JPH08247076A - Performance degradation diagnosing device for compressor - Google Patents

Abstract

PURPOSE: To provide a performance degradation diagnosing device of a compressor to judge a performance degradation degree of the compressor in accordance with an oscillatory waveform of the compressor and to efficiently and precisely judge compression performance, especially refrigerating capacity by determining an unsteady component with fluid oscillation of the compressor as its exciting source. CONSTITUTION: This device is constituted by furnishing a sensor 4, an amplifier 5, a waveform recorder 6, a filter 7, an A/D converter 8, a waveform processor 9, an indicator 10, a recorder 11 and a quantifying means of an unsteady component of an oscillatory waveform in the waveform processor 9.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a non-destructive abnormality diagnosis of a compressor, and more particularly to a method and apparatus for diagnosing an abnormality in the compression performance of a compressor by vibration.

[0002]

2. Description of the Related Art As a compressor, there is one that compresses a refrigerant such as used in a refrigerating machine. This compressor is usually connected to a refrigeration system, and its basic function is to suck a low-temperature low-pressure refrigerant gas, perform adiabatic compression in the compressor, and discharge a high-temperature high-pressure refrigerant gas.

However, the compression element portion of the compressor is made of metal, and there are many sliding portions between the metals. For this reason, when foreign matter such as dust is mixed in the sliding portion, the foreign matter causes damage to the sliding portion and causes a decrease in compression performance (for example, refrigerating capacity) due to refrigerant leakage during compression. The sliding parts may be scratched, causing sliding parts to wear and resulting in poor compression.

Further, due to a mistake in assembling the compressor, a large clearance may be assembled, and a specified compression performance may not be output due to refrigerant leakage during compression.

Therefore, in order to judge the deterioration of the compression performance and the compression failure due to these refrigerant leaks, the compressor was separated from the refrigeration system and evaluated by the calorie measuring device.

However, this is expensive and time-consuming, and it was practically impossible to manage the number. Although it is possible to detect the temperature in the freezer / refrigerator compartment, various causes may be considered not only in the performance of the compressor, but also in the refrigerant gas, heat exchanger, heat insulation performance, temperature detection method, etc. It could not be diagnosed correctly.

Therefore, a method for efficiently and accurately diagnosing the degree of performance deterioration of the compressor has been desired.

In order to meet these demands, there is a method of determining whether or not the compressor is operating abnormally by detecting a current value such as current or gas pressure and a pressure value.

Further, in order to detect an abnormality inside the compressor at an early stage, a method for detecting and diagnosing vibrations, sounds and AE signals transmitted to the outside of the compressor has been devised.

For example, Japanese Unexamined Patent Publication No. 62-75095 and Japanese Unexamined Patent Publication No. 2-205728 have been proposed, and Japanese Unexamined Patent Publication No. 62-75095 detects abnormalities such as a cylinder damage of a compressor and dust clogging. Therefore, there is a method of detecting noise or vibration generated from the compressor, and comparing it with a reference value from the ratio of the maximum value and the minimum value of the integrated fluctuation amount of the signal for each section divided in the wavelength direction to determine an abnormality. .

Further, in Japanese Patent Application Laid-Open No. 2-205728, in order to detect defects such as scratches and wear of the sliding portion, an AE signal generated from the compressor is detected, and the AE signal is envelope-detected. There has been devised a method of judging an abnormality by comparing with a reference value from the number of generated AE signals exceeding a certain threshold value generated per cycle or the intensity of the rotation tuning component.

[0012]

However, in the above-mentioned configuration, the current detection cannot determine that the compressor does not deteriorate comparatively in strength enough to cause a compression failure, and the compression due to gas leakage during compression occurs. It cannot detect subtle changes in performance.

The same applies to gas pressure detection. Further, according to the method disclosed in Japanese Patent Laid-Open No. 62-75095, it is possible to make a judgment only by the abnormality such as the audible feeling such as the foreign matter mixed in the compressor. Basically, the fluctuation of the mechanical vibration is extracted. The compression performance of can not be judged.

Further, A of Japanese Patent Laid-Open No. 205205/1990
Even the diagnosis using the E signal quantifies the characteristics of the AE signal due to the damage of the sliding portion, and there is a problem that the compression performance of the compressor cannot be diagnosed.

In view of the above problems, it is an object of the present invention to provide a compressor performance deterioration diagnosing device which can efficiently and accurately diagnose the degree of deterioration of the compression performance of the compressor.

[0016]

In order to achieve the above object, a compressor performance deterioration diagnosing device of the present invention provides a vibration waveform generated by being vibrated by a fluid vibration generated in a process of compressing a refrigerant in a compressor. A sensor for detecting, an amplifier for amplifying the vibration waveform, a waveform recording device for recording the signal output from the amplifier, a filter for removing a noise component of the signal output from the waveform recording device or the amplifier, and An A / D converter for A / D converting the signal output from the filter and the characteristics of the signal output from this A / D converter are subjected to waveform processing, and the degree of performance deterioration is determined from this waveform processing value. It is composed of a waveform processor, a display for displaying these results, and a recording device for recording these results. It is to determine the performance degradation degree by quantifying the stationary component.

Further, in the quantification process of the non-stationary component of the fluid vibration performed by the waveform processor, means for dividing the input vibration waveform for each cycle, and means for averaging the waveforms divided for each cycle. A means for calculating a difference waveform between the divided waveform and the waveform obtained by the arithmetic mean processing, a means for calculating a square of the difference, and a means for obtaining a dispersed waveform by arithmetic mean processing of the squared waveforms. , Means for obtaining the dispersion power by integrating this dispersion waveform, means for obtaining the vibration power by performing the root mean square of the vibration waveform, and means for calculating the ratio between the dispersion power and this vibration power to obtain the specific variance value, The degree of performance deterioration is determined.

Further, as means for dividing the vibration waveform input to the waveform processor into cycles, the vibration waveform is divided into cycles using autocorrelation analysis processing.

In the means for integrating the dispersed waveform in the waveform processor, after smoothing the dispersed waveform,
It is to integrate.

Further, in the means for integrating the dispersed waveform performed by the waveform processor, the integration is performed while the valve of the compressor is open.

Further, in the means for determining the integral range of the dispersed waveform, the integral range while the valve of the compressor is open is provided with a threshold value for the dispersed waveform, and is obtained from the intersection.

[0022]

According to the present invention, with the above-described structure, the sensor detects the vibration waveform up to the ultrasonic region where the fluid vibration transmitted to the outer shell of the compressor is used as the vibration source, and the detected signal is a small signal. Therefore, this signal is amplified by the amplifier. This signal is then recorded on the waveform recorder for later processing. The filter removes the noise component generated from other than the compressor of the signal output from the waveform recording device or the amplifier, so that the low frequency component and the high frequency component are removed. An analog signal of the filtered signal is converted into a digital signal by an A / D converter and is output to a waveform processor equipped with a DSP.
In this waveform processor, the unsteady vibration generated by the influence of the fluid vibration of the input signal is quantified. Further, the compression performance of the compressor is determined from the database showing the compression performance (for example, refrigeration capacity) corresponding to this quantified numerical value. These results can be output on a screen or the like by a display unit to check the degree of performance deterioration. In addition, the result is stored in a recording device such as a hard disk to store data.

Further, in the waveform processor, the input vibration waveform is divided for each cycle, and the divided vibration waveforms are added and averaged. Next, the difference between this average waveform and the waveform divided for each cycle is obtained, and then the waveform obtained by squaring the difference is calculated for each waveform divided for each cycle.
Further, the calculated waveforms are added and averaged, and the average of the dispersed waveforms of the divided waveforms with respect to the average waveform of one cycle is calculated. Then, the waveform is integrated to calculate the dispersion power. On the other hand, the root mean square is performed on the input vibration waveform to obtain the vibration power. Finally, the ratio between the calculated vibration power and the dispersion power is calculated, and the specific dispersion value is obtained. As a result, the unsteady component of the fluid vibration corresponding to the compression performance of the compressor is quantified, and the compression performance can be accurately determined.

Further, autocorrelation analysis calculation is performed in order to determine the cycle pitch for dividing the vibration waveform input to the waveform processor into cycles. As a result, the rotation cycle appears as a peak of the correlation analysis waveform, and a pitch between peaks represents one cycle pitch. Based on this pitch, the vibration waveform is divided into cycles. As a result, the error of each waveform due to the division is eliminated, and the accurate division of the waveform can be performed.

Further, while the dispersion power is calculated from the dispersion waveform, the dispersion waveform is once smoothed, and then integrated to calculate the dispersion power. This makes it possible to clarify the integration range and simplify the data, improve the calculation speed, and facilitate the handling of the data.

Further, when the dispersion waveform is integrated, it is integrated while the valve of the compressor is open, so that the dispersion power related to the fluid vibration corresponding to the compression performance can be accurately extracted, and the compression performance can be improved. Deterioration can be accurately diagnosed.

Further, in determining the opening range of the compressor valve, the opening range of the valve is determined from the intersection of the threshold value and the dispersion waveform. As a result, the opening range of the valve can be accurately detected.

[0028]

EXAMPLE A compressor used in a refrigerator or the like will be described in detail as an example.

(Embodiment 1) A compressor performance deterioration diagnosing device according to an embodiment of the present invention will be described below with reference to the drawings.

In FIG. 1, reference numeral 1 is a compressor for compressing CFC gas and the like, and 2 is a four-point welded portion fixing a compression element to a closed case 3 which is an outer shell. Reference numeral 4 denotes a sensor for detecting a vibration waveform, which can detect a vibration component up to about 100 KHz from the fundamental wave of the compressor. The sensor 4 is installed at one point of the four-point welded portion 2 where the signal from the inside of the compressor 1 is most strongly transmitted. Reference numeral 5 is for amplifying the signal detected by the sensor 4, and reference numeral 6 is a waveform recording device for recording this signal. 7 is a waveform recording device 6 and
It is a filter for removing the noise of the signal from the amplifier 5. Reference numeral 8 is an A / D converter for A / D converting the signal output from the filter 7. Reference numeral 9 denotes waveform processing for quantifying an unsteady component due to fluid vibration in the compression process of the compressor 1 from the A / D converted signal, and a performance deterioration degree from a database of performance deterioration corresponding to the quantified numerical value. It is a waveform processor equipped with a DSP or the like for determining. Reference numeral 10 denotes a display device that outputs and displays the results on a screen or a printer, and these results are recorded in a recording device 11 such as a hard disk. The database is also recorded in the recording device 11.

Here, the structure and vibration waveform of the compressor will be described. 2 and 3 show the structure of the compressor 1, in which the electric motor 12 and the compression element 13 are arranged inside the sealed case 3. The compression element 13 mainly includes the crankshaft 14 and the crankshaft 14 that transmit the rotational movement of the electric motor 12.
A main bearing 15, a sub bearing 16, a cylinder 17, a roller 18, a vane 19, and a valve 20 housed inside the sub bearing 16. Lubricating oil 21 is lubricated for sliding between the respective components of the compression element 13.
Reference numeral 22 is a suction pipe and 23 is a discharge pipe.

The refrigerant passes through a suction pipe 22 from a cooling system (not shown) and is sucked from a suction port 23 of the cylinder 17 into a suction chamber 24 in the cylinder 17. Next, the rotation of the crankshaft 14 causes the roller 19 to rotate in the cylinder 17, whereby the volume is reduced and the refrigerant is adiabatically compressed to become a high-temperature high-pressure refrigerant.

Here, since the vane 18 follows the roller 19 and reciprocates when the roller 19 turns, the vane 18 serves as a partition in the cylinder 17 to form the suction chamber 24 and the compression chamber 25, and the refrigerant is efficiently used. Compresses well.

The high-temperature and high-pressure refrigerant gas pushes up the valve 20 and is guided into the muffler 27 through the discharge notch 26. After that, the muffler 27 is discharged from a discharge hole (not shown) into the discharge space 28 of the closed case 3, and is guided from the discharge pipe 23 to the cooling system.

FIG. 4 shows the relationship between the vibration waveform of one rotation measured by the sensor 4 and the behavior of mechanical parts in one rotation of the compression element 13 in the cylinder 17. The behavior of the mechanical parts is such that the states A to D are one rotation, and the compression of the refrigerant is repeated.

The state of A represents the state of the compression start point, and the vane 18 is most retracted in the cylinder 17. In this state, the low temperature and low pressure refrigerant has been sucked into the suction chamber 24.

This state is set at a rotation angle of 0 degree and is particularly called top dead center. At this top dead center, a large shock wave is generated in the vibration waveform due to the collision between the vane 18 and other parts.

In the state of B, the rotation is advanced by 180 degrees,
The vane 18 is in the most protruding state to the cylinder 17,
Called bottom dead center. In this state, the suction chamber 24 is reduced in volume to become the compression chamber 25, and a new suction chamber 24 'is formed. In addition, the vane 18 is unstablely supported by the vibration waveform, and a small shock wave is generated.

In the state of c, the rotation is advanced by 270 degrees, and the gas pressure load of the compression chamber 25 is higher than the pressing force of the valve 20 to push up the valve 20 and the compressed refrigerant is discharged to the outside of the cylinder 17. Is.

The state of (d) represents a state in which compression has been completed by performing one rotation. In the vibration waveform, a shock wave is generated from 270 degrees as the starting point to the top dead center, and it is considered that the gas pressure load is generated because the peripheral parts are sliding while being pressed against the supporting parts. . Further, although the refrigerant is theoretically sealed in the cylinder 17 by the oil seal between the parts, the present situation is that there is some refrigerant leakage during one rotation.

Particularly, since the refrigerant becomes high temperature and high pressure from the rotation angle of 270 degrees to the top dead center, the refrigerant is discharged to the outside of the cylinder as the valve is opened, and the refrigerant leaks to the outside of the compression chamber 25 remarkably. .

Therefore, the gas pressure load of the refrigerant, which greatly contributes to the generation of the shock wave of the vibration waveform, is greatly affected by the release of the gas and the gas leakage. Further, although the shock wave of the vibration waveform looks periodic in macro, in the micro portion, the unsteady vibration is repeated for each rotation due to the fluctuation of the gas pressure load.

The unsteady component tends to be large when the refrigerating capacity of the compressor is large and small when the refrigerating capacity is small.

The operation of the embodiment configured as above will be described. The vibration component generated from the inside of the compressor 1 is transmitted inside the mechanical component and is transmitted to the four-point welded portion 2. The transmitted vibration is detected by the sensor 4 installed at one point of the four-point welded portion 2 and converted into a voltage signal or the like.

This signal is amplified by the amplifier 5, and the output signal is once recorded and reproduced by the waveform recording device 6 and output to the filter 7 or directly to the filter 7.

The signal input to the filter 7 is mainly mixed with noise from the low frequency range from the outside due to the condition of measuring the vibration of the compressor 1 and the like, and when A / D conversion is performed at the high frequency range. I am worried about the generation of aliasing noise.

Therefore, the filter 7 removes these noise components and outputs them to the A / D converter 8.

For example, in this case, the lower limit frequency is 300 Hz
And the upper limit frequency was set to 100 KHz, and filtering was performed. In the A / D converter 8, the input analog signal is converted into a digital signal and output to the waveform processor 9.

For example, in this case, the sampling frequency is 300 KHz. In the waveform processor 9, the waveform processing operation such as quantifying the non-stationary component of the above-mentioned vibration waveform is performed by the DS.
The compression performance is determined using P or the like, and then from a database showing the correspondence between these calculation results and the compression performance (for example, a database of the refrigeration capacity and the waveform processed value in this case). The determined result can be output and confirmed on the display unit 10.

Further, these results are stored in the recording device 11 and a database for each compressor is constructed. From the above, the degree of performance deterioration of the compressor can be efficiently determined by vibration.

(Second Embodiment) A second embodiment of the present invention will be described below with reference to the drawings.

In contrast to the first embodiment, the waveform processing means for quantifying the unsteady component of the vibration waveform performed by the waveform processor 9 is defined in order to accurately determine the degree of performance deterioration of the compressor.

In FIG. 5, the vibration waveform is input to the waveform processor 9 at 27. For example, in this case, a vibration waveform for 20 cycles was input. The input vibration waveform is divided at 28 for each cycle, and at 29, the arithmetic mean waveform of the vibration waveform for 20 cycles is calculated. Next, the difference between each of the vibration waveforms added and averaged in 30 and each of the vibration waveforms divided in 28 is obtained for each waveform, and in 31 the difference is squared for each vibration waveform.

Further, in 32, the averaging process is performed on each squared waveform to calculate the dispersed waveform. At 33, this dispersion waveform is integrated to calculate the dispersion power representing the non-stationary component of the vibration waveform. On the other hand, the root mean square of the vibration waveform input at 27 is performed at 34 to calculate the vibration power. Finally, at 35, the ratio between the dispersion power and the vibration power is calculated, and the specific dispersion value is obtained. The calculation result is output from 36.

FIG. 6 shows the vibration waveforms 37, 38 and the dispersion waveforms 39, 40 of the large refrigeration capacity product and the small refrigeration capacity product, and the size of the shaded area of the dispersion waveforms 39, 40 represents the dispersion power. It can be seen that this area is larger when the refrigerating capacity is larger. That is, the larger the dispersed power, the larger the refrigerating capacity.

Further, the reason why the ratio with the vibration power is calculated is to exclude the fluctuation of the dispersed power due to the load applied to the compressor or the model. From the above, by calculating the specific variance value by this waveform processing means, it is possible to efficiently determine the refrigerating capacity which is the compression performance.

(Embodiment 3) A third embodiment of the present invention will be described below with reference to the drawings.

In contrast to the second embodiment, means for obtaining the pitch number of one cycle for dividing the input waveform for 20 cycles into each cycle is defined.

FIG. 7 shows the relationship 41 of the autocorrelation analysis value with respect to the number of sample data. Correlation analysis calculation is performed while shifting the time axis of one time waveform of the waveforms of 20 cycles input to the waveform processor 9.

Therefore, a peak of the correlation analysis value appears every time the time axis shift is N times the cycle. This peak represents the average cycle of 20 cycles. As a result, as shown in FIG. 7, by counting the number of samples from the initial state to the first peak 42, the number of pitches in one cycle can be efficiently and accurately obtained. In this case, it was about 5,130 points.

(Embodiment 4) A fourth embodiment of the present invention will be described below with reference to the drawings.

In the second embodiment, the means for integrating the dispersed waveform 32 performed by the waveform processor 9 to calculate the dispersed power 33 is defined.

The dispersed waveform 32 has a very large number of samples, and if it is integrated as it is, the numerical value will be an enormous numerical value and the integration range will be unclear, which will cause an error. 33 was calculated. In this case, 60 points, 0.2 in time
Smoothing processing was performed at intervals of msec.

FIG. 8 shows the dispersed waveform 43 before the smoothing processing and the dispersed waveform 44 after the smoothing processing. After the processing, the shape of the dispersed waveform, which was much more frequent than that before the processing, becomes clear. I understand that. As a result, the integration range becomes clear, and the dispersion power can be calculated with high accuracy. In addition, the overflow due to the enormous number is eliminated and the processing speed is improved.

(Embodiment 5) A fifth embodiment of the present invention will be described below with reference to the drawings.

The integral range of the integral process performed by the waveform processor 9 of the second embodiment to calculate the dispersed power 33 from the dispersed waveform 32 is defined.

FIG. 9 shows the vibration waveform 45 and the dispersion waveform 46. While the valve in the vicinity of the rotation angle 270 degrees of the vibration waveform 45 is opened and closed, the fluid vibration as the vibration source fluctuates in every cycle due to the influence of the refrigerant gas release and the gas leakage, and the vibration waveform 45 is unsteady. This is the part where the component is particularly large.

A shock wave having a large top dead center is largely due to mechanical vibration, and since it has a sharp peak, it is a major cause of error in waveform processing. Therefore, by integrating in the integration range 50 between the valve open point 48 and the valve close point 49 where a large amount of unsteady components are generated, the accurate distributed power 33 corresponding to the compression performance is obtained.
Can be requested.

(Embodiment 6) A sixth embodiment of the present invention will be described below with reference to the drawings.

In the definition of the integration range of the waveform processing of the fifth embodiment, the recognition means of the integration range is specified.

In the distributed waveform 51 of FIG. 10, a threshold 52 is used as a trigger level to search from the top dead center peak 53 in the reverse rotation direction to find an intersection A54, and the rotation angle 2
By searching from around 70 degrees and finding the intersection B55, the integration range 56 can be efficiently and accurately found.

[0072]

As described above, according to the present invention, the sensor for detecting the vibration waveform generated by the fluid vibration generated in the process of compressing the refrigerant in the compressor, the amplifier for amplifying the vibration waveform, and the A waveform recording device for recording a signal output from an amplifier, a filter for removing a noise component of a signal output from the waveform recording device or the amplifier, and an A / D converter for A / D converting the signal output from the filter. An A / D converter, a waveform processor that performs waveform processing on the characteristics of the signal output from the A / D converter, and determines the degree of performance deterioration from the waveform processing value; and a display that displays these results. It is composed of a recording device that records these results.The degree of performance deterioration can be efficiently determined by quantifying the unsteady component of fluid vibration that occurs in the compression process of the compressor in the waveform processor. It can be.

Further, in the quantification process of the unsteady component of the fluid vibration performed by the waveform processor, the means for dividing the input vibration waveform for each cycle and the means for averaging the waveforms divided for each cycle. A means for calculating a difference waveform between the divided waveform and the waveform obtained by the arithmetic mean processing, a means for calculating a square of the difference, and a means for obtaining a dispersed waveform by arithmetic mean processing of the squared waveforms. , Means for integrating the dispersion waveform to obtain the dispersion power, means for obtaining the vibration power by performing the root mean square of the vibration waveform, and means for calculating the ratio between the dispersion power and the vibration power to obtain the specific variance value. ,
By performing the vibration waveform processing, it is possible to accurately determine the degree of performance deterioration.

Further, as a means for dividing the vibration waveform input to the waveform processor into cycles, the vibration waveform is divided into cycles using autocorrelation analysis processing, so that the vibration waveform is accurately divided into cycles. can do.

In the means for integrating the dispersed waveform in the waveform processor, after smoothing the dispersed waveform,
By integrating, the distributed power can be accurately obtained at high speed without error.

Further, in the means for integrating the dispersed waveform performed by the waveform processor, by integrating while the valve of the compressor is open, the dispersed power of the aperiodic component of vibration corresponding to the compression performance can be further obtained. It can be calculated.

In addition, in the means for determining the integral range of the dispersed waveform, the integral range while the valve of the compressor is open is determined by setting a threshold value on the dispersed waveform from the intersection point thereof, thereby making the integral range efficient and accurate. You can ask. As a result, it is possible to provide a compressor performance deterioration diagnosing device capable of efficiently and accurately diagnosing the compression performance of the compressor without measuring each by the calorie measuring device.

[Brief description of drawings]

FIG. 1 is a block diagram of a compressor performance deterioration diagnosing device showing a first embodiment of the present invention.

FIG. 2 is a cross-sectional view of a compressor that is a diagnosis target of the present invention.

FIG. 3 is a sectional view taken along the line AA ′ of the compressor which is a diagnostic target of the present invention.

FIG. 4 is a diagram of a mechanical behavior and a vibration waveform which are diagnostic targets of the present invention.

FIG. 5 is a waveform processing flowchart showing a second embodiment of the present invention.

FIG. 6 is a vibration waveform and dispersion waveform diagram for showing the second embodiment of the present invention.

FIG. 7 is an autocorrelation analysis diagram for showing the third embodiment of the present invention.

FIG. 8 is a distributed waveform diagram before and after smoothing processing for showing a fourth embodiment of the present invention.

FIG. 9 is a vibration waveform and dispersion waveform diagram for showing a fifth embodiment of the present invention.

FIG. 10 is a distributed waveform chart for showing a sixth embodiment of the present invention.

[Explanation of symbols]

 1 Compressor 2 4-point welded part 3 Sealed case 4 Sensor 5 Amplifier 6 Waveform recording device 7 Filter 8 A / D converter 9 Waveform processor 10 Display device 11 Recording device

Claims (6)

[Claims] 1. A sensor for detecting a vibration waveform generated by being excited by fluid vibration generated in the process of compressing a refrigerant in a compressor, an amplifier for amplifying the vibration waveform, and a signal output from the amplifier is recorded. Waveform recording device,
A filter for removing a noise component of a signal output from this waveform recording device or an amplifier, an A / D converter for A / D converting the signal output from this filter, and an output from this A / D converter Waveform processing is performed on the characteristics of the signal to be processed, and a waveform processor that determines the degree of performance deterioration from this waveform processing value, a display that displays these results, and a recording device that records these results. A performance deterioration diagnosis device for a compressor, wherein a degree of performance deterioration is determined by quantifying an unsteady component of fluid vibration generated in a compression process of the compressor in a waveform processor. 2. In the quantification process of the non-stationary component of the fluid vibration performed by the waveform processor, means for dividing the input vibration waveform for each cycle, and means for averaging the waveforms divided for each cycle. A means for calculating a difference waveform between the divided waveform and the waveform obtained by the arithmetic mean processing, a means for calculating a square of the difference, and a means for obtaining a dispersed waveform by arithmetic mean processing of the squared waveforms. , A ratio dispersion value is obtained by means for integrating the dispersion waveform to obtain the dispersion power, means for obtaining the vibration power by performing the root mean square of the vibration waveform, and means for calculating the ratio between the dispersion power and the vibration power. The performance deterioration diagnosing device for a compressor according to claim 1, wherein the vibration waveform processing is performed by means. 3. The compressor according to claim 2, wherein the vibration waveform input to the waveform processor is divided into cycles by using an autocorrelation analysis process as means for dividing the vibration waveform into cycles. Performance deterioration diagnostic device. 4. The performance deterioration diagnosing device for a compressor according to claim 2, wherein the means for integrating the dispersed waveform in the waveform processor performs the smoothing processing on the dispersed waveform and then integrates it. 5. The performance deterioration diagnosing device for a compressor according to claim 2, wherein the means for integrating the dispersed waveform performed by the waveform processor integrates while the valve of the compressor is open. 6. The dispersion waveform integration range determining means, wherein the integration range while the valve of the compressor is open is set to a threshold value for the dispersion waveform, and the integration range is determined from the intersection thereof. Compressor performance deterioration diagnosis device.