DE4023663C2 - Method for diagnosing the mechanical properties of a machine that has rotating components - Google Patents

Description

The invention relates to a method for diagnosing the mechanical properties of machines according to the generic term of claim 1. Such a method is known from the US Patent 43 80 172 is known, in which is indicated To temporarily change the operating status of the machine and the resulting change in frequency spectrum of machine vibrations as an indicator of egg to use any damage.

The invention has for its object in the preamble of the method specified in claim 1 so that automated monitoring in a simple manner using a parameter is possible.

This task is with the measure specified in claim 1 men solved. Before  partial further developments are indicated in the subclaim give.

The 1st harmonic is always the rotational frequency vibration or with gear drives the meshing frequency.

The invention is explained with reference to FIGS. 1 to 4, wherein

Fig. 1 is a basic block diagram of an arrangement for the detection and analysis of machine vibration,

Fig. 2 and 4, amplitude spectra of the machine vibration and

Fig. 3 represent a flow chart for evaluating the machine vibrations.

According to the exemplary embodiment for evaluating machine vibrations according to FIG. 1, structure-borne noise signals are measured on a machine housing G and processed in an evaluation unit A. The analog structure-borne sound signal is pre-filtered and digitized in a transient recorder TR of the evaluation unit A, a signal processor SP carrying out a further band limitation of the frequency band and calculating the frequency function (for example by Fast Fourier Transformation (FFT)). The individual harmonic frequencies are selected on the spectral level. The 1st / 2nd, 1st, 2nd, 3rd. Harmonics filtered out and processed according to the characteristic size equation. As can be seen from FIG. 2, the relation of the change in the increments of the i-th (i = 0.5; 2; 3) harmonic frequencies to the changes in the 1st harmonic during the T-follow-up measurements according to the following characteristic equation for the characteristic variable R _{i is} determined:

The quantities A _i here represent the amplitudes of the harmonics of the fundamental frequency (i = 1, corresponds to A ₁ ) at times t (for example t = 7 for 7 days of a week). The parameter R _i gives the relation of the respective amplitudes of the various harmonics (i = 0.5; 2; 3) to the fundamental frequency (i = 1) over different measurement times. . According to the Figure 2 are:

A _it . . . i-th harmonic (last measurement t)
A _{i (t-1)} . . . i-th harmonic (penultimate measurement (t-1))

ΔA _0.5 = A _0.5t - A _{0.5 (t-1)} increase (change in the 0.5th harmonic
ΔA₁ = A _1t - A _{1 (t-1)} increase (change) in the 1st harmonic
ΔA₂ = A _2t - A _{2 (t-1)} increase (change) in the 2nd harmonic
ΔA₃ = A _3t - A _{3 (t-1)} increase (change) in the 3rd harmonic
ΔA _i = A _it - A _{i (t-1)} ΔA _i . . . Increase in the i-th harmonic (i = 0.5; 1; 2; 3)

It is thus possible to determine which harmonic has grown faster in a given period of time, in order to be able to discover impending damage (e.g. shaft break or gear wear) in the early stages. In Fig. 3, a flow chart of a computer program for performing the parameter determination is shown, the application of which is shown in three cases, namely the wave crack detection, gear wear detection and the detection of damage to blades, for example a turbine.

The calculated parameters R _i are observed as a function of time, ie in the trend. Process parameters that are not discussed here are also included in the final statement.

Application: Damage to guide or rotor blades, for example a turbine

The signal analysis carried out in the evaluation unit A according to FIG. 1 of the vibrations measured on a turbine consists of the following stages:

• 1) Calculation of the rotating blade sound according to the basic frequency f ₁ , multiplied by the number of moving blades, and of the rotating vane sound according to the basic frequency f ₁ , multiplied by the number of the guide blades.
• 2) The relationship of the change in the rotating blade sound with the amplitude A ₂ or the rotating vane sound with the amplitude A ₃ to the changes in the rotational frequency amplitude A ₁ during the T-follow-up measurements according to the parameter equation (1) is examined: whereA _it = amplitude of the blade rotation sound (i = 2) (rotational frequency × blades) or the guide blade rotation sound (i = 3) at the time of measurement t),
  A₁ = rotational frequency amplitude,
  T = represent the following measurement (following time periods, e.g. days).
  If the relation R _i <1, then the i-th amplitude A _i grows faster than the rotational frequency amplitude A ₁ and therefore makes it possible to make a statement about damage to the guide and / or moving blades.
• 3) In addition, when analyzing the signal on blades, for example a turbine, the phase of the blade's rotary sound is also examined in order to obtain reliable information about damage. The change in phase ϕ ₂ of the rotating blade or the phase of the rotating blade during the follow-up measurements to the first measurement is investigated according to the following equation: where here represents _i = phase of the rotating blade sound (at i = 2) or phase of the rotating blade sound (i = 3).

If the relation F _i <1, then there is a phase increase and thus a turbine damage.

The analysis of the changes in the parameters R _i and F _i facilitates troubleshooting for guide vanes and rotor blades in turbomachinery.

Industrial applicability

The invention is particularly useful in automated surveillance of machines with rotating parts, for example in force works, applicable.

Claims (3)

1. A method for diagnosing the mechanical properties of a machine having rotating components, in which the vibrations caused by these are detected as vibration signals on the machine in order to detect damage and / or wear on the rotating components, which are pre-filtered and digitized and after a frequency transformation in the frequency domain can be evaluated on the basis of the amplitudes of the fundamental frequency and of harmonics, characterized in that the pre-filtering and digitization are carried out in a transient recorder and that a first parameter R _{i is determined} using the following equation: where A _{1t is} the amplitude of the fundamental frequency and
A _{it is} the amplitude of the i-th (0.5th, second, third) harmonics in the measurement t,
A _{1 (t-1)} and
A _{i (t-1) are} the amplitudes of the fundamental frequency and the i-th harmonic in the measurement (t-1) and
T is the number of measurements following the first measurement. 2. The method according to claim 1, characterized in that for evaluating the detected vibration signals on blades, in particular on turbines, a second parameter F _{i is} additionally determined according to the following equation: whereinϕ _it and
ϕ _{i (t-1)} the phase of the rotating vane sound (if i = 2) or the phase of the vane rotating sound (i = 3) for measurements t and t-1
represent. 3. The method according to claim 1 or 2, characterized records that structure-borne noise signals as vibration signals be recorded.