic mode comprising: supplying time-variable heater power to at least one of the one or more heaters; measuring a transient response of the sensor device to the heater power using the means for determining a response of the sensor device to heater power; and determining the second value of the fluid parameter based on the measured transient response.

14. The thermal sensor device of claim 13, wherein the processing circuitry comprises an oscillator for supplying a clock signal at a reference frequency, wherein the processing circuitry is configured to measure the transient response in the second mode of operation relative to the reference frequency, and wherein the processing circuitry is configured to output the clock signal.

15. The thermal sensor device of claim 13, wherein the processing circuitry is configured to supply the heater power to the same heater or heaters in both the first mode of operation and the second mode of operation.

16. The thermal sensor device of claim 13, wherein the means for determining a response of the sensor device to heater power comprise one or more temperature sensors, and wherein the processing circuitry is configured to measure the responses of the one or more temperature sensors to the heater power in both the first mode of operation and the second mode of operation.

17. A method for detecting faults of a thermal sensor device comprising one or more heaters, means for determining a response of the sensor device to heater power being supplied to the one or more heaters, and processing circuitry for supplying the heater power and for processing the response of the sensor device to the heater power in order to determine, based on said response, at least one value of a fluid parameter of a fluid in thermal contact with the sensor device, the method comprising: a) operating the thermal sensor device in a first mode of operation to determine a first value of the fluid parameter; b) operating the thermal sensor device in a second mode of operation to determine a second value of the fluid parameter; and c) deriving a fault indicator value based on a comparison of the first and second values of the fluid parameter.

18. The method of claim 17, wherein the first mode of operation is a steady-state mode comprising: supplying heater power to at least one of the one or more heaters; measuring a steady-state response of the sensor device to the heater power using the means for determining a response of the sensor device to heater power; and determining the first value of the fluid parameter based on the measured steady-state response, and wherein the second mode of operation is a dynamic mode comprising: supplying heater power to at least one of the one or more heaters; measuring a transient response of the sensor device to the heater power; and determining the second value of the fluid parameter based on the measured transient response.

19. The method of claim 17, comprising a step of measuring at least one auxiliary parameter, wherein the auxiliary parameter is taken into account in the determination of the first and second values of the fluid parameter, the first value having a different correlation with the auxiliary parameter than the second value.

20. The method of claim 19, wherein the at least one auxiliary parameter is ambient temperature, pressure and/or humidity of the fluid.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Preferred embodiments of the invention are described in the following with reference to the drawings, which are for the purpose of illustrating the present preferred embodiments of the invention and not for the purpose of limiting the same. In the drawings,

(2) FIG. 1 shows, in a schematic top view, a first embodiment of a thermal sensor device as it is known from the prior art;

(3) FIG. 2 shows, in a schematic perspective view, part of the sensor element of the thermal sensor device of the first embodiment;

(4) FIG. 3 shows, in a schematic sketch, a second embodiment of a thermal sensor device as it is known from the prior art;

(5) FIG. 4 shows a schematic block diagram of an embodiment of processing circuitry that may be used to implement the present invention;

(6) FIG. 5A shows a diagram illustrating a first, steady-state mode of operation of a thermal sensor device;

(7) FIG. 5B shows a diagram illustrating a second, dynamic mode of operation of a thermal sensor device;

(8) FIG. 6 shows a diagram illustrating a dependence of the change of a hydrogen concentration signal on temperature for measurements in steady-state and dynamic mode;

(9) FIG. 7 shows a diagram illustrating a dependence of the change of a hydrogen concentration signal on relative humidity for measurements in steady-state and dynamic mode;

(10) FIG. 8 shows a diagram illustrating a dependence of the change of a hydrogen concentration signal on pressure for measurements in steady-state and dynamic mode; and

(11) FIG. 9 shows a flow chart illustrating a sequence of steps according to an embodiment of the present invention.

DESCRIPTION OF PREFERRED EMBODIMENTS

(12) Thermal Sensor Device with Membrane

(13) FIGS. 1 and 2 illustrate a thermal sensor device 1 according to a first embodiment. The setup of such a sensor device is disclosed in U.S. Pat. No. 7,188,519B2.

(14) The thermal sensor device 1 comprises a microthermal sensor element 10 and processing circuitry 50 integrated on a common silicon chip 11. Contact pads 51 are provided for interfacing the flow sensor 1 with external circuitry. FIG. 2 shows a portion of the sensor element 10 in a perspective view.

(15) The silicon chip comprises a stack of dielectric layers, metal layers and polysilicon layers. The processing circuitry is formed in this layer stack by a CMOS process. For creating the sensor element 10, a resistive heater 21, a first temperature sensor 31 and a second temperature sensor 32 are formed in or on the layer stack. In the region of the heater 21 and the temperature sensors 31, 32, an opening or recess 12 is etched into the silicon chip 11 from below such that a thin dielectric membrane 13 remains, the membrane spanning the opening or recess 12. At least a portion of the heater 21 and of each temperature sensor 31, 32 is arranged on or in the membrane. In the present example, each of the temperature sensors 31, 32 consists of a thermopile, one set of junctions being disposed on the membrane and the other set of junctions being disposed on the surrounding bulk chip material. Instead, another type of temperature sensor can be used, e.g., resistive temperature sensors.

(16) The sensor element 10 is connected to the processing circuitry 50. The processing circuitry provides heater current to the heater 21 and reads out the temperature sensors 31, 32. In addition, the processing circuitry may be configured to determine the resistance of the heater 21.

(17) Also connected to the processing circuitry 50 is a reference temperature sensor 41 for determining the temperature of the bulk material of the silicon chip 11 that surrounds the membrane. In thermal equilibrium, this temperature will be approximately equal to the ambient temperature of the surrounding fluid. Further auxiliary sensors may be connected to the processing circuitry 50 for determining further auxiliary parameters, such as a pressure sensor for sensing the pressure of the fluid or a relative humidity sensor for sensing relative humidity of the fluid.

(18) Thermal Sensor Device with Bridges

(19) FIG. 3 shows, in a schematic manner, a thermal sensor device according to a second embodiment. The setup of such a sensor device is disclosed in EP3367087A2.

(20) As in the first embodiment, the thermal sensor device comprises a microthermal sensor element 10 connected to processing circuitry 50. Again, for creating the sensor element 10, an opening or recess 12 has been formed in the silicon chip 11. However, instead of an integral membrane, a plurality of bridges span this opening or recess 12, the bridges being separated by voids. Similar to the membrane of the first embodiment, each bridge may be formed by a plurality of dielectric layers, metal layers and/or polysilicon layers patterned from a layer stack on the silicon chip 11.

(21) In the present example, five bridges are present. Three of the bridges are heater bridges, carrying heaters 21, 22, 23, respectively. The processing circuitry 50 supplies the heaters with heater currents Ih1, Ih2 and Ih3, respectively. The heater bridges carry heater temperature sensors 33, 34, 35 for measuring the resulting heater temperatures Th1, Th2 and Th3, respectively. In the alternative or additionally, the processing circuitry may be configured to determine the resistance of each heater element 21, 22, 23 for the purpose of determining heater temperatures. The other two bridges are sensing bridges, carrying temperature sensors 31, 32 for measuring temperatures Tm1 and Tm2, respectively. Each sensing bridge is arranged between two of the heater bridges. While in the present example three heater bridges and two sensing bridges are present, different numbers of heater and sensing bridges may be provided. For instance, only one single heater bridge an only one single sensor bridge may be provided. Furthermore, while in the present example all bridges have the same distance from each other, these distances may also be unequal.

(22) As in the first embodiment, a reference temperature sensor 41 for determining a reference temperature Tref that is indicative of the ambient temperature of the surrounding fluid is connected to the processing circuitry 50. Also connected to the processing circuitry are a pressure sensor 42 for determining a pressure parameter p of the fluid and a relative humidity sensor 43 for determining a humidity parameter RH of the fluid. The temperature sensor, the pressure sensor and the humidity sensor may be provided on the same chip or on a different chip than the sensor element 10.

(23) Processing Circuitry

(24) FIG. 4 illustrates, in a highly schematic manner, a block diagram of a possible embodiment of the processing circuitry 50 of the first or second embodiment. The processing circuitry comprises a processor (?P) 501, a non-volatile (e.g., Flash ROM) memory 502, and a volatile (RAM) memory 506. The processor ?P communicates with the memory devices 502, 506 via a bus 510. The non-volatile memory 502 stores, inter alia, plural lookup tables (LUT), only two such lookup tables 503, 504 being illustrated. The non-volatile memory 202 further stores a machine-executable program (Prog) 505 for execution in the processor ?P. Via a device interface (IF) 507, the processing circuitry 50 drives the heater elements 21-23 and communicates with the various integrated or external sensors 31-35 and 41-43. A wired or wireless input/output interface I/O 208 enables communication to the outside world. An oscillator (Osc) 509 provides a clock signal with frequency f.sub.clock to the processor 501. The oscillator may have an output contact that allows the clock signal to be directly read out by external circuitry. A dedicated alarm contact 511 enables the output of a binary value representing a Boolean alarm indicator value, as will be explained in more detail below.

(25) The processing circuitry 50 may be completely integrated on the same silicon chip as the sensing element 10, or at least parts of the processing circuitry 50 may be implemented separately from the sensing element 10.

(26) Operation

(27) In operation, the sensor element 10 is exposed to a fluid of interest. The processing circuitry provides heater power to the heaters 21-23 and measures the resulting temperatures of the temperature sensors 31-35 and/or the resulting resistances of the heaters 21-23. The processing circuitry also measures the reference temperature T.sub.ref, the pressure parameter p and the humidity parameter RH, using the auxiliary sensors 41-43.

(28) The processing circuitry carries out two different modes of operation, as illustrated schematically in FIGS. 5A and 5B.

(29) First Mode of Operation

(30) The first mode of operation is a steady-state mode, as illustrated in FIG. 5A. Heater power is switched on and provided for a sufficiently long time that the temperatures measured by the temperature sensors 31-35 and/or the resistances of the heaters 21-23 reach a steady state, and these temperatures and/or resistances are measured.

(31) For instance, in the first embodiment, a heater power P(t) may be applied to the heater 21. Initially, the heater power is zero. At some point in time, the heater power is switched on and is kept constant at a value P.sub.m. The resulting temperatures at the temperature sensors 31, 32 are measured. From these temperatures, a linear combination may be formed, for instance, their sum. In FIG. 5A, this sum is designated as T.sub.s(T). This sum will change over time until it reaches a steady-state value. In FIG. 5A, the difference between the steady-state values of this sum after and before switching on the heater power is designated as ?T.sub.s.

(32) Likewise, in the second embodiment, constant power may be applied to one or more of the heaters 21-23, and the steady-state responses of the temperature sensors 31-35 and/or of the heater resistances may be measured. Combinations of the measured values may again be formed. These combinations or, more generally speaking, intermediate values may of course be more complex than a simple sum. In both embodiments, instead of applying a predetermined power, a predetermined voltage or current may be applied, or the heater power may be regulated to obtain a predetermined heater temperature.

(33) From the measured values and/or from the intermediate values, the processing circuitry 50 determines a first value of a fluid parameter associated with the fluid of interest. To this end, the processing circuitry may use one or more of the lookup tables.

(34) For instance, in the first embodiment, lookup table 503 may correlate ?T.sub.s to thermal conductivity of the fluid of interest. This correlation may have been determined beforehand by calibration measurements. Using the lookup table 503, the processing circuitry 50 may determine a first value for the thermal conductivity of the fluid, the first value being based to the measured value of ?T.sub.s.

(35) Second Mode of Operation

(36) The second mode of operation is a dynamic mode, as illustrated in FIG. 5B. Heater power is varied as a function of time. The transient response of the temperature sensors 31-35 and/or of the resistances of the heaters 21-23 is measured.

(37) For instance, in the first embodiment, a heater power P(t) may again be applied to heater 21. In the second mode of operation, the heater power P(t) now varies periodically, in the present example, sinusoidally. The transient response of the sum signal T.sub.s(t) from temperature sensors 31, 32 is now measured. In FIG. 5B, this transient response has been normalized such that its amplitude in the graph is the same as the amplitude of the heater power P(t). A time lag ?t or, equivalently, a phase difference between the heater power P(t) and the sum signal T.sub.s(t) is determined.

(38) In the second embodiment, one or more time lags or phase differences between heater power and the response of the temperature sensors or heater resistances may likewise be determined.

(39) From the measured time lags or phase differences and/or from intermediate values that have been calculated therefrom, the processing circuitry 50 determines a second value of the fluid parameter. To this end, the processing circuitry may use one or more additional lookup tables.

(40) For instance, in the first embodiment, lookup table 504 may correlate the time lag ?t to thermal conductivity. Again, this correlation may have been determined beforehand by calibration measurements. Using the lookup table 504, the processing circuitry 50 may determine a second value for the thermal conductivity of the fluid, the second value being based to the measured value of ?t.

(41) Determination of Concentration or Mixing Ratio

(42) If the fluid is a mixture of known constituents, knowledge of the thermal conductivity of the fluid allows an inference about the mixing ratio of the fluid or, equivalently, about the concentration of one of the constituents of the fluid.

(43) For instance, if the fluid is a mixture of hydrogen and air, knowledge of the thermal conductivity of the fluid allows an inference about the hydrogen concentration because the thermal conductivity of hydrogen is much larger than the thermal conductivity of air.

(44) Correction for Auxiliary Parameters

(45) The processing circuitry may correct the first and second values of the fluid parameter for variations of auxiliary parameters like ambient temperature, pressure and/or humidity of the fluid.

(46) For instance, the processing circuitry may correct first and second values of the thermal conductivity of the fluid, as determined by the first and second modes of operation of the sensor device, for deviations of the auxiliary parameters from standard conditions, as determined by the auxiliary sensors 41-43. In this manner, the processing circuitry may determine first and second thermal conductivity values at standard conditions.

(47) As another example, if the processing circuitry determines a concentration of a constituent of a mixture, the exact correlation between thermal conductivity and concentration may only be known at standard conditions, and therefore correction of the first and second values for deviations of the auxiliary parameters from standard conditions may be a prerequisite for obtaining a sufficiently precise value of the concentration.

(48) This is explained in more detail with reference to FIGS. 6 to 8. FIG. 6 illustrates the dependencies of the apparent hydrogen concentration in a mixture of hydrogen and air on temperature, as determined by the first and second modes of operation, if no temperature correction is applied. The broken line shows the temperature dependence of the 35 uncorrected apparent hydrogen concentration value as determined by the first mode of operation (steady state), while the solid line shows the temperature dependence of the uncorrected apparent hydrogen concentration value as determined by the second mode of operation. As can be seen from FIG. 6, the uncorrected apparent concentration values determined by both modes of operation strongly depend on temperature. For instance, if the device has been calibrated at 25? C. and the true hydrogen concentration is 0% mol, a measurement at 0? C., using the first mode of operation, would yield an apparent hydrogen concentration of approximately 2%. This may be unacceptable, and correction (temperature compensation) may therefore be needed.

(49) As also can be seen from FIG. 6, the dependencies of the apparent concentration values on temperature are different for the first and second modes of operation. Therefore, different corrections (compensations) should be applied to the values determined by the first and second modes of operation.

(50) Similarly, the determined hydrogen concentration values also depend on relative humidity and pressure. Again, the dependencies are different for the first and second modes of operation, and different corrections (compensations) should therefore be applied to the values determined by the first and second modes.

(51) Determination of Average

(52) The processing circuitry 50 may calculate an average of the first and second values of the fluid parameter and output the average to external circuitry through the I/O interface 508.

(53) Determination of Fault Indicator Value

(54) As outlined above, the processing circuitry determines first and second values of the same fluid parameter associated with the fluid of interest, the first value being determined by a steady-state mode of operation while the second value is determined by a dynamic mode of operation. Theoretically, these values should be identical. Substantial deviations between these values therefore indicate sensor faults such as drifts or contaminations.

(55) This opens up the possibility for the processing circuitry to calculate a fault indicator value. In the simplest case, the fault indicator value is simply the difference between the first and second values or the absolute value of this difference. However, the fault indicator value may also be a more complex function of these values.

(56) For instance, if the first and second values are concentration values, the fault indicator value may be the absolute value of the difference between these concentration values.

(57) It is to be noted that the fault indicator value will not only reflect faults of the sensor element 10, but also malfunctions of the auxiliary sensors 41-43 because of the different corrections that are applied to the first and second values of the fluid parameter. While this might on first sight seem disadvantageous because it might not be possible to distinguish between faults of the sensor element 10 and faults of the auxiliary sensors 41-43, this is actually an advantage in safety-relevant applications: Any fault in the sensor device, regardless of its cause, will be reflected by the fault indicator value. Monitoring the fault indicator value therefore enables highly reliable failure detection regardless of the cause of the failure. Once a failure has been detected, appropriate measures may be taken, such as replacing the sensor device.

(58) Determination of Alarm Indicator Value

(59) The processing circuitry may monitor the fault indicator value and set an alarm indicator to an alarm value if the fault indicator value meets certain criteria. For instance, the alarm indicator may be a Boolean variable, whose value is set to True once the fault indicator value has exceeded a predetermined threshold or once the fault indicator value has started to increase more rapidly between subsequent determinations than expected. The processing circuitry may output the alarm indicator value via the dedicated alarm contact 511.

(60) Predictive Maintenance

(61) The processing circuitry 50 may monitor the fault indicator value at different times and extrapolate an expected future value of the flow indicator or predict a predicted time to failure using past values. To this end, the ROM 502 may comprise a memory portion reserved for storing past fault indicator values, and the processing circuitry may read such past values and use them together with the most recent value to carry out the extrapolation or prediction. The processing circuitry may output the extrapolated future value or the predicted time to failure to external circuitry via the I/O interface 508.

(62) Flow Chart

(63) The above-described operation of the sensor device is summarized as a flow chart in FIG. 9. As in the above examples, it is assumed that the fluid parameter determined by the sensor device is a concentration of a constituent of a gas mixture.

(64) In step 301, the sensor element 10 is exposed to the gas mixture. In step 302, the sensor device is operated to carry out a steady-state measurement to determine a first value c.sub.static of the concentration. In step 303, the sensor device is operated to carry out a dynamic measurement to determine a second value c.sub.dynamic of the concentration. Each of steps 302 and 303 includes an appropriate correction for deviations from standard conditions, using the auxiliary sensors 41-43, as described above. In step 304, an average concentration c.sub.av=(c.sub.static+c.sub.dynamic)/2 is calculated. In step 305, a fault indicator F=|c.sub.static?c.sub.dynamic| is calculated. In step 306, the average concentration and the fault indicator are outputted through the I/O interface to external circuitry for further processing. In decision step 307, it is checked whether the fault indicator F has exceeded a threshold F.sub.crit. In the affirmative, an alarm is triggered by setting the alarm indicator value accordingly. In step 308, the current fault indicator value is stored in memory. In step 309, past flow indicator values are retrieved from memory, and the expected remaining lifetime (predicted time to failure) is computed. In step 310, the expected lifetime is outputted via I/O interface 508.

(65) Modifications

(66) While preferred embodiments of the invention have been described, it is to be understood that the invention is not limited to these embodiments, and that various modifications are possible without leaving the scope of the present invention.

(67) In particular, while the invention has been explained by the way of example of the determination of a concentration or mixing ratio, the thermal sensor device may also be configured to determine other fluid parameters, including other material parameters that are associated with the composition of the fluid as well as physical parameters associated with the fluid, in particular, its flow rate.

(68) The invention is not limited to the above-described examples of thermal sensor devices. The invention is applicable to any thermal sensor device that comprises at least one heater and means for determining a response to heater power being applied to the heater.

(69) While in the above embodiments it was assumed that the entire processing circuitry is integrated on the same silicon chip as the heaters and temperature sensors, some of the functionalities of the processing circuitry may also be implemented externally, e.g., in the form of software that is executed on external hardware. In particular, it is conceivable that the on-chip control circuitry only implements drivers for the heaters and readout circuitry for the temperature sensors and/or resistance values, while all further computations are carried out by external circuitry.