EP1877739A2

EP1877739A2 - Method and system for diagnosing mechanical, electromechanical or fluidic components

Info

Publication number: EP1877739A2
Application number: EP06754960A
Authority: EP
Inventors: Alf PÜTTMER
Original assignee: Siemens AG
Current assignee: Siemens AG
Priority date: 2005-05-04
Filing date: 2006-05-02
Publication date: 2008-01-16
Also published as: WO2006117375A3; WO2006117375A2; DE102005020900B3; CN101171499A; CN101171499B; US20090090185A1; US7703326B2

Abstract

The invention relates to a diagnosis method and system comprising a sensor (1) for structure-borne noise which emits a measuring signal (3) in a bandwidth-filtered manner due to its mechanical resonant frequency, its capacity (C0) and an inductance (L1). The system also comprises an evaluation device (4) in which the measuring signal (3) is subsampled and an error message signal (9) is generated if the intensity of the subsampled measuring signal exceeds a defined threshold value. The inventive system requires only a small number of electronic components and only little power supply capacity, thereby allowing diagnosis to be carried out without complication in existing devices, especially in field devices of process instrumentation for which only a limited amount of operating power is available. The inventive method and system are especially advantageous for the recognition of valve leakage with an electropneumatic position controller.

Description

description

Method and system for the diagnosis of mechanical, electromechanical or fluidic components

The invention relates to a method for the diagnosis of mechanical, electromechanical or fluidic components, in particular of a valve which can be actuated by a positioner via a drive, according to the preamble of claim 1 and a system for diagnosis of such compo ^¬ nents according to the preamble of the claim 4th

By analyzing structure-borne noise, features can be obtained that contribute to the detection of faults or faults in mechanical, electromechanical or fluidic components. For example, from EP 1216375 Bl, a diagnostic system for a actuated by a position controller via a drive valve is known in which the Intensi ^¬ ty of the structure-borne sound signal in a spectral range above 50 kHz for detecting a leakage in the valve is used. However, the known diagnostic system requires a spec ^¬ tral analysis of the measurement signal and thus a considerable on ^¬ wall of electronics and computing power. Associated with this is an increased electrical power consumption of the evaluation device. An integration of the diagnostic method As an added ^¬ Liche function in existing equipment is rarely possible because the extra power is often not, has become available ^¬ supply. This is divided in particular for field devices of automation technology set ^¬ such as transmitters or manipulated, the case. These often have to meet the requirements of explosion protection or are supplied with the required auxiliary power via a 4 to 20 mA interface or a PROFIBUS interface. The diagnostic system then has to be laboriously accommodated in an additional device. From US-PS 5 477 729 a transducer for structure-borne noise is known, which is suitable for measuring high-frequency acoustic signals up to about 2 MHz.

The invention has for its object to provide a method and a system for the diagnosis of mechanical, electro-mechanical or fluidic components, which are characterized by a low cost.

To solve this problem, the new method of the type mentioned in the characterizing part of claim 1 features. In claim 4 is a system for carrying out the method, in the dependent claims, developments of the invention are described.

The invention has the advantage that a much lower switching and energy expenditure for implementing the procedural ^¬'s proceedings required than previously. This advantage is achieved that are combined to be implemented functions and greatly simplified by playing a component simultaneously fulfills several functions at ^¬. Thus, a sensor for structure-borne sound in addition to the egg ^¬ tual conversion of the sound signal into an electrical signal at the same time the function of a Bandpassfilte- tion. This is achieved in a simple manner, be that the mechanical resonance frequency and the capacitance of the transducer and an inductance in a suitable way to each other abge ^¬ true. Thus, a measurement signal is already generated by the transducer for structure-borne noise, which has predominantly signal components in the frequency range relevant for the respective application. Further filter elements are therefore not mandatory. For further processing, a subsampling of the measurement signal is performed. As a result, in the area of the analysis for feature extraction, a minimization of the required computing power takes place. Likewise, sub-sampling causes a reduction in the speed requirements imposed on the analog-to-digital conversion. A Sub-scan means that the measuring signal is detected with a clotting ^¬ Geren sampling rate for the digital wide development than the Messsig- would be required according to the known Nyquist-Shannon sampling theorem for the analysis of the frequency of interest shares Nals. In order to minimize the computing power, sub-sampling is preferably performed at a sampling rate which is orders of magnitude lower. This measure is based on the finding that the sound intensity in the frequency range of interest can also be determined with a strong undersampling because of the bandpass filtering and that thus can be dispensed with a complex Fast Fourier transform for analyzing and viewing the frequency range of interest. Since only signal components in the frequency range of interest are present in the bandpass-filtered measurement signal, the intensity determined with a sub-sampling corresponds to that of the measurement signal in the frequency range of interest.

The intensity of the subsampled measurement signal is now compared in a simple manner with a threshold to a

To obtain a statement as to whether an error state of the mechanical, electro-mechanical or fluidic component is present or not. This threshold can be predetermined, for example, by manual input or by a prior measurement and analysis in a good condition.

Due to the few required electronic components, the low required computing power and electrical power is now an integration of the diagnosis with body sound measurement and signal analysis in existing devices, such as sensors or actuators of Prozessinstrumentie ^¬ tion, in particular an integration in a control valve that with a positioner can be actuated via a drive, possible. This integration is particularly simple if a microcontroller is already present in the device which, in addition to its previous tasks, readily requires the calculations necessary for evaluating the measurement signal can take over. The new diagnostic method and system is therefore characterized by a particularly low cost, which must be applied to carry out the diagnosis.

A particularly low computational complexity, and associated therewith, a particularly low energy requirement of the calculations exporting microcontroller is achieved when the gleiten- de average of the amounts of the individual samples is true as a parameter of the intensity of sub-sampled measurement signal be ^¬ and is compared with the predetermined threshold value.

Prior to subsampling, signal amplification of the signal produced by the structure-borne sound pickup may be made with additional bandpass filtering in an electronic circuit having only one operational amplifier. This has the advantage that a better selection of the signal components in the frequency range of interest is made possible without having to substantially increase the energy requirement for the diagnosis. The optional operational amplifier is used on the one hand for signal amplification and on the other hand through its circuitry for bandpass filtering. Thus, the number of additional electronic components is reduced to a minimum. In an improved way, the bandpass filtering leaves only the signal components that are in the

Related to the phenomenon to be detected.

In a particularly advantageous manner, the new diagnostic method and system for leakage detection in control valves applicable bar, since this is a frequency range of the measurement signal of interest, which is above 50 kHz and reflects the strength of Kavi ^¬ tion noise. For a more detailed Erläute ^¬ tion of an arrangement for valve diagnosis using acoustic emission analysis ^¬ and the related advantages will be made to the already initially mentioned EP 1216375 Bl. With reference to the drawings, in which an embodiment of the invention is shown, the invention and refinements and advantages are explained in more detail below. Show it:

FIG. 1 is a block diagram of a diagnostic system;

2 shows a diagnostic system with optional active bandpass filter and

FIG. 3 is a timing diagram of a bandpass filtered measurement signal.

A transducer 1 for structure-borne noise has, according to FIG. 1, a piezoceramic 2 which is provided with electrodes. In the electric equivalent circuit diagram of the piezoelectric ceramic 2 has a capacity of CO which, if can be modified by additional factors Kondensa ^¬ required. To the piezoceramic 2, an inductance Ll is connected in parallel, which, as indicated in Figure 1 with broken lines, is integrated in the transducer 1. Alternatively, the inductance can be realized as getrenn ^¬ tes component. In order for the transducer 1 to emit a measuring signal 3 which essentially contains only signal components in a frequency range of interest, the mechanical resonance frequency of the piezoceramic 2, the capacitance CO and the inductance L 1 are suitably matched to one another. The measuring signal 3 is fed to an analog input ADC IN of a microcontroller 4. The microcontroller 4 forms an evaluation device in which a subparagraph sampling of the measuring signal 3 made and a Fehlermeldesig ^¬ is generated NAL 9, when the intensity of sub-sampled measurement signal 3 exceeds a predetermined threshold. The intensity is determined by calculating a simple characteristic K corresponding to the average of the magnitudes of the individual samples. The formula is: With

Ν - number of accumulated samples i = 1 ... Ν - summing index and X ₁ - ith sample, where i = 1 corresponds to the last sample, i = 2 to the penultimate sample, and so on.

The diagnostic system according to FIG. 1 is integrated in an electropneumatic positioner for a valve which can be actuated by a pneumatic drive. The microcontroller 4 is the microcontroller already present in the positioner. It is particularly clear that the new diagnostic system can be supplemented with very little effort in an existing positioner. In principle, only the transducer 1 for structure-borne noise and an analog input of the microcontroller 4 is required. The required changes in the program of the microcontroller 4 for carrying out the diagnostic method are ^¬ due to the simple calculations of comparatively small scale. The calculations require only a small part of the existing computing power of the microcontroller 4.

The gain of the measuring signal 3 is carried out by utilizing the resonance rise of the resonance circuit, which is det by the capacity CO of the piezoelectric ceramic and the inductor Ll gebil ^¬.

FIG. 2 shows a diagnostic system which is expanded by an optional amplifier circuit 5. The already explained with reference to Figure 1 circuit parts are provided in Figure 2 with the same reference numerals. The additional amplifier circuit 5 consists in its core of an operational amplifier 6, the supply ^terminals are connected to a positive power supply voltage VCC or to ground. On the reference input of the operational amplifier 6, the half versor ^¬ supply voltage VCC / 2 is performed. In the input path of the surgical Amplifier 6 is a series circuit of a resistor Rl and a capacitor Cl arranged. In the feedback branch is a parallel circuit of a resistor R2 and a capacitor C2. With a correct tuning of the operational amplifier is obtained a band-pass effect in a simple manner with this circuit 6, which is precisely the signal ^¬ shares amplified in the frequency range of interest. The optional operational amplifier 6 is thus used on the one hand to amplify the measurement signal 3 and on the other hand through its circuitry for bandpass filtering. Thus, the number of electronic components is reduced to a minimum.

In the described use of the diagnostic system for detecting valve leakage, the structure-borne sound pickup 1 and the electronic circuit 5 are specifically optimized for high sensitivity to flow noise while insensitivity to the working noise of pumps or similar adjacent components. The mounting of the pickup 1 is carried out continuously at a preparatory ^¬ ended smooth outside surface be on the valve housing with a Schrau ^¬. For a reliable acoustic coupling a tem ^¬ peraturbeständiges coupling grease between the valve body and Sen ^¬ sor provides. Alternatively, the attachment can be made with good acoustic coupling to the housing of the positioner. In addition to the transducer 1 for structure-borne noise, no additional sensors are necessary. An adaptation of the evaluation to changing load ^¬ conditions such as pressure and stroke rate can be done automatically, without any parameters must be set or that a calibration would be necessary to a Gutzustand. The alarm thresholds can also be set manually by a user.

FIG. 3 serves to illustrate the principle of undersampling, which leads to a considerable reduction of the power requirement associated with the evaluation. Shown is a curve 7 of the bandpass-filtered measuring signal with 100 ab- probes obtained at a sampling rate which is conventionally adapted to the frequency range of interest. On the abscissa the number of the sample, the so-called sample, is plotted on the ordinate its amplitude. In a sub-sample having a sample rate that is one order of magnitude lower, every tenth sample is included in the further processing. The samples of the sub-sampling taken into account in this case are marked by dots in the course of the measurement signal, for example the sample value 8. The calculated mean value, which is compared with a predetermined threshold value, is shown as a horizontal line 10 in FIG. It will be appreciated that in moving averaging of the amounts of the samples, whether a sub-sample is made or not, a similar characteristic is calculated as the average. An essential requirement of this effec ^¬ TES is that the signal components of the measurement signal are limited by band-pass filtering in the frequency range of interest. It can thus be dispensed with a complex Fast Fourier analysis for viewing the frequency range of interest. Can be used, this leads to a materiality ^¬ significant reduction of the necessary for the implementation of the diagnostic method computation complexity and thus the Strombe ^¬ needs for supplies, so that the new method of diagnosis even in field devices th, where only a limited amount is available for use with auxiliary energy.

Claims

claims

1. A method for diagnosing mechanical, electromechanical see or fluidic components, in particular of a

Valve which can be actuated with a positioner via a drive, with a transducer (1) for structure-borne noise which is comparatively insensitive in the area of low-frequency operating noise, but sensitive in the higher-frequency range of the fault noise, and with a device (4) for evaluating the recorded measurement signal ( 3), characterized in that the transducer for structure-borne noise, the measurement signal (3) due to its mechanical resonance frequency, its capacitance (CO) and an inductance (Ll) outputs bandpass filtered that in the evaluation (4) Unterabtas ^¬ tion of the measuring signal (3 ) is performed and that a Feh ^¬ lermeldesignal (9) is generated when the intensity of the subsampled measurement signal exceeds a predetermined threshold.

2. The method according to claim 1, characterized in that as a parameter for the intensity of the measurement signal (3) of the moving average from the amounts of the individual sample values ^{¬ is} determined.

3. The method according to claim 1 or 2, characterized in that before subsampling a signal amplification with additional bandpass filtering in an electronic circuit with only one operational amplifier (6) is made.

4. System for the diagnosis of mechanical, electromechanical or fluidic components, in particular of a valve which can be actuated by a positioner via a drive, with a transducer (1) for structure-borne noise, the comparatively unempfind ^¬ lich in the field of low-frequency operating noise, in the higher frequency range the error sounds but is sensitive, and with a device (4) for evaluation the recorded measurement signal (3), characterized in that the transducer (1) for structure-borne noise is designed such that it outputs the measurement signal (3) bandpass filtered due to its mechanical resonance frequency, its capacitance (CO) and an inductance (Ll), that the Auswerteein ^¬ direction to carry out a sub-scan of the measurement signal and is provided for generating an error message signal (9) when the intensity of the sub-sampled measurement signal exceeds a predetermined threshold.