JP2003299242A - Motor overload estimation system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To qualitatively estimate the overload of a motor easily in real time. <P>SOLUTION: This system includes a current detection means 1 for detecting the load current of the motor. Based on a differential output obtained by differentiating the load current output detected by the detection means 1 and a traveling average derived from the load current, such a generating pattern that the moor is in an overload condition is identified to be estimated, thus easily applying the estimation of the overload of the motor to analysis and failure processing in real time. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a motor overload prediction system, and more particularly to a large-scale industrial equipment such as a crusher, a processing machine, a high-speed stirring / kneading device ... The present invention relates to a motor overload prediction system suitable for overload prediction.

[0002]

2. Description of the Related Art Motors used in large-scale industrial equipment, such as crushers, processing machines, high-speed stirring / kneading devices, etc., have a built-in overload protection circuit using a thermistor.

Therefore, if an overload is generated on the rotating shaft of the motor and the rotating shaft is locked, the overload protection circuit operates and stops. Then, with a crusher or the like, it is necessary to take out all of the contents before restarting, so there is a problem that a great deal of effort is required for restoration.

As one method for solving this problem, a method using a meter relay is used.

A meter relay is a kind of limit operation type relay, which opens a contact when a set current value is reached.

Therefore, in the above-mentioned crusher or the like, for example, before the overload current is reached, the output of the meter relay is used to control the loading conveyor and the closing regulation is performed to prevent the locking of the motor. ing.

However, since this control is an on / off control, there is a problem that it cannot be well controlled depending on the material to be put.

That is, in the case of highly viscous material such as waste soil, even if the meter relay is turned off and the application is stopped, the load remains and the current flows too much, resulting in an overload problem. . Therefore, it was necessary to set the meter relay setting value to a low value.

On the other hand, in the case of rock or the like, it can be destroyed in an instant, so even if an overload occurs, it is instant so there is no need to stop it, and the set value of the meter relay can be set to a large value.

However, since the crushed material is a mixture of these, it is difficult to set the current value exactly according to the crushed material.

Therefore, in place of the meter relay, an overload protection device [called a shock relay (R)] that can open the contact at a set time when the set current value is exceeded is used. .

When an overload protection device of the type that sets such a delay time is used, for example, since the device will not be stopped even if an instantaneous overcurrent flows, it is more difficult than the setting of the current value of the meter relay. Therefore, adjustment is considered to be relatively easy.

However, even when the above-mentioned overload protection device is used, the time must be set depending on the experience and feeling of the operator.

Therefore, in place of such an overload protection device, it is conceivable to use the load current of the motor to characterize the overload phenomenon in real time to predict the overload so that the abnormality can be dealt with.

[0015]

However, in order to analyze the overload prediction in real time by using the load current of the motor and to perform the abnormality treatment, the waveform matching method and the area of the load level and the time axis are required. Judgment methods can be considered, but in order to carry out those methods, it is necessary to accumulate a large amount of data in advance in order to identify innumerable combination conditions which are normal and which are abnormal.

Further, at this time, since the load current contains a large amount of noise of a high frequency component, there is also a problem that it is necessary to remove them.

Therefore, an object of the present invention is to make it possible to easily analyze the overload prediction in real time and take an abnormal action.

[0018]

In order to solve the above-mentioned problems, according to the present invention, a current detection means for detecting a load current of a motor and a transient value based on a differential value of a load current output detected by the current detection means are used. The processing means for predicting the load is adopted, and the processing means adopts a configuration that causes an overload when differential outputs below or above a preset level occur at preset intervals. is there.

By adopting such a configuration, the pattern of occurrence of the overload state of the motor is identified and predicted from the differential value (change amount) of the load current of the motor. That is, the level of the differential output is proportional to the degree of load (the overload is instantaneously generated in the rotating shaft, and the level increases as the lock is applied). On the other hand, since the level generation cycle is proportional to the period in which the motor receives a load, for example, if a differential output of a level higher than a set value is frequently detected (when completely locked, the interval becomes 0). ), The largest overload condition can be predicted.

At this time, a calculating means for deriving the moving average of the load current is provided, and the differential output of which the moving average derived by the calculating means is equal to or more than a preset value and equal to or less than a preset level. It is possible to employ a configuration in which an overload occurs when is continued for a set period of time or more, or when a differential output of a set level or more is generated at a preset interval.

If such a configuration is adopted, an increase in the level of the average value of the load current can be detected, and from that increase,
For example, it is possible to detect even a crushed material having a high viscosity such as waste soil having a high water content, in which the load current gradually increases. Therefore, it is possible to detect the type that cannot be detected by the differential output, so that it is possible to predict the overload regardless of the property of the crushed material.

[0022]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

As shown in FIG. 1, the overload prediction system comprises a current detection means 1, a low-pass filter 2, a differentiation means 3, a moving average calculation means 4 and an analysis / abnormality processing means 5.

The current detecting means 1 is capable of detecting the load current in a non-contact manner such as CT, and is attached to the power line of the motor 6 as shown in FIG. 1, for example.

The low-pass filter 2 has a cutoff characteristic as shown in FIG. 2 (for example, f ₀ = 0.25H).
Z), noise is removed from the output of the current detecting means 1.

The differentiating means 3 is for differentiating the output from the low-pass filter 2 from which noise is removed,
The differentiated output is output to the analysis / abnormality processing means 5 in the subsequent stage.

The moving average calculation means 4 processes the output of the low-pass filter 2 and outputs the moving average of the load current in real time. The calculation result is the analysis / abnormality processing means 5.
Is output to.

The processing means 5 has a comparison means for determining a differential output and a comparison means for determining a moving average.

That is, when the processing means 5 sequentially compares the level of the differential output input from the differentiating means 3 with a preset threshold value, and detects a differential output having a level lower than the threshold value. , Measure how long that low level lasts. Then, the continuation period is compared with the set period, and when the period becomes short, it is determined that the abnormality has occurred.

By doing so, the level of the differential voltage is proportional to the steep rise of the load current I _L of the motor 6, and is therefore proportional to the degree of the load (the higher the level of instantaneous locking, the higher the level). On the other hand, the level continuation period is proportional to the period in which the motor 6 receives a load, and therefore, for example, if a differential voltage at a level higher than a set level is frequently detected (if the cycle becomes short), the overload state is detected. Can be predicted.

At the same time, the values of the moving average are also compared, and when the average value exceeds the set threshold value, it is determined to be abnormal.

That is, the moving average does not suddenly increase even if the load current I _L of the motor 6 suddenly increases, and when the load current I _L gradually increases, the moving average also increases. Therefore, by monitoring this increase, the motor 6 is subjected to a load and stopped due to a highly viscous crushed property such as waste soil with high water content, and the load current I _L gradually increases. It is possible to predict a state of overload that is going on.

Therefore, here, as will be described later, when the duration of the previous differential output exceeds the set value, the moving average value is compared with the set value to detect an excess that cannot be detected by the previous differential output. The load can be predicted.

Since the overload can be predicted in this way, it can be dealt with before the overload occurs in the rotary shaft and the locked state is reached.

Next, it was decided to verify whether or not this system can correctly predict the overload by using the data of the actual crusher 7.

Specifically, as shown in FIG. 3, the load current I _{L of the} motor is detected by the current detection means 1 provided in the start control board 8 of the idal crusher 7, and the detected signal is A /
After being converted to a digital value by the D converter 9 (sampling frequency 1 Hz), it is input to a digital computer (personal computer) 10 and digitally processed.

Therefore, the digital computer 10 is provided with, for example, a low-pass filter 2 formed by a digital filter, a differentiating program 3 provided as differentiating means, and a program 4 provided as moving average calculating means.

The low-pass filter 2 is composed of IIR-type sixth-order non-polar amplitude Chebyshev characteristic, so that the number of calculations is small to narrow the width of the pass band, and the number of multiplications is small with a small number of coefficients. It is suitable for real-time processing.

The differentiation program 3 is, for example, DD
A very general one using a difference such as an A (digital differential analyzer) algorithm is used.

Similarly, the moving average program 4 is also very general, and conversion data a ₀ , a ₁ , a ₂ , a ₃ , a ₄ , a which are successively output from the A / D converter 9 are obtained. ₅ , a
₆ , a ₇ , a ₈ . ． ． For example, (a ₀ + a ₁ + a ₂ + a ₃ ... + a _n ) / n = moving average A1 (a ₁ + a ₂ + a ₃ ... + a _n + a _{n + 1} ) / n = moving average A2 ...・ ・ ・

Further, the digital computer 10 detects abnormalities in real time from the differential values and the moving average, which are input one after another, and predicts overload. The threshold value is set as follows, and the differential value is set. Is calculated.

That is, since the differential value I _L 'of the load current I _{L in} the normal state as shown in FIG. 4A is as shown in FIG. 4B, a threshold value is set and the period is below a certain level. Is being measured. Conversely, it is also possible to measure a period of a certain level or higher. A timer may be prepared for this measurement, but a timer program or an internal timer can be used to perform this measurement with high accuracy.

[0043] Experiment inputs the data of the load current I _L as measured in the data and overload state of the load current I _L measured by the Aidaru grinder 7 in normal operating conditions to the computer 10 of the system, its We examined the results and confirmed that they could be predicted.

This will be described below with reference to the drawings.

FIG. 5 (a) shows the load current I _L when the crusher is in the normal processing state (normal 1), and FIG. 5 (b) is its first derivative I _L '. Further, (c) is a moving average A _V. The moving average _AV is set to an average value for 40 seconds. Therefore, no data is output for 40 seconds after activation.
At this time, the period T at the threshold value = 10 of the first derivative I _L '
f (3 places in the figure) was classified into 50 seconds ≧ Tf> 25 seconds (2 places) and Tf ≧ 50 seconds (1 place). The moving average _AV was less than 250.

Similarly, FIG. 6A is also for normal processing (normal 2), and the period Tf at the threshold value = 10 of the primary differential value from (b) and (c) is ( Here, at two locations, the rising after the startup is ignored), 50 seconds ≧ Tf> 25 seconds and Tf ≧ 50 seconds. The moving average was also less than 250.

By the way, FIG. 7A shows the load current I _L when there is no load (normal 3), and the cycle of the first derivative I _L 'at the threshold value = 10 from (b) and (c). Tf is Tf ≧ 50 seconds.

The moving average A _V was the same, less than 250.

FIG. 8A shows the load current I _L when the crusher 2 is tripped and stopped (abnormality 1), and the threshold value of the first derivative I _L 'from (b) and (c) = 10. All the cycles Tf (9 places in the figure) were classified into 25 seconds ≧ Tf> 5 seconds.

The moving average A _V was less than 250.

FIG. 9 (a) shows the load current I _L when the crusher 2 is stopped before the trip and the blockage inside the machine is confirmed (abnormal 2), and the threshold value of the first derivative I _L '= 10. Cycle T
Although f was innumerable, all were 25 seconds ≧ Tf> 5 seconds.

[0052] Further, the moving average A _V was less than 250.

FIG. 10A shows the load current I _L in the locked state, and the threshold value of the first-order differential value from (b) and (c) =
The period Tf at 10 was (2 places), 25 seconds ≧ Tf> 5 seconds and Tf ≧ 50 seconds.

At this time, the moving average _AV was 250 or more.

The above results are summarized and tabulated in FIG.
It became like.

[0056] By monitoring the level and period Tf and the moving average A _V clearly load current I _L from the result, it is possible to distinguish the abnormal state. Therefore, it was possible to predict overload.

Although the motor of the crusher has been described in this embodiment, it is obvious that the invention is not limited to this and can be applied to any motor. In particular,
For example, it is most suitable for use as a motor for large industrial equipment such as a processing machine, a high-speed stirring / kneading device, ...

[0058]

As described above, the present invention can be applied to the overload prediction easily in real time for analysis and abnormality processing.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment.

FIG. 2 is a diagram showing a filter characteristic of the embodiment.

FIG. 3 is a block diagram of an embodiment.

FIG. 4 is an operation explanatory view of the embodiment.

FIG. 5 (a), (b), (c) waveform chart showing experimental results

FIG. 6 (a), (b), (c) waveform chart showing experimental results

7 (a), (b), (c) Waveform diagram showing experimental results.

8A, 8B, and 8C are waveform charts showing experimental results.

9 (a), (b), (c) waveform chart showing experimental results

10 (a), (b), (c) Waveform diagram showing experimental results.

FIG. 11: Table of experimental results

[Explanation of symbols]

1 current detector 2 low pass filter 3 differentiating means 4 moving average computation unit 5 processor 6 motor 7 Aidaru crusher 9 A / D converter 10 a digital computer A _V moving average I _L load current

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Atsuhiko Shirai             2-35 Jimmucho, Yao-shi Kubotakyu Co., Ltd.             Houji Temple Factory F-term (reference) 5G044 AA01 AC01 AD01 AE01 CA01                       CB01 CE08                 5H570 AA17 AA23 BB09 BB10 CC05                       DD03 DD04 DD08 JJ03 JJ16                       JJ23 JJ26 KK06 LL02 MM04

Claims (2)

[Claims] 1. A current detecting means for detecting a load current of a motor, and a processing means for predicting an overload based on a differential value of a load current output detected by the current detecting means, wherein the processing means is preset. An overload prediction system for a motor, which predicts an overload when a differential output equal to or lower than a predetermined level occurs at a preset interval. 2. A calculation means for deriving a moving average of the load current, wherein a differential output of which the moving average derived by the processing means is a preset value or more and a preset level or less is set. The motor overload prediction system according to claim 1, wherein an overload is predicted when the differential output continues for a predetermined period of time or more or when a differential output of a set level or more occurs at a preset interval.