WO2008142075A1 - Diagnostic method for thermal mass flow measuring devices - Google Patents

Abstract

The invention relates to a method for the diagnosis of a device for measuring a process variable of a flowing fluid (1), wherein a heatable heating element (4, TH) is heated using an alternating current or voltage signal, and dissipates the generated heat at least partially to the fluid (1), and a temperature sensor (TF) is placed in thermal contact with the fluid (1), and wherein the course of a heating and/or cooling process occurring in the temperature sensor (TF) is measured, and the state of the temperature sensor (TF) and/or of the heating element (4, TH), particularly a decontamination and/or a coating, is diagnosed from the same. The invention further relates to a method for the simultaneous throughflow measurement of a fluid (1) and the diagnosis of a device for measuring a mass flow of a flowing fluid (1) by using an adaptive signal control.

Description

 DIAGNOSTIC PROCEDURES FOR THERMAL MASS FLOW GAUGES

DESCRIPTION

Technical area

The invention relates to the field of thermal mass flowmeters. In particular, it relates to a method for diagnosing a mass flow meter according to the preamble of the independent claims.

State of the art

The measuring principle of thermal mass flow meters is based on the cooling of a heating element by the fluid flowing around the heating element. The decisive factor here is the heat transfer α from the outside of the heating element into the flow, which is the ratio between the introduced heating power density q and the temperature difference T ₀ -T _F between the surface of the heating element and

the fluid gives α = - - -. This heat transfer is directly dependent on

konstant und z.B. aus einer Kalibrierung bekannt sind.the flow around the heating element and can by an equation of the form α = A (pv) ^m + B. The heat transfer is thus directly dependent on the mass flow pv. The constants A, B and m are determined by the sensor structure and the thermal properties of the fluid. From the temperature difference between heater and fluid as well as the heating power supplied to the heater, the mass flow can now be calculated under the assumption that the constants

constant and known eg from a calibration.

In the case of thermal mass flow meters, in general, as described in DE 361 7770, the measurement is purely static with a constant heat output or a constant temperature difference between the heater and the flow. From the writings GBl 345324, WO03 / 0581 78, however, embodiments are also known in which a pulsed operation takes place. There pulsed operation is used with the aim of reducing the energy consumption for the flow measurement, or as in WO2006 / 01 8366, to determine both the flow rate and the fluid temperature using only one temperature sensor.

For the measurement methods described above, it is always assumed that the thermal contact between the heating element or the temperature measurement associated with the heating element and the sensor surface remains constant over the life of the measuring device. However, deposits on the sensor or changes in the sensor affect the thermal contact and change it. But this also changes the measuring characteristics of the sensor and leads to incorrect measurements in the flow measurement. As a result, the output measuring signal can deviate significantly from the actual flow rate. In order to be able to detect such changes, it is proposed in WO 2004/001 349 that, when operated with a heated and an unheated temperature sensor, the function of the to exchange both temperature sensors and to determine the flow also when the function of the sensors is changed. A deviation of the flow rates in normal and exchanged operation indicates a possible malfunction of the sensor. However, this detection requires a different change of the sensors or a different degree of deposition on the two sensors. In the case of the same deposit or change in the sensors, correct operation would be indicated despite a strongly deviating measurement signal. Also caused by the change of the two sensors an interruption in the measurement, so that a continuous operation of the meter is not possible.

Presentation of the invention

The present invention seeks to alleviate at least some of the above problems. The object is achieved by a method for diagnosing a device according to the independent claims. Further advantages, features, aspects and details of the invention will become apparent from the dependent claims, the description and the figures.

According to one aspect of the invention, a method for the diagnosis of a device for measuring a process variable of a flowing fluid is proposed. The device for measuring the process variable comprises a heating element, which is acted upon by an alternating current or voltage signal and thereby heated. As an alternating signal is here to understand the regular change between different current or voltage levels, which can be given by any function such as sinusoidal staircase or pulse sequences. Under alternating should also be understood a current or voltage, the a DC or DC voltage component and an AC or AC voltage component. The heat generated by the heating element is then released to the flowing fluid. Furthermore, the device comprises a sensor which is in thermal contact with the fluid, ie the sensor is typically surrounded by the fluid. The heating and cooling process in the sensor caused by the heating element is measured in the sensor. From the course of the measured curves, the state of the temperature sensor and / or the state of the heating element can be determined. Heating element and temperature sensor can be two separate components. Heating element and Temeperatursensor but may well be united in a single component, for example in the form of a heated Temsperatursensors with the heated and the temperature can be measured simultaneously. With the diagnostic method according to the invention, it is now possible in situ during operation of the device to determine the functionality and thus aging, soiling, deposits, erosion on the heating element and / or on the temperature sensor. By determining the functionality, the measurement of the process variable in the device can be corrected during the measurement.

According to a further aspect of the invention, a method is proposed with the steps defined in claim 1 0, in which a flow measurement of a fluid is made in a device and at the same time a diagnosis of the same device is created. The simultaneous flow measurement and diagnosis takes place in situ and thus during operation of the device. Any removal of heating element or sensor for this purpose is not necessary. Simultaneous flow measurement and diagnostics are performed with the measurement and diagnostic function determination at the same time or at least during the same measurement cycle.

The inventive method for simultaneous flow measurement and for the diagnosis of the measuring device thus allows to check the functionality of the measuring device during the measuring operation. This is particularly advantageous in creeping or continuously occurring measurement errors that cause no complete malfunction of the device and thus would remain undetected over a long period of time until a revision of the device.

In embodiments, the measured process size of the fluid is the mass flow or the temperature of the fluid. In embodiments, the diagnosed condition of the temperature sensor or state of the heating element provides information about the operability, i. about degree of soiling or coating of temperature sensor or

Heating element. This makes it possible to monitor the temperature sensor and the heating element, thus avoiding errors when measuring the mass flow or the temperature.

In embodiments, the alternating current or voltage signal supplied to the heating element is a quasi-periodic signal. A quasi-periodic signal is characterized by the fact that the period duration is finite, thus interruptions in the period occur and the period length can change after interruption of the period. In still other embodiments, the current or voltage signal may be a periodic, in particular a sinusoidal or cosinusoidal signal. Brief description of the drawings

They show schematically:

Fig. 1; Model thermal mass flowmeter

Fig. 2; Simulation of the frequency response (amplitude) of the heater without a) and b) with 1 00 μm coating

Fig. 3: Simulation of the frequency response (phase) of the heater without a) and b) with 100 μm coating

Fig. 4; Simulation of the change in amplitude relative to a

Reference amplitude (no coating, v = 1 0 m / s) a) due to a change in the flow velocity of 1 - 100 m / s, b) due to a change in the coating from 1 - 100 μm, excitation power: sinusoidal (P = I + sin (2 pi ft)) with 1 Hz (2 Hz)

Fig. 5; Control structure with adaptive signal control

Fig. 6; Algorithm of adaptive signal control

Fig. 7; Course of the temperature by a one-dimensional rod at different excitation frequencies of the heat source

The reference numerals used in the drawings and their meaning are listed in the list of reference numerals. Basically, the same or equivalent parts are provided with the same or similar reference numerals in the figures. For the understanding of the invention non-essential parts are not shown in part. The described

Exemplary embodiments are exemplary of the subject matter of the invention and have no limiting effect. Ways to carry out the invention

In order to detect measurement errors of a thermal mass flow meter, it is proposed to change the heating power of the heater 4 in time and to evaluate the response to this change in the temperature signal with respect to the contamination of the sensor. In particular, the difference in the behavior of higher and low frequency signals in the sensor is exploited for this purpose. The low-frequency components (<0.1-l Hz depending on the sensor structure) in the temperature signal are determined both by the heat transfer on the outside, which is given by the flow, and the thermal conductivity of the sensor. In contrast, higher-frequency components are primarily determined by the heat capacity or the coupling of heat capacity and thermal conductivity of the sensor and of the fluid. Since the heat capacity, in particular of gases, is very small in comparison to the heat capacity of a solid, as used for the sensor element, the gas and thus the flow hardly influence the high-frequency signals. On the other hand, a contamination or surface abrasion applied to the surface influences it very much. In the high-frequency measurement signal, this causes a change in the amplitude and the phase for the individual frequencies, which can be detected by the evaluation of individual frequencies or a comparison of different frequencies. This can be done both in the frequency range, as described by the direct comparison of different frequencies, as well as in the time domain by a corresponding evaluation of the time course of the signal.

Figure 1 shows schematically the cross section through a layer-like (hatched shown) constructed heating element 4, on which a coating 2 has deposited and which is flowed by a fluid 1 in the arrow direction. The associated electrical equivalent circuit diagram illustrates the heat transfer from the location of the heat generation 5 by the heating element 4 in the fluid. 1 The introduced heating power P is attenuated due to the structure (layers) of the heater 4, a coating 2 and the fluid 1, symbolized by attenuators RiCi, R2C2, R3C3, RcCc and the

Heat transfer resistance RM between the surface of the heater 4 and the fluid 1.

In practice, the heating power of the heating element 4 is now selectively varied. This can be done on the one hand in a special measurement in offline mode, but it can also be measured online by modulating on the static measurement signal a time-varying signal. This can be filtered out again in a later step for flow measurement. Or if the temperature sensor already works in a pulsed mode for the flow measurement, these changes in the heating power can additionally be utilized for the coating detection. The shape in which the heating power P is varied may be selected differently. It is important only that relevant frequency components are included or at least in addition to the static fundamental signal f = 0 Hz a higher-frequency signal is thus present with a frequency f> 0 Hz. The sensor reacts to this variation of the heating signal with a temperature change at the sensor, which is analyzed for coating detection as described above.

This type of diagnosis is particularly interesting for thermal mass flow sensors for gases, since the measurement signal is very strongly influenced by a coating or change in the structure. For liquids, the diagnosis is more difficult but can still be used. Further applications are pure temperature sensors. In these, any coating, debris or contamination will initially cause no change in the sensor signal, it will especially only the response time influenced. But with very thick deposits or interruptions in thermal contact between the sensor element and the medium to be measured, this diagnosis may be advantageous. With this diagnosis it is possible to detect coatings on the sensor or erosions from the surface of the sensor. Likewise, however, fractures within the sensor, which arise, for example, as a result of tensions or chemical changes in the sensor, can be detected. This process even reacts much more sensitively to these because they are closer to the sensor element. For very sensitive measurement technology or at

In addition, even fluid properties of the fluid, such as its heat capacity, could be measured and liquids could be output as an additional diagnostic function.

The sensor shows even at low frequencies depending on the coating 2 a different frequency response as shown in Figure 2 and Figure 3.

FIG. 2 shows the simulation of the frequency response (amplitude) of the sensor in FIG

Graph (a) without coating and in graph b) with a 1 00 μm

Coating. The relative amplitude is plotted against the frequency in Hz. A significant change in amplitude between coated sensor and uncoated sensor is in the range between

0.3 Hz and 1 2 Hz detectable. An appropriate simulation of the

Frequency response (phase φ) of the sensor was made in Figure 3, wherein

Graph a) the simulation for a sensor without coating and graph b) the

Simulation with 1 00 μm coating on the sensor shows. It is the relative phase shift φ plotted against the frequency f.

In particular, in the range 1 -1 0 Hz in this example, the amplitude and phase response of the coated sensor deviate significantly from the uncoated sensor. In contrast, these frequencies become influenced by a change in the flow rate only slightly. This behavior is used to determine the coating without disturbing the flow measurement.

The principle of the measurement is based on exciting the sensor with a frequency in which e.g. the heating power is varied sinusoidally, and then the amplitude or the phase is analyzed by the resulting temperature signal. Alternatively, the sensor can be excited with several frequencies or a complete frequency spectrum and then the interesting frequency (s) is filtered out. The amplitude or phase of this frequency is measured once in the original state of the sensor and used as a reference. Later, this amplitude or phase is again measured during operation and compared with the initial state. Since the amplitude or phase is changed by the coating state, a change in the amplitude indicates the coating of the sensor. The frequency should be chosen so that it is only slightly influenced by a change in flow.

This process is shown in FIG. Here, the change in amplitude at a frequency of 1 Hz is represented by a coating. It clearly shows how this amplitude A / Ao changes with increasing coating thickness d _coa t. If the material of the coating is known, it is even possible to directly determine the coating _thickness d _coa t. At the same time once the flow Ffi _{ow has} been varied within the general flow range. This shows that even for this frequency, the change in the flow Ffow _{ow has} only a minimal effect on this amplitude A / O _o and a change in the amplitude A / Ao is thus a clear indication for the coating of the sensor. In addition, the influence of a flow change but also by the measurement at other eg lower frequencies can be corrected, as they react more to the flow. For example, in the static case the flow is measured and with this value the result for the coating is corrected.

In the example presented here, the interesting frequencies, which are most sensitive to a coating, are in the range of 1 -1 0 Hz. In general, this range for a general sensor is in the range of the "natural frequency" of the sensor, which follows

can calculate, where a = - the thermal diffusivity or

Thermal diffusivity and λ is the thermal conductivity, p the density and c the specific heat capacity of the sensor, d represents the distance between the heating wires and the surface. In the example shown here, this frequency is about 5 Hz.

Furthermore, a localization of malfunction occurring at the sensor can be made with the diagnostic method. For this purpose, various measurements are made by means of the temperature sensor, wherein the AC component of the heating power generated in the heating element is excited in each case with a different frequency. It is used that due to the periodic component in the heating signal, the propagating amplitude of the spatial temperature distribution is frequency-dependent and the propagating amplitude attenuates faster at higher frequencies than at lower frequencies. At higher frequencies, therefore, there is a greater localization of the amplitude around the heating element. Such a behavior is illustrated in FIG. 8 for a one-dimensional rod-shaped body and can be calculated for the respective body by means of the heat conduction equation. The Temperature T _re ι is applied over the distance three to the heating source for selected excitation frequencies f / b = 0.1, f / b = 1, f / b = 1 O and f / b = 100. Solutions of the associated heat equation are in this case of the form

T (x) = A ₁ * e ^a + A ₂ * e ^{~ a} with κ = (l + i) * J-

V b and where f is the exciting frequency, K is the wave vector, x is the distance and b is a material-dependent constant. It can be seen that higher frequencies lead to a larger real part of the wave vector K and thus to a greater attenuation of the heat propagation. In particular, it can thereby determine the heat dissipation through the support of the sensor element and detect a change in the heat conduction, which is e.g. is caused by a change in a wall temperature, or a change in the heat conduction in the holder itself, since the amplitude and phase of the heat wave at low frequencies are changed by this heat transfer, while they are independent of it at high frequencies.

As a further diagnosis, cracks or breaks within the sensor can be detected. Here again, one chooses a frequency small enough that the heat wave reaches to the crack or break point and compares this with a value at such a high frequency at which this interruption of the heat flow no longer plays any role.

The device advantageously stores the behavior of the sensor at different flow rates and later compares it with the values during operation. A discrepancy between different measured values can then be used to localize the cause of the error and appropriate corrections to the flow signal can be made. The flow rate determination for thermal mass flow meters is performed in the low frequency range between O Hz to several Hz. Since the diagnosis, for example, the determination of a coating on the temperature sensor or the heating element to be performed in this frequency range, eliminates the possibility of filtering to separate the useful signals for the flow measurement of the useful signals for the diagnosis. Thus, no separation of the signals can be used in the conventional case, which would allow a simultaneous flow measurement with simultaneous diagnosis.

Figure 5 shows an inventive control loop 1 0 on the means of a

Algorithm allows independent and simultaneous measurement of flow rate and diagnosis. Diagnosis may be, for example, the detection of a coating on the heatable temperature sensor T _H. The control loop 10 has a first regulator R, an amplification unit V, a heatable temperature sensor T _H , a temperature sensor T _F for measuring the fluid temperature and an adaptive signal regulator A.

The resulting temperature difference between the temperature of the heatable temperature sensor T _H and the temperature of the fluid is an input to the controller R. The controller R, which is for example a proportional integral controller, regulates the temperature difference, so that the temperature of the heater T _H a assumes a constant higher temperature than that of the fluid 1, the latter being measured via the temperature sensor T _F. The required power at the output of the regulator R is adjusted so that the

Temperature increase corresponds to the reference value ref at the input of the regulator R. In order to characterize additional properties, the value for the output power of the regulator R becomes an alternating one Power signal P _a it superimposed by means of amplification unit V, which may consist of several periodic components. The output of the amplification unit V provides a time-varying power 1 1. The acted upon by the time-varying power 1 1 heater of the heatable temperature sensor T _H changes its temperature 1 2 in correlation with the applied power. This temperature 1 2 is measured by means of temperature sensor T _H. Depending on the properties of the heater and its coating 2, the properties of the surrounding fluid and its speed with respect to the heater, this correlation is different. The output signal of the heatable temperature sensor T _H is corrected by means of the output signal 1 3 of the adaptive signal regulator A. This corrected signal serves as input signal for the adaptive signal regulator A and as input variable for the regulator R. The adaptive signal control algorithm of the adaptive signal regulator A divides new the time-varying signal at the output of the heated temperature sensor T _H in two components: a DC signal (DC signal) and an AC signal (AC signal), the latter consists of periodic components. The DC signal is primarily a measure of the primary flow signal. The comparison of the superimposed AC power signal P _a to the AC content at the output of the heatable temperature sensor T _H is a measure of the properties of the heater and the surrounding media. The subtraction of these ascertained AC components supplies the overtemperature of the heater T _H , which has been adjusted by the superimposed signals, with respect to the temperature of the fluid.

Figure 6 shows an algorithm of adaptive signal control. The input signal 1 4 is present in the time domain continuously or time / value-discreet. With recursive use of the algorithm it corresponds to the Error signal between signal and compensation. The signal in the time domain is decomposed by product demodulation 15 into its Fourier components. This can also be done by means of a Fourier transformation, which becomes interesting with several simultaneous frequency components. By applying the "Adaptive Control Law", the Fourier coefficients x are transformed into vectors u by vector transformation.The matrix A contains the information about the system and is subject to stability criteria.The compensation signal is calculated in the recursive case by summing with the old compensation signal Fourier transformation or product modulation is calculated a compensation signal 1 3, which is given back into the measuring system.

LIST OF REFERENCE NUMBERS

a Curve for an uncoated sensor b Curve for a sensor with deposits Ci, C2, C3. Cc capacity

P heating power

Pait age. Power signal at the input of the amplification unit V

Ri, R2, R3, Rc resistance

RM resistance ref reference value

T _H heatable temperature sensor

T _F temperature sensor

1 fluid

2 deposition, contamination, coating on the heater 3 direction of heat transport

4 heating element

5 Place of heat generation in the heating element

1 0 control loop

1 1 Power signal at the output of the amplification unit 1 2 Temperature signal at the output of the heatable temperature sensor

1 3 correction signal u _. the adaptive signal controller

1 4 input signal of the algorithm

1 5 Product Modulation

Claims

, Method for diagnosing a device for measuring a process variable of a flowing fluid (1), wherein a heatable Heating element (4, T _H ) is heated with an alternating current or voltage signal and at least partially releases the heat generated to the fluid (1) and a temperature sensor (T _F , T _H ) is thermally brought into contact with the fluid (1) , and wherein the course of taking place in the temperature sensor heating and / or Abkühlvorganges is measured and from the condition of the temperature sensor (T _F , T _H ) and / or the heating element (4, T _H ) is diagnosed. 2. Method for diagnosing a device for measuring a Process variable according to claim 1, characterized in that the mass flow and / or the temperature is measured as the process variable and that from the state of the temperature sensor (T _F , TH) and / or the heating element (4, T _H ), the functionality, in particular pollution and / or a coating of Temperature sensor (T _F , T _H ) and / or the heating element (4, T _H ), is determined. 3. A method for diagnosing a device for measuring a process variable according to one of claims 1 or 2, characterized in that the alternating current or voltage signal is a quasi-periodic signal. 4. Method for diagnosis according to one of claims 1 to 3, characterized in that the alternating current or voltage signal is a periodic, in particular a sinusoidal signal. 5. A method for diagnosis according to one of the preceding claims, characterized in that from a DC signal or a low frequency quasi DC signal, the temperature of the fluid (1) is determined and / or the flow of the fluid (1) is determined. 6. Method for diagnosis according to one of the preceding claims, characterized in that the signal in the initial state, in particular in an unpolluted state of the heating element (4, TH) and the temperature sensor (T _F , T _H ) is determined and as a desired signal with an actual signal during operation and a deviation between the setpoint and actual signal Functional signal or a signal correction is determined. 7. Method for the diagnosis according to one of the preceding claims, characterized in that a phase and / or an amplitude of the desired signal is compared with a phase and / or an amplitude of the actual signal during operation. 8. Method for diagnosis according to one of the preceding claims, characterized in that the heating element (4, T _H ) is a heatable first temperature sensor (T _H ) and the Temperature sensor is a heatable second temperature sensor (T _H ), in particular that the first and second temperature sensor are interchangeable operable, for example, upon reversal of the flow direction of the fluid. 9. Method for the diagnosis according to claim 7, characterized in that the thickness of the coating is determined from the comparison of the actual signal and desired signal. 1 0. A method for simultaneous flow measurement of a fluid and diagnosis of a device for measuring a mass flow of a flowing fluid, wherein a heatable heating element (T _H ) is acted upon by a periodic current or voltage signal, wherein the course of the temperature sensor (T _F ) taking place Heating and / or Abkϋhlvorganges is measured and wherein a periodic portion of a signal which resulting from the course of the temperature sensor (TF) heating and / or cooling process is subtracted by means of an algorithm. 1 1. A method according to claim 1 0, characterized in that the signal is Fourier transformed and the low-frequency component for flow measurement and the higher-frequency component for monitoring the functionality, in particular the degree of contamination and optionally for determining a corrected flow measurement signal is used.