JPH0138251B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for diagnosing rotating machines, particularly those used in large-scale plants such as LNG plants that are highly public.

Rotating machines account for a very high proportion of large plants and equipment such as LNG plants, and their role is also important. When abnormalities or breakdowns occur in these rotating machines, it is natural that there will be economic losses, but if the accidents escalate, it will develop into a major social problem. For example, LNG plants are used as a supply source for city gas, cold power generation equipment, etc., so LNG is
If a failure occurs in the LNG pump that pumps LNG, it will cause a large loss to the user and also cause damage to general users.

Mechanical diagnostic technology in Japan began with the opening of the Shinkansen,
Although serious efforts began in the 1950s and 1960s, when the problem of pollution and contamination began to receive attention, the current situation is that the criteria for making such decisions have not yet been generalized.

The present invention was made against this background.
The objective is to establish failure prediction technology for rotating machinery in plants, improve plant reliability, reduce maintenance costs, and homogenize maintenance techniques.

Embodiments of the present invention will be described below with reference to the drawings.

To carry out the machine diagnosis method of the present invention, an acceleration detector is installed at an appropriate position on a rotating machine, and based on the signal from the detector, a microcomputer system or the like is used to detect the acceleration in a predetermined frequency range at a certain time. A vibration spectrum is created, and the spectrum is stored and displayed on a required display device such as a television monitor.

If we add the subscript n to the nth spectrum in a given frequency domain, the nth acceleration spectrum is α(n), the nth velocity spectrum is υ(n), and the nth amplitude spectrum is x(n). The n-th angular velocity is ω(n), the n-th frequency is (n), the spectrum width is Δ, the sampling frequency is _s , the number of sampled data is N _s , and the relationship between the above spectra is as follows. Become.

υ(n)=α(n)/ω(n) x(n)=υ(n)/ω(n)=α(n)/ω(
n) ² ω (n) = 2π (n) (n) = n・Δ Δ = ( _s / 2) / (N _s / 2) The following is a relatively low frequency region where the speed of rotating machinery is about 30000 rpm = 500 rps A case will be explained in which failure diagnosis in a vehicle is performed using a vibration spectrum.

Possible causes of failure that can be diagnosed in a relatively low frequency range include bent shafts, rubbing and cracking of the pump impeller, and eccentricity of the rotating body.

In the frequency domain, only the fundamental wave component will be analyzed, and the number of sampling data will be 1024.
If the full width of the spectrum is 500 Hz and the sampling frequency is 1 KHz, the total number of spectra will be 512 (n = 1 to 512), and the spectral width (1000/2)/(1024/2) ≒ 0.98 Hz. .

FIG. 1 shows the distribution of the vibration spectrum (x) of the rotating machine obtained under the above conditions, with the horizontal axis representing the vibration frequency (maximum 500 Hz) and the vertical axis representing the amplitude value (μm).

First, constant standards of vibration levels D ₁ and D ₂ obtained through past results, experiments, etc. are set and input into the microcomputer system. This vibration level D ₁ , D ₂
For example, let D ₁ = 50 μm and D ₂ = 75 μm, and if the sampling value over the entire predetermined wave number range of the vibration spectrum (x) is between D ₁ and D ₂ , be careful, and if it exceeds D ₂ , it is considered abnormal. judge.

Next, an interval for diagnosis is determined, and an amplitude spectrum (x) in the frequency range is created every predetermined period of time and displayed on the display device at the same time as the lines displaying the vibration levels D ₁ and D ₂ . do.

This display allows the vibration state of the rotating machine at the time of diagnosis to be immediately determined.

However, the amplitude spectrum at a certain diagnosis is the first one.
As shown in the figure, if the vibration level exceeds _D2 and enters the abnormal region, it is assumed that there is a failure or serious problem, and necessary maintenance is performed.

When diagnosing a failure by displaying such an amplitude spectrum, the frequency of the amplitude spectrum that has entered the abnormal region can be easily traced, so the location of the failure can be easily determined.

When a defect occurs, it is thought that the defect will grow over time and lead to a failure.If the vibration level exceeds _D1 at a certain diagnosis point and enters the caution area, the diagnosis period should be shortened and the time of failure can be determined. Be sure to understand.

In the above embodiment, the amplitude spectrum (x) was selected as the diagnostic index, but it goes without saying that fault diagnosis can be performed using the same method even if the acceleration spectrum (α) or the speed spectrum (υ) is selected. .

FIG. 2 is a block diagram of a preferred apparatus for carrying out the invention, which apparatus will now be briefly described.

In the figure, 1 is a sensor for detecting acceleration, and 2 is a low-pass filter for reducing aliasing errors when A/D converter 3 converts the analog signal (acceleration) from sensor 1 into a digital signal. , the signals digitized by the A/D converter 3 are input to a master microcomputer system 4. A diagnostic condition setting device 5 is connected to the master microcomputer system 4 to set and input the diagnostic interval and vibration levels D ₁ and D ₂ , as well as input the arithmetic processor 6 , main memory 7 , auxiliary memory 8 , diagnostic display section 9 , and alarm. The device 10 is connected to the master microcomputer system 4.

The master microcomputer system 4 performs averaging calculations to extract the characteristics of the signal from the A/D converter 3, digit number adjustment calculations to prevent underflow and overflow, etc., and then processes the main memory 7, arithmetic processor 6, Frequency analysis (FFT) is performed together with the auxiliary memory 8 to convert a signal on the time axis into a vibration spectrum on the frequency axis. Here, the arithmetic processor 6 performs butterfly operations (multiplication, addition) in frequency analysis and square root operations for obtaining a spectrum. The auxiliary memory 8 stores a microprocessor, a main memory 7, and an operating system that assists in programming the arithmetic processor 6, and also stores arithmetic results (amplitude spectra). The master microcomputer system 4 displays the calculation results on the diagnostic display section 9 according to the diagnostic conditions inputted by the diagnostic condition setting device 5, and drives the alarm device 10 as necessary, as well as the display, alarm, and automatic stop system. A starting signal 17 is output.

Also, master microcomputer system 4
is connected to a power source 12 via a semiconductor relay 11, and is configured to be energized and driven only at a required time (diagnosis time).

A spray microcomputer system 13 is connected to the master microcomputer system 4, and the spray microcomputer system 13 is equipped with a memory 14 and a clock 15.
Further, a battery backup circuit 16 connected to the power supply 12 is connected to the slave microcomputer system 13 and the clock 15.

The slave microcomputer system 13 is protected against power outages by the battery backup circuit 16 and is always in operation. Based on the results measured by the clock 15 and based on the diagnostic conditions from the master microcomputer system 4, the slave microcomputer system 13 Drives the semiconductor relay 11 to control the diagnostic interval,
In addition, necessary data is memorized and stored in preparation for a power outage after the diagnosis is completed, and the master microcomputer 4 is stopped via the semiconductor relay 11 after the diagnosis is completed, and then activated 30 minutes before the next diagnosis. .

As described above, according to the present invention, frequency analysis (FFT) makes it possible to select sufficiently fine frequencies over the entire predetermined frequency range, and it is possible to estimate the presence or absence of an abnormality in rotating machinery for each famous part. It is possible to establish a failure diagnosis technique, and to know the time of failure with certainty, thereby increasing the reliability of the plant and significantly reducing economic losses.

[Brief explanation of drawings]

FIG. 1 is a diagram of an amplitude spectrum, which is one of the indicators for fault diagnosis according to the present invention, and FIG. 2 is a block diagram of a preferred apparatus for carrying out the present invention. 1 is a sensor, 4 is a master microcomputer system, 9 is a diagnostic display section, and 13 is a slave microcomputer system.

Claims (1)

[Claims] 1. Detect the acceleration of the rotating machine, perform frequency analysis based on the detection result, and create at least one of an acceleration spectrum, a speed spectrum, and an amplitude spectrum in a predetermined frequency range for each required period, covering the entire predetermined frequency range. A method for diagnosing a rotating machine in which an abnormality in the rotating machine is determined by determining whether the spectrum exceeds a vibration level determined experimentally or empirically at each sampling point set at small frequency intervals. .