Method for determining a quantitative deposit indicator, and coriolis mass flow meter for carrying out the method

A method using device and material-specific parameters addresses the lack of a uniform fouling model in Coriolis mass flow meters, enabling accurate deposit detection and maintenance, enhancing fouling management across different configurations.

WO2026082515A1PCT designated stage Publication Date: 2026-04-23ENDRESS HAUSER FLOWTEC AG
View PDF 9 Cites 0 Cited by

Patent Information

Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
ENDRESS HAUSER FLOWTEC AG
Filing Date
2025-10-08
Publication Date
2026-04-23

AI Technical Summary

Technical Problem

Current methods lack a uniform model to describe fouling formation across different configurations and types of Coriolis mass flow meters, making it difficult to apply findings on fouling formation effectively.

Method used

A method to determine a quantitative deposit indicator using device and material-specific parameters, including viscosity and material damping properties, to assess the amount of deposit in measuring tubes, and a Coriolis mass flow meter equipped with a measuring and operating circuit to control this method.

Benefits of technology

Enables accurate determination of deposit indicators, allowing for timely maintenance and preventing incorrect results due to gas loading, thereby improving fouling detection and cleaning efficiency across various Coriolis mass flow meter configurations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure EP2025078973_23042026_PF_FP_ABST
    Figure EP2025078973_23042026_PF_FP_ABST
Patent Text Reader

Abstract

A method (100) is used to determine a quantitative deposit indicator (hrel) for a Coriolis mass flow sensor having at least one oscillatory measuring tube for conducting a medium, wherein the deposit indicator depends on an amount of deposit in the at least one measuring tube, wherein a device parameter (A) which is specific to the Coriolis mass flow meter is provided, wherein the device parameter (A) describes a relationship between a viscosity (h) of a medium conducted in the at least one measuring tube and a media-specific damping contribution (Dh) for damping an oscillating mode of the at least one measuring tube and in particular for determining a measured viscosity value (hm) on the basis of the media-specific damping contribution (Dh), wherein the deposit (C) comprises a deposit material with a material-specific damping property which, in dependence on the amount of deposit, brings about a damping of the oscillating mode of the at least one measuring tube, wherein a material parameter (M) which represents the material-specific damping property is provided, wherein the method comprises: determining a measured damping value (Dm) (110) of the at least one oscillating mode; determining a deposit-specific damping contribution (Dc) (120) on the basis of the measured damping value (Dm) of the at least one oscillating mode; providing the material parameter (M) specific to the deposit; and determining the deposit indicator (hrel) (139), on the basis of the deposit-specific damping contribution, the material parameter and the device parameter.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Method for determining a quantitative deposit indicator and Coriolis mass flow meter for carrying out the method

[0002] The present invention relates to a method for determining a quantitative deposit indicator and a Coriolis mass flow meter for carrying out the method.

[0003] Detecting deposits in the measuring tubes of Coriolis mass flow meters is important for plant operators because it may necessitate cleaning the measuring devices and, if applicable, the wetted components of the plant that communicate with them. Deposits in Coriolis mass flow meters have been addressed, among others, in DE 10 2017 101 923 A1, DE 10 2005 050 898 A1, DE 10 2020 132 949 A1, and DE 10 2022 134 589 A1. In these cases, an increase in the damping of a vibration mode is typically interpreted as an indication of deposit formation.

[0004] However, a model of the relationship between damping and fouling is currently lacking, making it impossible to describe fouling formation with a uniform model across a wide variety of fouling types and configurations of Coriolis mass flow meters. This would allow, for example, the application of findings on fouling formation to different configurations of Coriolis mass flow meters or to different fouling types. The object of the invention is therefore to remedy this lack.

[0005] The problem is solved according to the invention by the method according to independent claim 1 and the Coriolis mass flow meter according to independent claim 18.

[0006] The method according to the invention serves to determine a quantitative deposit indicator (hrei) for a Coriolis mass flow sensor with at least one oscillating measuring tube for guiding a medium, wherein the deposit indicator depends on the amount of a deposit in the at least one measuring tube, wherein a device parameter (A) is given which is specific for the Coriolis mass flow sensor, wherein the device parameter (A) describes a relationship between a viscosity (n) of a medium guided in the at least one measuring tube and a medium-specific damping contribution (Dr) for damping a vibration mode of the at least one measuring tube and in particular for determining a viscosity measurement value ( m) based on the media-specific damping contribution (Dr,), wherein the coating (C) has a coating material, wherein the coating material has a material-specific damping property, which, depending on the amount of coating, causes damping of the vibration mode of the at least one measuring tube, wherein a material parameter (M) is given that represents the material-specific damping property, wherein the method comprises: determining a damping measurement value (D m ) of at least one vibration mode; determining a surface-specific damping contribution (D c ), based on the damping measurement value (D m ) of at least one vibration mode; providing the material parameter (M) specific to the coating; and determining the coating indicator (h) re i) based on the surface-specific damping contribution, the material parameter and the device parameter.

[0007] In a further development of the invention, the device parameter (A) is viscosity-independent.

[0008] In a further development of the invention, the material parameter (M) is independent of the device.

[0009] In a further development of the invention, determining the coating-specific damping contribution (D) includes c ): Clean up damping measurement value (D m ) by a self-damping contribution (Do) of at least one measuring tube.

[0010] In a further development of the invention, determining the coating-specific damping contribution (D) includes c ): Clean up damping measurement value (D m ) for a media-specific damping contribution (Dr.).

[0011] In a further development of the invention, the correction of the damping measurement value (D) comprises m ) to determine the media-specific damping contribution (D ), the determination of the media-specific damping contribution (D^) based on a viscosity value (n) for the medium.

[0012] In a further development of the invention, the method further comprises: checking whether a gas loading of the medium is negligible; and discarding the damping measurement value if this is not the case.

[0013] The at least one vibration mode exhibits a frequency fluctuation that depends on the gas loading, and according to a further development of the invention, checking whether the gas loading is negligible comprises the following steps: determining a fluctuation indicator value (Fl) which represents the frequency fluctuation; comparing the fluctuation indicator value (Fl) with a fluctuation reference value (FR); and evaluating the gas loading as negligible if the fluctuation indicator value (Fl) is less than at least one fluctuation reference value (FR).

[0014] According to one embodiment of this further development of the invention, determining the fluctuation indicator value Fl comprises determining at least one amount of a fluctuation measurement NFF normalized to the third power of the frequency, i.e.: NFF = |(df / dt) / f 3 !

[0015] In particular, determining the fluctuation indicator value Fl involves filtering a data stream of normalized fluctuation measurements NFF with a low-pass filter to obtain a mean normalized fluctuation value MNF; wherein, in particular, the fluctuation indicator value Fl comprises the mean normalized fluctuation value MNF.

[0016] According to a further embodiment of this invention, determining the fluctuation indicator value Fl comprises: determining a data stream of fluctuation measurement magnitudes FF = |(df / dt)|; filtering the data stream of fluctuation measurement magnitudes with a low-pass filter to obtain a fluctuation mean value MFF; and normalizing the fluctuation mean value MFF by the cube of an associated frequency value to obtain a normalized fluctuation mean value NMF, i.e., NMF = MFF / f 3 ; in particular, the fluctuation indicator value Fl includes the normalized fluctuation mean NMF.

[0017] In one embodiment of this further development of the invention, the comparison with the fluctuation reference value can have a hysteresis function to avoid excessively frequent changes between the detection of a negligible gas load and a non-negligible gas load. For this purpose, if, according to the immediately preceding fluctuation indicator values, there is no or at most a negligible gas load, a comparison can be made with a first fluctuation reference value FR. If, on the other hand, according to the immediately preceding fluctuation indicator values, there is a non-negligible gas load, a comparison is made with a second fluctuation reference value that is lower than the first fluctuation reference value FR, wherein the second fluctuation reference value is, in particular, no more than 90% of the first fluctuation reference value FR.This means that the classification of the gas load as negligible is only determined after a previously determined non-negligible gas load has been significantly reduced.

[0018] In a further development of the invention, the material parameter represents a loss factor of the material, in particular for a frequency range of 10 A 2 Hz to 3 A 3 Hz. In one embodiment of this further development of the invention, the following applies for a ratio R := M / V between the loss factor V and the material parameter: 14 < R < 4, wherein the loss factor V applies in particular for a frequency range below a lowest natural frequency of vibrations of the coating at a temperature of 300 K or at an effective medium temperature at which the method is carried out.

[0019] In a further development of the invention, a literature value of the loss factor tan(8) for the material of the coating is used as a material parameter.

[0020] In a further development of the invention, a reference value of the loss factor tan(8) is used as a material parameter, wherein the reference value was determined on the basis of a reference measurement. In the reference measurement, damping values ​​of a vibration mode of at least one measuring tube of a Coriolis mass flow meter were measured as a function of the relative coating thickness h. rei determined, and a surface-specific damping contribution Dc of the surface material as a function of the relative ballast thickness h rei The damping of the vibration mode was determined, from which the reference value can be derived, for example according to:

[0021] 4'

[0022] M =

[0023] 1 - (1 - h rei

[0024] In a further development of the invention, one or more measured values ​​selected from a list comprising a media temperature measurement; a measuring pipe temperature measurement and a media density measurement are included in the determination of the self-damping contribution and / or media damping contribution and / or the coating indicator.

[0025] In a further development of the invention, the coating indicator comprises a relative coating thickness value hrei in relation to the inner radius of the measuring tube, wherein the relative coating thickness value hrei is related to a monotonically increasing function of the coating-specific damping contribution D. c , is calculated, in particular with a function of the type: where A is the device parameter; where M is the material parameter; and where, in particular, the surface-specific damping contribution D c , is calculated according to

[0026] DC — Dm Do - Dr,., or where the coating indicator is a function f(Dc ) is calculated, which determines the behavior of the function h re i(D c ) at least over an initial fit range between h rei = 0 and an upper control value of h rei = 0.2, in particular at least over a second fit range between hrei = 0 and an upper control value of h rei = 0.3 approximated, so that for a coefficient of determination R 2 a fit of the function f(D c ) to the function h re i(D c ) Regarding the second fitness area: 1 - R 2 < 1.5%, for example 1 - R 2 < 0.5%, especially 1 - R 2 < 0.25%, or where R is the coefficient of determination 2 a fit of the function f(D c ) to the function h re i(D c ) Regarding the first fitness area: 1 - R 2 < 1.0%, for example 1 - R 2 < 0.3%, especially 1 - R 2 < 0.1%. Suitable function types for f(D cExamples include square root functions, linear functions, or quadratic functions.

[0027] In a further development of the invention, the coating indicator comprises a linear function of the relative coating thickness value, wherein the coating indicator is in particular proportional to the relative coating thickness value.

[0028] In a further development of the invention, the method further comprises: comparing the coating indicator with at least one limit value; and outputting a warning signal if the comparison detects an exceedance of the limit value.

[0029] In a further development of the invention, the method further comprises: initiating a maintenance measure in response to the warning signal, wherein the maintenance measure in particular is a cleaning of the at least one measuring tube.

[0030] The Coriolis mass flow meter according to the invention comprises: a Coriolis mass flow meter with at least one oscillating measuring tube for guiding a medium; and a measuring and operating circuit which is configured to control the method according to one of the preceding claims.

[0031] The invention will now be explained in more detail with reference to the exemplary embodiments shown in the drawings. These show:

[0032] Fig. 1a: A diagram showing damping measurements as a function of the relative

[0033] Coverage thickness h rei represents various types of Coriolis mass flow meters.

[0034] Fig. 1b: A diagram showing surface indicators, their values ​​as a function of the

[0035] Damping measurements from Fig. 1a are shown above, along with the relative thicknesses of the coverings corresponding to the damping measurements.

[0036] Fig. 2a: A schematic cross-section of a measuring tube with a coating.

[0037] Fig. 2b: A schematic cross-section of a measuring tube under the influence of a viscous substance

[0038] Mediums. Fig. 3: A flowchart of an embodiment of the invention.

[0039] Procedure.

[0040] Fig. 4: A schematic representation of an embodiment of a mass flow meter according to the invention.

[0041] Fig. 5a: A diagram showing the course of frequency fluctuations in a

[0042] schematically depicts an episode of gas loadings.

[0043] Fig. 5b: A diagram showing the course of damping measurements during the

[0044] The episode of gas loadings is schematically represented, for which the frequency fluctuations are shown in Fig. 5a.

[0045] Fig. 1a shows the results of an experiment in which the measuring tubes of various Coriolis mass flow meters from the applicant's portfolio were coated with different relative thicknesses of a wax-like coating. The corresponding damping measurements of the bending vibration mode, or F1 mode, were recorded. The data marked with an asterisk are from a Coriolis mass flow meter designated Promass F 50. The data marked with an asterisk are from a Coriolis mass flow meter designated Promass F 25. The data marked with a circle are from a Coriolis mass flow meter designated Promass Q 50. In this experiment, the measuring tubes were filled with air to observe the effect of the coating without the influence of a damping medium.In principle, the method according to the invention can also be operated with measuring tubes filled with liquid and viscous media.

[0046] The diagram in Fig. 1a shows that the same relative coating thickness has very different effects on the damping of the bending vibration mode for the different configurations of the Coriolis mass flow meters. Therefore, it is not possible to directly infer the relative coating thickness from an observed damping measurement using a uniform model. The method according to the invention provides a remedy for this, as shown in Fig. 1b. The coating indicators C(D) shown here m ) were determined using the inventive method as a function of the damping measurement values ​​D m determined from Fig. 1a and over the causative relative coating thickness h reiThe data is plotted. A clear linear or proportional relationship is evident, which now allows for a simple interpretation of the damping measurements. For an explanation of the modeling according to the invention, reference is made to Figures 2a and 2b. The modeling of the damping of a vibration mode of a measuring tube 210 by a coating 212 with relative coating thickness h rei The invention is based on an analogy to viscous damping by a medium 214, whose damping effect can be described by the relative thickness e of the Stokes layer. The analogy is helpful insofar as device-specific parameters, which are required to describe the relationship between vibration damping and the viscosity of a medium, are established for various types of Coriolis mass flow meters and, based on the above analogy, can now be adopted for the characterization of deposits.

[0047] Assuming a coating with relative coating thickness h rei and with a surface-specific loss factor tan(8) a surface-specific damping contribution D c The relative thickness of the covering, which is proportional to the area moment of inertia of the covering, can be expressed as: where k' is initially a factor yet to be defined, which incorporates device-specific properties.

[0048] As a general rule:

[0049] 1 k' h rpi ~ - — D, r eL 4 tan 5 '

[0050] The relative thickness of the Stokes layer, given by the Stokes number St, is given by: where Dri is a viscosity-specific damping contribution, and where Ai describes a device-specific factor in the relationship between Stokes number and viscosity-specific damping, where the factor is further divided by the square of the natural frequency f of the vibration mode, which depends on the medium density, and the density p of the damping medium. In practice, it has been found that the medium density remains constant at 1000 kg / m³. 3 can be applied and f is the device-specific associated natural frequency.

[0051] By comparing the last two equations, the following can be deduced by analogy:

[0052] An embodiment 100 of the method according to the invention will now be explained with reference to Fig. 3.

[0053] The procedure 100 begins with the acquisition 110 of a damping measurement value D m .

[0054] The next step is determining 120 of a surface-specific damping contribution D.c based on the damping measurement value D m This includes correcting the damping measurement value D (122). m to achieve a self-damping contribution D o of the measuring tube, as well as the correction 124 of the attenuation measurement value D m to determine a viscosity-specific damping contribution D, such that the surface-specific damping contribution D c is given as:

[0055] De — Dm — Do ■ Dr,.;

[0056] The determination of the relative coating thickness follows, according to section 130. where the device parameter A is given by: A = k', and where the loss factor tan(8) is used for the material parameter M. The loss factor describes the ratio between the imaginary and real parts of the complex modulus of elasticity of the lining material, where the imaginary part indicates the damping influence of the lining material. If available, the loss factors can be obtained from material data literature, or they can be experimentally determined for the lining types of interest and then made available in a database. To experimentally determine a loss factor, essentially a lining-specific damping contribution is determined for a given relative lining thickness hrei, from which the material parameter M or the loss factor can then be determined according to the following equation:

[0057] If necessary, the coating indicator can be further normalized (140), for example, such that the maximum permissible relative coating thickness results in a coating indicator with the value 1. Finally, a limit value comparison (150) can follow, whereby if the limit value is exceeded by the coating indicator, a maintenance requirement, in particular a cleaning requirement, is signaled.

[0058] The embodiment of a Coriolis mass flow meter 200 according to the invention, shown in Fig. 4, comprises a sensor 202 inside which two measuring tubes 212 run parallel to each other and are mounted to oscillate relative to each other. The Coriolis mass flow meter 200 further comprises an electronics housing 204 with a measuring and operating circuit 206 for driving a vibration exciter to excite the measuring tube oscillations, for acquiring vibration sensor signals, and for carrying out the method according to the invention. The damping values ​​are determined, in particular, on the basis of an excitation current for exciting the measuring tube oscillations in relation to the achieved vibration amplitude.

[0059] The diagrams shown in Fig. 5a and 5b relate to the verification of whether the gas loading of a liquid medium is negligible. A gas loading influences at least one vibration mode in two ways: firstly, through a considerable gas-specific damping contribution, and secondly, through a frequency fluctuation. The gas-specific damping contribution D g , in order to determine the damping measurement value D m The amount that would need to be corrected cannot be readily determined quantitatively during ongoing measurements. Therefore, determining the surface-specific damping contribution D is not possible. cThis is not practical when gas loading is present. However, frequency fluctuations make it possible to detect gas loading and then suspend the determination of a deposit indicator for the duration of a non-negligible gas loading. An example of the implementation of gas loading detection is now explained using Fig. 5a, which shows a fluctuation indicator value Fl for at least one oscillation mode over an episode of gas loading with a gas volume fraction of approximately 0.1% to 0.2% relative to a base of 0%. The fluctuation indicator values ​​Fl are, on the one hand, unfiltered fluctuation measurements represented by individual points with normalized NFF = 1 / f. 3■ |df / dt| of the frequency of at least one vibration mode, and on the other hand, mean normalized fluctuation values ​​MNF determined on the basis of the NFF, for example with a low-pass filter, which are shown as solid lines. The fluctuation indicator values ​​Fl are compared with a fluctuation reference value FR, which is shown as a dashed line in the diagram. Exceeding the fluctuation reference value FR is interpreted as an indicator of the presence of a non-negligible amount of gas bubbles, where determining the deposit indicator based on damping measurements would lead to incorrect results. Therefore, in this case, the damping measurements D m to discard what is shown in Fig. 5b by setting D m The value is indicated as zero. The decision to discard the damping measurements D represented by the solid line is made here. mThe value is determined based on the mean normalized fluctuation values ​​MNF from Fig. 5a. However, if the normalized fluctuation measurements NFF occasionally fall below the fluctuation reference value FR at low gas loading, this can falsely suggest situations of negligible gas loading. The subsequent use of the damping measurements D m Determining a surface coating indicator would lead to significantly inflated results, as can be seen from the data points of the damping measurements in Fig. 5b, for which the decision to reject was made based on the unfiltered, normalized fluctuation measurements NFF from Fig. 5a. Filtering reliably suppresses this source of error when determining surface coating indicators.

Claims

Patent claims 1. Method (100) for determining a quantitative coating indicator (h re i) for a Coriolis mass flow sensor with at least one vibrating measuring tube for guiding a medium, wherein the coating indicator depends on an amount of a coating in the at least one measuring tube, wherein a device parameter (A) is given which is specific for the Coriolis mass flow sensor, wherein the device parameter (A) describes a relationship between a viscosity (r|) of a medium guided in the at least one measuring tube and a medium-specific damping contribution (Dr,) for damping a vibration mode of the at least one measuring tube and in particular for determining a viscosity measurement value (r| m) based on the media-specific damping contribution (Dr,) serves, wherein the coating (C) has a coating material, wherein the coating material has a material-specific damping property which, depending on the amount of coating, causes damping of the vibration mode of the at least one measuring tube, wherein a material parameter (M) is given that represents the material-specific damping property, wherein the method comprises: Determining a damping measurement value (D) m ) (110) of at least one mode of vibration; Determining a surface-specific damping contribution (D) c ) (120) based on the damping measurement value (D m ) of at least one vibration mode; Providing the material parameter (M) specific to the coating; and Determining the coating indicator (h re i) (139), based on the surface-specific damping contribution, the material parameter and the device parameter.

2. Method according to claim 1, wherein the device parameter (A) is viscosity-independent.

3. Method according to claim 1 or 2, wherein the material parameter (M) is independent of the device.

4. Method according to one of the preceding claims, wherein the determination of the surface-specific damping contribution (D) c ) (120) includes: Clean (122) attenuation measurement (D m ) to provide a self-damping contribution (D o ) of at least one measuring tube.

5. Method according to one of the preceding claims, wherein the determination of the surface-specific damping contribution (D) c ) includes: Clean (124) attenuation measurement (D m ) to achieve a media-specific damping contribution (Dr,)- 6. Method according to claim 5, wherein the clearing (124) of the damping measurement value (D m) around the media-specific damping contribution (D ), the determination of the media-specific damping contribution (Dr,) based on a viscosity value (n) for the medium includes.

7. Method according to any of the preceding claims, further comprising: Check whether the gas loading of the medium is negligible; and Discard the damping measurement if this is not the case.

8. Method according to one of the preceding claims, wherein the material parameter represents a loss factor of the material, in particular for a frequency range of 10 A 2 Hz to 3 A 3 Hz.

9. Method according to one of the preceding claims, wherein furthermore one or more measured values ​​selected from a list comprising a media temperature measurement; a measuring pipe temperature measurement and a media density measurement are included in the determination of the self-damping contribution and / or media damping contribution and / or the coating indicator.

10. Method according to any of the preceding claims, wherein the coating indicator has a relative coating thickness value h rei in relation to the inner radius of the measuring tube, where the relative coating thickness value h rei with a monotonically increasing function of the surface-specific damping contribution D c , is calculated, in particular with a function of the type: where A is the device parameter; where M is the material parameter; and where, in particular, the damping contribution D c , calculated according to D c = D m - D o - D.

11. Method according to one of the preceding claims, wherein the coating indicator comprises a linear function of the relative coating strength value, wherein the coating indicator is in particular proportional to the relative coating strength value.

12. Method according to any one of the preceding claims, further comprising: Comparing the surface area indicator with at least one limit value; and Output of a warning signal if the comparison detects that the limit has been exceeded.

13. The method of claim 11, further comprising: Initiating a maintenance measure in response to the warning signal, wherein the maintenance measure includes in particular cleaning of the at least one measuring tube.

14. The method of claim 7, or of a claim dependent on claim 7, wherein the vibration mode has a frequency fluctuation dependent on the gas loading, and wherein checking whether the gas loading is negligible comprises: Determining a fluctuation indicator value (Fl) which represents the frequency fluctuation; Comparing the fluctuation indicator value (Fl) with a fluctuation reference value (FR), and The gas loading is considered negligible if the fluctuation indicator value (Fl) is less than at least one fluctuation reference value (FR).

15. The method of claim 14, wherein determining the fluctuation indicator value Fl comprises determining at least one amount of a fluctuation measurement NFF normalized to the third power of the frequency, i.e.: NFF = |(df / dt) / P|, 16. Method according to claim 15, wherein determining the fluctuation indicator value Fl comprises filtering a data stream of normalized fluctuation measurements NFF with a low-pass filter to obtain a mean normalized fluctuation value MNF; wherein in particular the fluctuation indicator value Fl comprises the mean normalized fluctuation value MNF.

17. Method according to claim 14 where determining the fluctuation indicator value Fl includes: Determining a data stream of fluctuation measurement values ​​FF = |(df / dt)|; Filtering the data stream of fluctuation measurement values ​​with a low-pass filter to obtain a fluctuation mean value (MFF); and normalizing the fluctuation mean value (MFF) to the cube of an associated Frequency value to obtain a normalized fluctuation mean NMF, i.e., NMF = MFF / f 3; in particular, the fluctuation indicator value Fl includes the normalized fluctuation mean NMF.

18. Coriolis mass flow meter, comprising: a Coriolis mass flow sensor with at least one oscillating measuring tube for guiding a medium; and a measuring and operating circuit configured to control the method according to any of the preceding claims.

Citation Information

Patent Citations

  • in-line measuring device

    DE102005050898A1

  • exhaust system

    DE102017101923A1

  • Method for identifying deposits in a process plant using a Coriolis mass flow sensor

    DE102020132949A1

  • Method for determining at least one coating property on a wall of a measuring tube

    DE102022134589A1

  • Method to monitor and control processes with on-line quartz crystal microbalance sensors

    EP0829010B1