KR102560570B1 - Oscillation detection method of plant process control loop and oscillation detection apparatus using the same - Google Patents

Abstract

The present invention relates to a method for diagnosing oscillation of a plant process control loop and an oscillation diagnosis apparatus using the same. The method for diagnosing oscillation of a plant process control loop according to an embodiment of the present invention includes a process identified through the plant process control loop. performing FFT (Fast Fourier Transform) on the value data; extracting a peak value from an FFT result value for the process value data; Deriving a result of applying a first limit value for magnitude in the extracted peak value; and detecting and diagnosing the occurrence of oscillation in the plant process control loop.

Description

OSCILLATION DETECTION METHOD OF PLANT PROCESS CONTROL LOOP AND OSCILLATION DETECTION APPARATUS USING THE SAME}

The present invention relates to a method for diagnosing oscillation in a plant process control loop and an oscillation diagnosis device using the same, and more particularly, to an oscillation diagnosis device that automatically detects oscillation regardless of single/multiple, stationary/non-stationary oscillation. The present invention relates to an oscillation diagnosis method of a plant process control loop and an oscillation diagnosis device using the same for detecting and diagnosing the occurrence of oscillation in a plant process control loop using time-series process data by diagnosing the oscillation period and the oscillation period.

In the plant process control loop, process value oscillation may occur due to excessive tuning of controllers, stiction of on-site valves, disturbance, and the like. 1 is a diagram showing an example of oscillation.

Oscillation of plant process values deteriorates plant stability and reduces facility operating efficiency. In particular, since dozens or more of plant process control loops are intricately connected to each other, oscillation of one control loop may affect other related control loops.

As such, diagnosing the occurrence of oscillation has become an important issue in plant process control loops. That is, in diagnosing occurrence of oscillation, since it is impossible for a manager to individually monitor numerous plant process control loops implemented in a plant, a method for automatically diagnosing oscillation is required.

In addition, oscillations may appear in various forms such as single/multiple, stationary/non-stationary. That is, to diagnose the occurrence of oscillation, it is necessary to be able to monitor and diagnose various types of oscillation.

Here, the stationary signal is a signal in which the mean, variance, and covariance of the process value data distribution do not change with time, and the non-stationary signal is a signal in which the mean, variance, and covariance of the process value data distribution change with time. .

Previously, US Patent Registration No. 5,719,788 has been proposed for determining oscillation using an attenuation ratio of an autocorrelation function (ACF) of input data. Here, US Patent No. 5,719,788 defines the ratio of the first valley (b) and the second mountain (a) of the ACF in FIG. 2 as the oscillation index 'R', and when the oscillation index 'R' is greater than 0.5, Diagnose that rationing has occurred. 2 is a diagram illustrating an ACF damping ratio measurement method.

In addition, Korea Patent Registration No. 1,267,681 has previously been proposed that diagnoses oscillation by determining the regularity of the zero point passing interval of the ACF after obtaining the ACF of the process value. Here, Korean Patent Registration No. 1,267,681 calculates the regularity index 'r' using the average (Tp) of the zero-passing period and the standard deviation (σ _Tp ) of the zero-passing period in FIG. If the index 'r' is greater than 1, it is diagnosed that oscillation has occurred. 3 is a diagram for explaining an ACF zero passing period determination method.

By the way, U.S. Patent No. 5,719,788 and Korean Patent No. 1,267,681 are non-stationary composed of sine waves of three frequencies (0.159, 0.318, 0.477 Hz) and a ramp signal with a period of 1800 seconds as shown in FIG. When multiple oscillation signals are applied as test signals, oscillation can be misdiagnosed. 4 is a diagram illustrating a multi-oscillation signal.

First, in U.S. Patent No. 5,719,788, when the multiple oscillation signal of FIG. 4 is applied, the oscillation index R=B/A=0.2641 as shown in FIG. . 5 is a diagram showing a result of determining an ACF attenuation ratio of a test signal.

Next, in Korean Patent Registration No. 1,267,681, when the multiple oscillation signal of FIG. 4 is applied, the regularity index r=0.1454 as shown in FIG. 6 is a diagram showing the result of determining the ACF zero pass period of the test signal.

Therefore, in diagnosing oscillation, there is a need to prepare a method for automatically diagnosing whether or not oscillation occurs and an oscillation period regardless of single/multiple, stationary/non-stationary oscillation.

US Patent Registration No. 5,719,788 (Registered on February 17, 1998) Korean Registered Patent Publication No. 10-1267681 (registered on May 20, 2013)

An object of the present invention is to automatically diagnose oscillation and oscillation cycles regardless of single/multiple, stationary/non-stationary oscillation, thereby preventing the occurrence of oscillation in a plant process control loop using time-series process data. An oscillation diagnosis method of a plant process control loop for detecting and diagnosing, and an oscillation diagnosis device using the same are provided.

An oscillation diagnosis method of a plant process control loop using an oscillation diagnosis apparatus according to an embodiment of the present invention includes the steps of performing FFT (Fast Fourier Transform) on process value data identified through a plant process control loop; going through a preprocessing process of measuring the length of the process value data by taking an absolute value of the FFT result value for the process value data, and extracting a peak value from the FFT result value; Deriving a result of applying a first limit value for magnitude in the extracted peak value; and detecting and diagnosing occurrence of oscillation in the plant process control loop.

According to an embodiment, after the step of deriving a result of applying the first threshold value, deriving a result of applying a second threshold value to an adjacent frequency component; may further include.

The detecting and diagnosing step may include detecting and diagnosing the occurrence of oscillation in the plant process control loop using the number of frequencies identified through a result of applying the second threshold value.

In the detecting and diagnosing step, if the number of frequencies is 0, it is determined that no oscillation has occurred, if the number of frequencies is 1, it is determined that a single oscillation has occurred, and if the number of frequencies is greater than 1, it is determined that multiple oscillations have occurred. It may be judged by the case where this occurred.

The process value data may be any one or more time series data selected from among single, multiple, stationary, and non-stationary data.

In the step of extracting the peak value, an absolute value of the FFT result value may be subjected to a preprocessing process.

The first limit value may be set to a value obtained by adding the 'average of absolute values of FFT result values' and 'three times the standard deviation of absolute values of FFT result values'.

The second threshold may be set to 0.001 to 0.003 Hz.

In addition, an oscillation diagnosis apparatus according to an embodiment of the present invention includes at least one processor; and a memory for storing computer readable instructions, wherein the instructions, when executed by the at least one processor, cause the oscillation diagnosis apparatus to process value data identified through a plant process control loop. FFT (Fast Fourier Transform) is performed, a peak value is extracted from the FFT result value for the process value data, a result of applying a first limit value for the size is derived from the extracted peak value, and the plant process control It may be to detect and diagnose the occurrence of oscillation on the loop.

The instructions, when executed by the at least one processor, cause the oscillation diagnosis device to derive a result of applying a second threshold value for an adjacent frequency component after deriving a result of applying the first threshold value. it may be

The commands, when executed by the at least one processor, cause the oscillation diagnosis device to detect and diagnose the plant process by using the number of frequencies identified through the result of applying the second limit value. It may be to detect and diagnose the occurrence of oscillation for the control loop.

The commands, when executed by the at least one processor, cause the oscillation diagnosis device to determine that oscillation does not occur if the number of frequencies is 0 during the detection and diagnosis, and the number of frequencies is 1 , it may be determined that a single oscillation occurs, and if the number of frequencies exceeds 1, it may be determined that multiple oscillations occur.

The instructions, when executed by the at least one processor, may cause the oscillation diagnosis apparatus to undergo a pre-processing process of taking an absolute value from the FFT result value when extracting the peak value.

The present invention detects the occurrence of oscillation in a plant process control loop using time-series process data by automatically diagnosing oscillation and oscillation cycle regardless of single/multiple, stationary/non-stationary oscillation, and can be diagnosed.

In addition, the present invention detects and diagnoses the occurrence of oscillation using time-series process data of the plant, and can provide information on the frequency of oscillation as well as recognizing the state of oscillation.

In addition, the present invention can be used to analyze the correlation between complexly linked plant control loops by checking oscillation frequency information.

In addition, the present invention can contribute to fast plant stable operation by finding the control loop that is the root cause when oscillation occurs in multiple processes.

In addition, the present invention can automatically monitor and diagnose whether dozens or hundreds of control loops in a plant are oscillated and provide the information to a manager, thereby promoting plant stability and efficient operation.

1 is a diagram showing an example of oscillation;
2 is a diagram explaining an ACF damping ratio measurement method;
3 is a diagram explaining an ACF zero-passing period determination method;
4 is a diagram showing a multi-oscillation signal;
5 is a diagram showing ACF damping ratio determination results of a test signal;
6 is a diagram showing the result of ACF zero passing period determination of a test signal;
7a and 7b show an oscillation diagnosis method of a plant process control loop according to an embodiment of the present invention;
8 is a diagram showing a normal distribution and composition ratio;
9 is a diagram showing FFT result values of multiple oscillation signals in an oscillation diagnosis device;
10 is a diagram showing the result of extracting the maximum value from the FFT result value of FIG. 9;
11 is a view showing a result of applying a first limit value to the extraction result of FIG. 10;
12 is a view showing a result of applying a second limit value to the application result of FIG. 11;
13 is a diagram showing oscillation diagnosis results for various oscillation signals.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, detailed descriptions of well-known functions or configurations that may obscure the gist of the present invention will be omitted in the following description and accompanying drawings. In addition, it should be noted that the same components are indicated by the same reference numerals throughout the drawings as much as possible.

The terms or words used in this specification and claims described below should not be construed as being limited to ordinary or dictionary meanings, and the inventors are appropriately defined as terms for describing their invention in the best way. It should be interpreted as meaning and concept consistent with the technical idea of the present invention based on the principle that it can be done.

Therefore, the embodiments described in this specification and the configurations shown in the drawings are only one of the most preferred embodiments of the present invention, and do not represent all of the technical ideas of the present invention. It should be understood that there may be equivalents and variations.

In the accompanying drawings, some components are exaggerated, omitted, or schematically illustrated, and the size of each component does not entirely reflect the actual size. The present invention is not limited by the relative sizes or spacings drawn in the accompanying drawings.

When it is said that a certain part "includes" a certain component throughout the specification, it means that it may further include other components without excluding other components unless otherwise stated. In addition, when a part is said to be "connected" to another part, this includes not only the case of being "directly connected" but also the case of being "electrically connected" with another element interposed therebetween.

Singular expressions include plural expressions unless the context clearly dictates otherwise. Terms such as "comprise" or "having" are intended to indicate that a feature, number, step, operation, component, part, or combination thereof described in the specification exists, but that one or more other features, numbers, or steps are present. However, it should be understood that it does not preclude the possibility of existence or addition of operations, components, parts, or combinations thereof.

Also, the term "unit" used in the specification means a hardware component such as software, FPGA or ASIC, and "unit" performs certain roles. However, "unit" is not meant to be limited to software or hardware. A “unit” may be configured to reside in an addressable storage medium and may be configured to reproduce on one or more processors. Thus, as an example, “unit” can refer to components such as software components, object-oriented software components, class components and task components, processes, functions, properties, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays and variables. Functionality provided within components and "parts" may be combined into fewer components and "parts" or further separated into additional components and "parts".

Hereinafter, with reference to the accompanying drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily carry out the present invention. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings.

7A and 7B are oscillation diagnosis methods of a plant process control loop according to an embodiment of the present invention.

As shown in FIGS. 7A and 7B , the method for diagnosing oscillation of a plant process control loop according to an embodiment of the present invention automatically detects single/multiple and stationary/non-stationary oscillations. It detects and diagnoses the occurrence of oscillation in the plant process control loop using time-series process data by diagnosing whether or not the oscillation occurs and the oscillation cycle, and is performed by the oscillation diagnosis device.

Here, the oscillation diagnosis apparatus includes at least one processor and a memory for storing computer readable instructions. In this case, when the at least one processor executes computer readable instructions stored in the memory, the oscillation diagnosis apparatus performs the oscillation diagnosis method of the plant process control loop according to an embodiment of the present invention.

Such an oscillation diagnosis apparatus checks the periodicity of an oscillation signal through a Fast Fourier Transform (FFT) that converts a time domain signal into a frequency domain signal.

When an input signal oscillates at a specific frequency, the oscillation diagnosis apparatus detects a value corresponding to the corresponding frequency to know the occurrence of oscillation and the oscillation period.

In addition, the oscillation diagnosis device must detect and diagnose oscillation occurrence only with an input signal without a manager's intervention when automatically monitoring and diagnosing oscillation occurrence of tens or hundreds of plant process control loops.

To this end, the oscillation diagnosis device obtains a valid peak value of the FFT value of the input signal, delivers it to the manager, applies a limit value for extracting only the peak value determined to be oscillation, and returns the resultant value to the manager. report.

Hereinafter, a process of detecting and diagnosing the occurrence of oscillation in the oscillation diagnosis apparatus will be described in detail with reference to FIGS. 7A and 7B.

First, the oscillation diagnosis device receives process value data, which is a time domain signal, and performs FFT (S101). Here, the process value data is any one or more time series data selected from among single, multiple, stationary, and non-stationary, and is converted into a frequency domain signal through FFT.

Thereafter, the oscillation diagnosis device undergoes a preprocessing process of measuring the length of the process value data by taking an absolute value from the FFT result value in order to extract a peak value through the FFT result value, and extracting only the peak value determined as oscillation. First and second thresholds are set (S102).

Here, the first threshold value is applied to discard a frequency component having a small size. At this time, the first threshold is 'mean of the absolute value of the FFT result value' [mean(F(s))] and 'three times the standard deviation of the absolute value of the FFT result value' [3*std(F(s))] is automatically set to the sum of That is, the first threshold is the sum of the average and 3σ, which is mean(F(s))+3*std(F(s)).

Referring to FIG. 8 , in the normal distribution, 99.7% of all variables are probabilistically distributed within a section separated by 3 times the standard deviation (3σ) in a positive/negative direction centered on the mean. 8 is a diagram showing a normal distribution and a composition ratio.

Assuming that the process value data follows a Gaussian distribution (i.e., a normal distribution), when the first limit value is applied, the oscillation diagnosis device takes only the first limit value, that is, a value greater than 'average + 3σ' In this case, data of 0.15% of the total variables are acquired with probability. This means that when the data acquisition cycle is set to 1 second and the oscillation algorithm application cycle is set to 3600 seconds (1 hour), 4 to 5 pieces of data (ie, 3600 seconds * 0.15% = 5.4) are stochastically present.

And, the second threshold is applied to discard adjacent frequency components. At this time, the second limit value is an empirically extracted value, and a value of 0.001 to 0.003 Hz is automatically applied. Here, it is an extremely rare case that the oscillation frequency is within the interval of 0.001 to 0.003 Hz in the plant process.

However, even if oscillations having multiple frequencies occur with a difference within 0.001 to 0.003 Hz, the oscillation diagnosis device detects the most influential oscillation by applying the second limit value to diagnose the plant process control loop. .

Then, the oscillation diagnosis device extracts the peak value of the FFT result value (S103). That is, the oscillation diagnosis apparatus may obtain a local maximum (peak) of the absolute value of the FFT result value.

For example, if the x+1th value is greater than or equal to the xth value, the xth value is discarded because the xth value cannot be a local maximum. And, if the x+1 th value is smaller than the x th value, the x+1 th value is compared with the x+2 th value, and if the x+1 th value is greater than or equal to the x th value, the x+1 th value becomes the maximum value. Leave as is. However, in this case, if the x+1 th value is smaller than the x+2 th value, the x+1 th value cannot be the maximum value, so the x+1 th value is discarded. The oscillation diagnosis device repeats this process to extract the maximum value from the FFT result value.

Thereafter, the oscillation diagnosis apparatus derives a result of applying the first limit value for magnitude in order to obtain an effective oscillation period for the peak value of the FFT result value (S104).

That is, the oscillation diagnosis apparatus discards the x-th value when the x-th value is less than or equal to the first threshold value, and leaves it as it is when the x-th value is greater than the first threshold value. As such, the oscillation diagnosis device repeatedly performs this process.

Then, the oscillation diagnosis apparatus derives a result of applying a first threshold value to the peak value of the FFT result value and then applying a second threshold value for the frequency contiguous interval (S105).

First, the oscillation diagnosis apparatus finds the maximum value within the result of applying the first threshold value before applying the second threshold value. At this time, the oscillation diagnosis device stores the frequency and magnitude of the maximum value in new variables. Further, the oscillation diagnosis apparatus discards all values at a distance corresponding to the second threshold value in a positive/negative direction based on the maximum value by applying the second threshold value.

In this way, the oscillation diagnosis device finds a new maximum value and repeats this process.

Thereafter, the oscillation diagnosis device reports a diagnosis result to a manager based on the remaining data (S106).

That is, the oscillation diagnosis apparatus detects and diagnoses the occurrence of oscillation through the result of applying the second threshold value. At this time, the oscillation diagnosis apparatus can determine the number of remaining result values (the number of frequencies, hereinafter referred to as 'N_F(s)') because the aforementioned process is a process of extracting the oscillation period.

Specifically, the oscillation diagnosis apparatus determines that oscillation does not occur if N_F(s)=0. And, if N_F(s)=1, the oscillation diagnosis device determines that a single oscillation has occurred and reports the occurrence frequency. In addition, if N_F(s)>1, the oscillation diagnosis apparatus determines that multiple oscillations have occurred and reports the occurrence frequency.

Meanwhile, the oscillation diagnosis apparatus will be described with reference to FIGS. 9 to 12 to be described later for a simulation verification result using a multi-oscillation signal, which is a test signal of FIG. 4 . As described above, the multi-oscillation signal of FIG. 4 is composed of sine waves with three frequencies (0.159, 0.318, and 0.477 Hz) and a ramp signal with a period of 1800 seconds.

9 is a diagram showing FFT result values of multiple oscillation signals in the oscillation diagnosis device, FIG. 10 is a view showing the result of extracting the maximum value from the FFT result value of FIG. 9, and FIG. 11 is the extraction result of FIG. 12 is a diagram showing a result of applying a second threshold to the application result of FIG. 11 .

First, the oscillation diagnosis apparatus derives an FFT result value by performing FFT on multiple oscillation signals as shown in FIG. 9 . Here, the manager can diagnose the occurrence of multiple oscillations by recognizing four peak values with the naked eye, but cannot find out this only with data. Therefore, the oscillation diagnosis apparatus extracts the peak value of the FFT result value and applies the limit value.

And, the oscillation diagnosis apparatus extracts a maximum value or a peak value from the FFT result value as shown in FIG. 10 .

Thereafter, the oscillation diagnosis apparatus applies a first limit value of amplitude from the maximum value extracted from the FFT result value as shown in FIG. 11 . That is, although a number of peak values are shown in FIG. 10, the oscillation diagnosis apparatus discards values below the first threshold value in FIG. 11 and detects only five peak values.

Thereafter, the oscillation diagnosis apparatus applies a second threshold value to the result of applying the first threshold value as shown in FIG. 12 . 12 is a case in which both the first limit value and the second limit value are applied.

That is, in FIG. 11, the oscillation diagnosis device discards those whose peak value intervals are within 0.002 Hz and returns only those with larger intervals, and accurately diagnoses four frequencies corresponding to an input signal having four frequencies.

13 is a diagram showing oscillation diagnosis results for various oscillation signals.

Referring to FIG. 13 , the oscillation diagnosis apparatus generates single/multiple, stationary/non-stationary oscillation signals, which are various test signals, and shows results of performing oscillation diagnosis. As such, the oscillation diagnosis device generates 11 test signals, diagnoses the occurrence of oscillation, and confirms the result of the corresponding oscillation frequency. Here, test signals 9 to 13 in FIG. 13 correspond to non-stationary signals whose frequency changes with time.

The oscillation diagnosis apparatus may diagnose occurrence of single/multiple, stationary/non-stationary oscillation, and provide oscillation period information.

On the other hand, conventionally, single oscillation occurrence can be diagnosed, but it is difficult to diagnose multiple, abnormal oscillation occurrences, and it is difficult to provide oscillation period information. That is, conventionally, only stationary signals can be diagnosed, but the oscillation diagnosis apparatus according to an embodiment of the present invention can diagnose oscillation status and oscillation frequency even for non-stationary signals.

Methods according to some embodiments may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. Program instructions recorded on the medium may be those specially designed and configured for the present invention or those known and usable to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CDROMs and DVDs, and magnetic-optical media such as floptical disks. Included are hardware devices specially configured to store and execute program instructions, such as magneto-optical media and ROM, RAM, flash memory, and the like. Examples of program instructions include high-level language codes that can be executed by a computer using an interpreter, as well as machine language codes such as those produced by a compiler.

Although the foregoing description has been focused on the novel features of the present invention as applied to various embodiments, one skilled in the art will be able to understand the above-described apparatus and method without departing from the scope of the present invention. It will be appreciated that various deletions, substitutions, and changes in the form and detail of are possible. Accordingly, the scope of the present invention is defined by the appended claims rather than by the foregoing description. All modifications that come within the scope of equivalence of the claims are embraced by the scope of the present invention.

Claims (16)

In the oscillation diagnosis method of a plant process control loop using an oscillation diagnosis device,
Performing FFT (Fast Fourier Transform) on the process value data confirmed through the plant process control loop;
going through a preprocessing process of measuring the length of the process value data by taking an absolute value of the FFT result value for the process value data, and extracting a peak value from the FFT result value;
Deriving a result of applying a first limit value set as a sum of 'average of absolute values of FFT result values' and 'three times the standard deviation of absolute values of FFT result values' for magnitude in the extracted peak value; and
After deriving a result of applying the first threshold value, deriving a result of applying a second threshold value set to 0.001 to 0.003 Hz for an adjacent frequency component; and
Detecting and diagnosing the occurrence of oscillation in the plant process control loop by using the number of frequencies identified as a result of applying the second threshold,
The detection and diagnosis step,
If the number of frequencies is 0, it is determined that no oscillation has occurred, if the number of frequencies is 1, it is determined that a single oscillation has occurred, and if the number of frequencies is greater than 1, it is determined that multiple oscillations have occurred A method for diagnosing oscillation in an in-plant process control loop. delete delete delete According to claim 1,
The process value data,
A method for diagnosing oscillation of a plant process control loop, which is any one or more time series data selected from single, multiple, stationary and non-stationary. delete delete delete As an oscillation diagnostic device,
at least one processor; and
A memory for storing computer readable instructions;
The instructions, when executed by the at least one processor, cause the oscillation diagnosis device to:
FFT (Fast Fourier Transform) is performed on the process value data confirmed through the plant process control loop,
A preprocessing process of measuring the length of the process value data by taking an absolute value from the FFT result value for the process value data, extracting a peak value from the FFT result value,
A result obtained by applying a first threshold value set as the sum of 'the average of the absolute value of the FFT result value' and 'three times the standard deviation of the absolute value of the FFT result value' for the magnitude in the extracted peak value,
After deriving a result of applying the first limit value, deriving a result of applying a second limit value set to 0.001 to 0.003 Hz for an adjacent frequency component,
Detecting and diagnosing the occurrence of oscillation in the plant process control loop using the number of frequencies identified through the result of applying the second threshold,
In the detection and diagnosis, if the number of frequencies is 0, it is determined that no oscillation has occurred, if the number of frequencies is 1, it is determined that a single oscillation has occurred, and if the number of frequencies is greater than 1, it is determined that multiple oscillations have occurred. An oscillation diagnosis device that determines that this occurs. delete delete delete According to claim 9,
The process value data,
An oscillation diagnostic device that is any one or more time series data selected from single, multiple, stationary and non-stationary. delete delete delete

Images