CN117874685B - Pressure data abnormality early warning method for direct-current variable-frequency electric half ball valve - Google Patents
Pressure data abnormality early warning method for direct-current variable-frequency electric half ball valve Download PDFInfo
- Publication number
- CN117874685B CN117874685B CN202410268889.5A CN202410268889A CN117874685B CN 117874685 B CN117874685 B CN 117874685B CN 202410268889 A CN202410268889 A CN 202410268889A CN 117874685 B CN117874685 B CN 117874685B
- Authority
- CN
- China
- Prior art keywords
- data
- window
- preset
- value
- length
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 31
- 230000005856 abnormality Effects 0.000 title claims abstract description 18
- 238000013507 mapping Methods 0.000 claims description 12
- 230000003044 adaptive effect Effects 0.000 claims description 10
- 230000002159 abnormal effect Effects 0.000 description 8
- 230000000630 rising effect Effects 0.000 description 7
- 238000001514 detection method Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 239000012530 fluid Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Landscapes
- Testing And Monitoring For Control Systems (AREA)
- Measuring Fluid Pressure (AREA)
Abstract
The invention relates to the technical field of data processing, in particular to a pressure data abnormality early warning method of a direct-current variable-frequency electric half ball valve; obtaining the initial window length of the data point according to the data difference characteristic in the preset window in the pressure data sequence; obtaining window fitting degree and optimal window length according to the data fluctuation characteristics in the initial window length range of the data points; obtaining a pressure data denoising sequence according to the data characteristics within the optimal window length range; obtaining a change characteristic value and a target data point according to the change characteristic of the data in the pressure data denoising sequence; and obtaining the data change concentration according to the distribution of the target data points. According to the data change concentration degree and the change characteristic value, the preset pressure threshold value is adjusted to obtain the self-adaptive threshold value; and the pressure of the electric half ball valve is pre-warned according to the self-adaptive threshold value, so that the reliability and timeliness of pre-warning are improved.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a pressure data abnormality early warning method of a direct-current variable-frequency electric half ball valve.
Background
The direct-current variable-frequency electric half ball valve is a common key element in an industrial control system and is used for adjusting the flow of a fluid medium, and accurate control and response can be realized by driving the half ball valve through the direct-current variable-frequency motor; in an industrial automation process, the normal operation of the half ball valve is critical to maintaining the stability and safety of the system.
The semi-ball valve needs to monitor the pressure of the fluid medium in real time in the operation process so as to ensure that the system operates in a normal working range; the traditional pressure anomaly detection method is to denoise the collected pressure data through a moving average method and set a pressure threshold value, and early warning is carried out when the denoised pressure data exceeds the pressure threshold value. However, when denoising is performed by using a moving average method, because the length of a denoising window is fixed, the denoising effect is not ideal, and the early warning reliability is affected; and the fixed pressure threshold value may cause untimely early warning, and early warning can not be performed in advance according to the pressure change trend, so that the accuracy and reliability of pressure anomaly detection are low.
Disclosure of Invention
In order to solve the technical problems of low accuracy and reliability of pressure anomaly detection caused by denoising collected pressure data through a moving average method and setting a pressure threshold value for early warning, the invention aims to provide a pressure data anomaly early warning method of a direct-current variable-frequency electric half ball valve, which adopts the following specific technical scheme:
acquiring a pressure data sequence of the electric half ball valve; obtaining the initial window length of the data point according to the data difference characteristic in the preset window of the data point in the pressure data sequence;
obtaining window fitting degree of the data points according to the data fluctuation characteristics in the initial window length range of the data points and the preset adjacent data points; obtaining an optimal window length of the data points according to the window fit degree and the initial window length; obtaining a pressure data denoising sequence through a moving average algorithm according to data characteristics within the optimal window length range of the data points;
Obtaining a change characteristic value and a target data point according to the change characteristic of the data in the pressure data denoising sequence; obtaining a data change concentration degree according to the distance characteristic and the quantity characteristic of the target data points; adjusting a preset pressure threshold according to the data change concentration degree and the change characteristic value to obtain an adaptive threshold;
And pre-warning the pressure of the electric half ball valve according to the self-adaptive threshold value.
Further, the step of obtaining the initial window length of the data point according to the data difference characteristic in the preset window of the data point in the pressure data sequence includes:
Calculating a differential sequence of data in a preset window of the data points in the pressure data sequence to obtain a neighborhood fluctuation sequence of the data points; calculating variance of the neighborhood fluctuation sequence and carrying out negative correlation mapping to obtain a stability index of a data point;
When the stability index exceeds a preset first stability threshold value, calculating and normalizing the difference value between the stability index and the preset first stability threshold value to obtain a first adjustment coefficient; calculating the product of the first adjustment coefficient and the length of a preset adjustment window to obtain the adjustment length of the first window; calculating the sum of the first window adjustment length and the length of a preset window and rounding down to obtain the initial window length of a data point;
When the stability index does not exceed a preset second stability threshold, calculating and normalizing the difference value between the preset second stability threshold and the stability index to obtain a second adjustment coefficient; calculating the product of the second adjustment coefficient and the length of a preset adjustment window to obtain the adjustment length of the second window; calculating the difference between the length of the preset window and the adjustment length of the second window, and rounding downwards to obtain the initial window length of the data point;
And when the stability index is in a closed interval of a preset second stability threshold value and a preset first stability threshold value, the length of the preset window is the initial window length of the data point.
Further, the step of obtaining window fitness of the data point according to the data fluctuation characteristics in the initial window length range of the data point and the preset adjacent data point comprises the following steps:
Calculating the variance of the data value in the initial window length of the data point to obtain the fluctuation characteristic value of the data point; calculating the variance of the fluctuation characteristic value of the preset adjacent data point corresponding to the data point to obtain the adjacent fluctuation degree; calculating variances of fluctuation characteristic values of the data points and preset adjacent data points to obtain overall fluctuation degrees;
Calculating the absolute value of the difference between the integral fluctuation degree and the adjacent fluctuation degree and carrying out negative correlation mapping to obtain an adjacent fluctuation difference value; mapping the negative correlation of the overall fluctuation degree to obtain an overall fluctuation characteristic value; and calculating the product of the adjacent fluctuation difference value and the integral fluctuation characteristic value to obtain the window fit degree of the data point.
Further, the step of obtaining an optimal window length for a data point based on the window fit and the initial window length comprises:
when the window fit degree of the data point exceeds a preset proper threshold value, taking the initial window length of the data point as the optimal window length of the data point;
And when the window fit degree of the data point does not exceed a preset proper threshold value, taking the window adjustment value as the optimal window length of the data point.
Further, the step of obtaining the window adjustment value includes:
Calculating the ratio of the window fit degree of a data point to the preset proper threshold value to obtain a window adjustment coefficient of the data point;
When the initial window length of the data point exceeds the length of a preset window, calculating the difference between the initial window length and the length of the preset window to obtain a window adjustment reference; calculating the product of the window adjustment reference and the window adjustment coefficient to obtain the adjustment length of the data point, calculating the sum of the adjustment length and the length of a preset window, and rounding down to obtain the window adjustment value of the data point;
when the initial window length of the data point does not exceed the length of a preset window, calculating the product of the initial window length and the window adjustment coefficient and rounding down to obtain the window adjustment value of the data point.
Further, the step of obtaining the pressure data denoising sequence according to the data characteristics within the optimal window length range of the data points through a sliding average algorithm comprises the following steps:
obtaining denoising data points of the data points through a moving average algorithm according to the data in the optimal window length range of the data points; and constructing denoising data points corresponding to each data point in the pressure data sequence according to the sequence to obtain the pressure data denoising sequence.
Further, the step of obtaining a variation characteristic value and a target data point according to the variation characteristic of the data in the pressure data denoising sequence comprises the following steps:
Calculating a differential sequence of the pressure data denoising sequence to obtain a change characteristic sequence; taking the mode with the numerical value in the positive number in the change characteristic sequence as a change characteristic value; and taking the data points of the positive numbers in the change characteristic sequence at the corresponding positions in the pressure data denoising sequence as target data points.
Further, the step of obtaining the data change concentration according to the distance characteristic and the quantity characteristic of the target data points comprises the following steps:
Calculating the quantity ratio of the data points in the pressure data denoising sequence to the target data points, and obtaining a fluctuation average distance value; calculating the average value of the time distances between any adjacent target data points to obtain a fluctuation distance characteristic value; and calculating and normalizing the difference value between the fluctuation average distance value and the fluctuation distance characteristic value to obtain the data change concentration.
Further, the step of adjusting the preset pressure threshold according to the data change concentration and the change characteristic value to obtain an adaptive threshold includes:
Calculating the ratio of the variation characteristic value to the maximum value in the variation characteristic sequence and carrying out negative correlation mapping to obtain a threshold adjustment weight; obtaining an adaptive adjustment value according to the product of the data change concentration, the threshold adjustment weight and a preset threshold adjustment amount; and calculating the difference value between the preset pressure threshold value and the self-adaptive adjustment value to obtain the self-adaptive threshold value.
The invention has the following beneficial effects:
according to the method, the initial window length is obtained, the denoising window can be determined according to the fluctuation characteristics of the data points in the pressure data sequence in the preset window, and the denoising accuracy and the prediction reliability are improved; the window fitting degree can be obtained to correct the initial window length according to the change characteristics of the fluctuation characteristics of the pressure data sequence, and the data denoising accuracy is further improved. According to the data characteristics in the optimal window length of the data points, the pressure data denoising sequence is obtained through moving average denoising, so that the abnormal fluctuation characteristics of the data can be highlighted while the data noise is reduced, and the early warning reliability is improved. The method comprises the steps of obtaining a change characteristic value and a target data point, wherein the change characteristic value and the target data point can represent the pressure rising degree and the position in a pressure data denoising sequence, and obtaining the data change concentration degree can represent the abnormal rapid rising degree of the pressure data. The pressure of the electric half ball valve is pre-warned according to the self-adaptive threshold value, the pressure threshold value can be adaptively adjusted according to the pressure change characteristic, and the accuracy, reliability and timeliness of the pressure abnormality pre-warning are improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for early warning of pressure data abnormality of a direct-current variable-frequency electric half ball valve according to an embodiment of the present invention.
Detailed Description
In order to further explain the technical means and effects adopted by the invention to achieve the preset aim, the following is a specific implementation, structure, characteristics and effects of the pressure data abnormality pre-warning method for the direct-current variable-frequency electric half-ball valve according to the invention, which are described in detail below with reference to the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The specific scheme of the pressure data abnormality early warning method of the direct-current variable-frequency electric half ball valve provided by the invention is specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flow chart of a pressure data anomaly early warning method of a direct current variable frequency electric half ball valve according to an embodiment of the invention is shown, and the method comprises the following steps:
Step S1, acquiring a pressure data sequence of an electric half ball valve; and obtaining the initial window length of the data point according to the data difference characteristic in the preset window of the data point in the pressure data sequence.
In the embodiment of the invention, the implementation scene is to denoise and detect the abnormality of the pressure data of the direct-current variable-frequency electric half ball valve; the traditional pressure anomaly detection method is to denoise the collected pressure data through a moving average method and set a pressure threshold value, and early warning is carried out when the denoised pressure data exceeds the pressure threshold value. However, when denoising is performed by using a moving average method, because the length of a denoising window is fixed, the denoising effect is not ideal, and the early warning reliability is affected; and the fixed pressure threshold value can lead to untimely early warning and can not early warn according to the pressure change trend. Therefore, in order to improve the accuracy and reliability of early warning, when the pressure has abnormal fluctuation, the early warning is timely carried out, the adjacent historical data at the current moment can be obtained, and the pressure threshold value is adjusted according to the change characteristics of the adjacent historical data.
Firstly, acquiring a pressure data sequence of an electric half ball valve, acquiring pressure data of the half ball valve through a pressure sensor, constructing the sequence, and enabling an implementer to automatically determine acquisition frequency and acquisition positions in an implementation scene according to the implementation scene; the pressure data sequence refers to the range of adjacent historical data at the current moment, the last bit of the data at the current moment in the historical data sequence reflects the adjacent historical pressure change characteristics at the current moment, and each acquisition moment corresponds to one pressure data sequence. In the embodiment of the invention, the length of the historical data sequence is 300 data points, an implementer can determine the length according to an implementation scene by himself, when the pressure data sequence corresponding to any moment does not meet the length in the initial stage of acquisition, a fixed preset pressure threshold value is used for early warning, and when the pressure exceeds the preset pressure threshold value, early warning is directly carried out, and the implementer can determine the length according to the implementation scene by himself. When the historical data sequence corresponding to any time meets the length, the pressure threshold value can be adjusted according to the change characteristics of the historical data sequence in order to improve the accuracy and the reliability of early warning at any time.
Before pressure early warning, the pressure data sequence needs to be denoised so as to improve early warning reliability, the moving average method is a traditional denoising method, denoising is realized by calculating a data average value in a denoising window, the denoising accuracy is influenced by the size of the denoising window, the data is smooth and excessive due to a larger denoising window, the data change trend is difficult to find in time, the denoising effect is possibly influenced by a smaller denoising window, and early warning errors occur; the initial window length of the data points can be obtained according to the data difference characteristics in the preset window of the data points in the pressure data sequence.
Preferably, in an embodiment of the present invention, acquiring the initial window length includes: calculating a differential sequence of data in a preset window of data points in the pressure data sequence, and obtaining a neighborhood fluctuation sequence of the data points; it should be noted that, the data point is taken as the end point of the window, the window length with the data length of 10 is taken as the preset window of the data point, and the implementer can determine according to the implementation scene; in the embodiment of the present invention, if the initial position of the data point in the pressure data sequence is the initial position, in order to meet the calculation condition, other pressure data adjacent to the historical time of the pressure data sequence may be used to participate in the calculation. The step of obtaining the differential sequence belongs to the prior art, and specific steps are not repeated. Calculating variance of the neighborhood fluctuation sequence and carrying out negative correlation mapping to obtain a stability index of the data point; the larger the data difference in the neighborhood fluctuation sequence, the more obvious the data fluctuation in the preset window of the data point, and the smaller the stationary index. When the fluctuation characteristic of the data in the preset window of the data point is weaker, longer window data can be used for carrying out moving average on the data point; when the fluctuation characteristic of the data in the preset window of the data point is stronger, smaller window data is needed to carry out moving average on the data point, so that excessive smoothing of the data is avoided, and the accuracy of abnormal early warning is reduced. When the stability index exceeds a preset first stability threshold value, calculating and normalizing the difference value between the stability index and the preset first stability threshold value to obtain a first adjustment coefficient; in the embodiment of the invention, the first stability threshold value is preset to be 0.7, and an implementer can determine according to an implementation scene; the larger the first adjustment coefficient, the larger the window meaning the moving average. Calculating the product of the first adjustment coefficient and the length of a preset adjustment window to obtain the adjustment length of the first window; calculating the sum of the first window adjustment length and the length of the preset window, and rounding down to obtain the initial window length of the data point; in the embodiment of the invention, the length of the preset adjusting window is 0.5 times of the length of the preset window. When the stability index does not exceed the preset second stability threshold, calculating and normalizing the difference value between the preset second stability threshold and the stability index to obtain a second adjustment coefficient; when the second adjustment coefficient is larger, this means that the smaller the stationary index is, the smaller window data is required for the moving average. Calculating the product of the second adjustment coefficient and the length of the preset adjustment window to obtain the adjustment length of the second window; calculating the difference between the length of the preset window and the adjustment length of the second window, and rounding down to obtain the initial window length of the data point; in the embodiment of the invention, the preset second stability threshold is 0.3, and the implementer can determine according to the implementation scene. When the stability index is in a closed interval of the preset second stability threshold value and the preset first stability threshold value, the length of the preset window is the initial window length of the data point. The more obvious the data fluctuation characteristic in the preset window of the data point is, the smaller the corresponding initial window length is, and the larger the corresponding initial window length is, otherwise the more the data fluctuation characteristic in the preset window of the data point is.
Step S2, obtaining window fitting degree of the data points according to the data points and the data fluctuation characteristics in the initial window length range of the preset adjacent data points; obtaining an optimal window length of the data points according to the window fit degree and the initial window length; and obtaining a pressure data denoising sequence through a moving average algorithm according to the data characteristics within the optimal window length range of the data points.
The initial window length is only obtained according to the data fluctuation characteristics in the preset window of the data point, and the pressure data sequence has certain change and trend, if the preset window of the data point and the data fluctuation characteristics in the adjacent preset window have certain difference, the initial window length needs to be adjusted, and the accuracy of the sliding average is improved. For example, when the data fluctuation characteristic in the initial window length of a data point is larger than the data fluctuation characteristic in the initial window length of an adjacent data point, which means that the data fluctuation at the moment is abnormal, the initial window length needs to be further reduced, the smoothness is reduced, the fluctuation characteristic near the data point is further highlighted, and if the fluctuation characteristic in the initial window length of the data point is similar to the fluctuation characteristic in the initial window length of the adjacent data point, the initial window length does not need to be reduced. The window appropriateness of a data point can be obtained from the data fluctuation characteristics in the initial window length range of the data point and the preset adjacent data point.
Preferably, in one embodiment of the present invention, obtaining window fitness includes: calculating the variance of the data value in the initial window length of the data point to obtain the fluctuation characteristic value of the data point; the larger the fluctuation feature value, the larger the data value difference within the initial window length of the data point. Calculating the variance of the fluctuation characteristic value of the preset adjacent data point corresponding to the data point to obtain the adjacent fluctuation degree; in the embodiment of the invention, the preset adjacent data point is the first 6 data points of the data point, an implementer can determine according to an implementation scene, the adjacent fluctuation degree is the variance of fluctuation characteristic values of 6 data points in the preset adjacent data points, and when the adjacent fluctuation degree is smaller, the data fluctuation characteristics in the initial window length corresponding to the preset adjacent data point are more similar; the fluctuation characteristic changes less, and the data change characteristic has a certain rule. Calculating variances of fluctuation characteristic values of the data points and preset adjacent data points to obtain overall fluctuation degrees; the overall fluctuation degree characterizes the difference degree of the data fluctuation characteristics in the initial window length of the data point and the preset adjacent data point; the larger the overall waviness, the larger the waviness characteristic change, and the more obvious the data waviness characteristic in the initial window length of the data point is compared with the data waviness characteristic in the initial window length of the preset adjacent data point.
Calculating the absolute value of the difference between the overall fluctuation degree and the adjacent fluctuation degree and carrying out negative correlation mapping to obtain an adjacent fluctuation difference value; when the difference between the overall waviness and the adjacent waviness is larger, the adjacent waviness difference value is smaller. Mapping the negative correlation of the overall fluctuation degree to obtain an overall fluctuation characteristic value; when the overall fluctuation feature value is smaller, the degree of difference of the data fluctuation feature in the initial window length of the data point and the preset adjacent data point is larger, and the difference of the data fluctuation feature in the initial window length of the data point and the initial window length of the preset adjacent data point is larger. Calculating the product of the adjacent fluctuation difference value and the integral fluctuation characteristic value to obtain the window fitting degree of the data point; when the window fit is smaller, the data fluctuation characteristic in the initial window length of the data point is more obvious than the data fluctuation characteristic in the initial window length of the preset adjacent data point; the initial window length of the data point needs to be changed, and the denoising and abnormality early warning accuracy is improved.
Further, the optimal window length of the data points can be obtained according to the window fit degree and the initial window length, and the method specifically comprises the following steps: when the window suitability of the data point exceeds a preset suitable threshold, the data in the initial window length of the data point can be accurately denoised, and the initial window length of the data point is taken as the optimal window length of the data point; in the embodiment of the invention, a proper threshold value is preset to be 0.6, and an implementer can determine according to implementation scenes. When the window fit of the data point does not exceed the preset proper threshold, the window adjustment value is taken as the optimal window length of the data point.
The step of obtaining the window adjustment value comprises the following steps: calculating the ratio of the window fitting degree of the data point to a preset proper threshold value to obtain a window adjustment coefficient of the data point; when the window adjustment coefficient is smaller, this means that the window length needs to be adjusted to a greater extent. When the initial window length of the data point exceeds the length of the preset window, calculating the difference between the initial window length and the length of the preset window to obtain a window adjustment reference; calculating the product of a window adjustment reference and a window adjustment coefficient to obtain the adjustment length of the data point, calculating the sum of the adjustment length and the length of a preset window, and rounding down to obtain the window adjustment value of the data point; when the initial window length of the data point does not exceed the length of a preset window, calculating the product of the initial window length and the window adjustment coefficient and rounding downwards to obtain a window adjustment value of the data point; the window adjustment value is less than the initial window length of the data point so that the denoised value of the data point is more reflective of the fluctuation characteristics in the vicinity of the data point.
Further, after the optimal window length of the data point is obtained for sliding average, the pressure data denoising sequence can be obtained through a sliding average algorithm according to the data characteristics within the optimal window length range of the data point, which specifically comprises the following steps: obtaining a denoising data point of the data point through a moving average algorithm according to the data in the optimal window length range of the data point; constructing denoising data points corresponding to each data point in the pressure data sequence according to the sequence to obtain a pressure data denoising sequence; it should be noted that, the sliding average method belongs to the prior art, and specific calculation steps are not repeated.
S3, obtaining a change characteristic value and a target data point according to the change characteristic of the data in the pressure data denoising sequence; obtaining a data change concentration degree according to the distance characteristic and the quantity characteristic of the target data points; and adjusting the preset pressure threshold according to the data change concentration degree and the change characteristic value to obtain the self-adaptive threshold.
The pressure data denoising sequence is used for pressure abnormality early warning, errors in early warning caused by noise data can be avoided, and early warning accuracy is improved. When the change degree of the pressure data is larger and the pressure value is more and more close to the preset pressure threshold value, the possibility of larger abnormal operation of the half ball valve is indicated, if the fixed preset pressure threshold value is used, early warning is possibly not timely, and detection is late, so that in order to further improve the early warning reliability, the self-adaptive adjustment of the pressure threshold value can be carried out according to the data change characteristics in the pressure data denoising sequence. Firstly, according to the change characteristics of data in the pressure data denoising sequence, a change characteristic value and a target data point are obtained.
Preferably, in one embodiment of the present invention, acquiring the change feature value and the target data point includes: calculating a differential sequence of the pressure data denoising sequence to obtain a change characteristic sequence; the differential sequence obtaining step belongs to the prior art, the specific obtaining step is not repeated, and the change characteristic sequence reflects the change degree of the adjacent moment data in the pressure data denoising sequence. Taking the mode with the numerical value in the change characteristic sequence being positive as a change characteristic value; the change characteristic value represents the most common value of the pressure rising degree at the adjacent moment, and the smaller the change characteristic value is, the smaller the pressure rising degree is, the more common is the pressure stable operation condition in the hemispherical valve work. And taking the data point of the positive number in the change characteristic sequence at the corresponding position in the pressure data denoising sequence as a target data point, wherein the target data point represents the data point at the moment when the pressure rise occurs, for example, if the difference value of the data points obtained at the third acquisition moment and the second acquisition moment in the pressure data denoising sequence is positive, the data point at the third acquisition moment is the target data point.
Further, when the time interval of occurrence of the target data points is shorter and the target data points are more, which means that the pressure data has abnormal rising trend in a short period, the data change concentration degree can be obtained according to the distance characteristic and the quantity characteristic of the target data points, which specifically comprises: calculating the quantity ratio of data points in the pressure data denoising sequence to the target data points, and obtaining a fluctuation average distance value; when the fluctuation average distance representation value is smaller, which means that the number of target data points is larger, in the embodiment of the invention, the time distance between two adjacent acquisition moments is set to be 1 in unit length, and then the ratio of the number of data points in the pressure data denoising sequence to the number of target data points can reflect the average time distance between the target data points. Calculating the average value of the time distances between any adjacent target data points to obtain a fluctuation distance characteristic value; the smaller the fluctuation distance characteristic value is, the more concentrated the distribution of the target data points is. And calculating and normalizing the difference between the fluctuation average distance value and the fluctuation distance characteristic value to obtain the data change concentration, wherein when the fluctuation distance characteristic value is smaller than the fluctuation average distance value, the distribution of target data points is more concentrated, the possibility of abnormal rapid rising of pressure is more likely to occur, and the data change concentration is larger.
After the data change concentration of the pressure data denoising sequence is obtained, the preset pressure threshold value can be adjusted according to the data change concentration and the change characteristic value to obtain the self-adaptive threshold value, which comprises the following steps: calculating the ratio of the variation characteristic value to the maximum value in the variation characteristic sequence and carrying out negative correlation mapping to obtain a threshold adjustment weight; when the variation characteristic value is smaller than the numerical value maximum value in the variation characteristic sequence, the threshold value adjusting weight is larger, which means that the pressure data is normally changed stably, when the pressure data is increased abnormally and rapidly, the preset pressure threshold value is required to be reduced to a larger extent, and then the early warning reliability and response speed are improved; on the contrary, when the threshold value adjusting weight is smaller, the change characteristic value is larger, and the degree of pressure rising in normal operation is also larger, so that when pressure data rises, the preset pressure threshold value does not need to be reduced to a larger degree, and the early warning is avoided being excessively sensitive and the early warning error is generated. Obtaining a self-adaptive adjustment value according to the product of the data change concentration, the threshold adjustment weight and the preset threshold adjustment amount; the preset threshold adjustment amount can be determined by an implementer according to the pressure characteristics in an implementation scene, so as to adjust the change amount of the pressure threshold; when the data change concentration and the threshold value adjustment weight are larger, the self-adaptive adjustment value is larger, the preset pressure threshold value is required to be reduced, the pressure abnormality early warning is timely carried out, the loss caused by the abnormality is reduced, and the early warning reliability is improved. And calculating a difference value between the preset pressure threshold value and the self-adaptive adjustment value to obtain the self-adaptive threshold value. When the pressure data rises abnormally and rapidly, the smaller the self-adaptive threshold value is, the more early warning can be improved, and the reliability of the early warning is improved; the formula for obtaining the adaptive threshold includes:
In the method, in the process of the invention, Representing the value of the adaptive threshold value,Indicating that the preset pressure threshold value is to be set,Indicating the preset threshold adjustment amount,Representing the concentration of the data changes,The characteristic value of the change is indicated,Represents the maximum value of the values in the sequence of variation characteristics,Represents an exponential function with a base of a natural constant,Representing the threshold adjustment weight value,Representing the adaptive adjustment value.
And S4, early warning is carried out on the pressure of the electric half ball valve according to the self-adaptive threshold value.
After the self-adaptive threshold is obtained, the pressure of the electric half ball valve can be pre-warned according to the self-adaptive threshold, and the self-adaptive threshold is a pressure threshold corresponding to a data point at the last moment in the pressure data sequence, and when the pressure data after denoising of the data point at the last moment exceeds the self-adaptive threshold, the pressure abnormality pre-warning is carried out. The pressure data sequences corresponding to different acquisition moments can be calculated to obtain an adaptive threshold value as a pressure early warning threshold value of the acquisition moment, so that the accuracy reliability and timeliness of the pressure early warning are improved.
In summary, the embodiment of the invention provides a pressure data abnormality early warning method of a direct-current variable-frequency electric half ball valve; obtaining the initial window length of the data point according to the data difference characteristic in the preset window in the pressure data sequence; obtaining window fitting degree and optimal window length according to the data fluctuation characteristics in the initial window length range of the data points; obtaining a pressure data denoising sequence according to the data characteristics within the optimal window length range; obtaining a change characteristic value and a target data point according to the change characteristic of the data in the pressure data denoising sequence; and obtaining the data change concentration according to the distribution of the target data points. According to the data change concentration degree and the change characteristic value, the preset pressure threshold value is adjusted to obtain the self-adaptive threshold value; and the pressure of the electric half ball valve is pre-warned according to the self-adaptive threshold value, so that the reliability and timeliness of pre-warning are improved.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
Claims (4)
1. The pressure data abnormality early warning method of the direct-current variable-frequency electric half ball valve is characterized by comprising the following steps of:
acquiring a pressure data sequence of the electric half ball valve; obtaining the initial window length of the data point according to the data difference characteristic in the preset window of the data point in the pressure data sequence;
obtaining window fitting degree of the data points according to the data fluctuation characteristics in the initial window length range of the data points and the preset adjacent data points; obtaining an optimal window length of the data points according to the window fit degree and the initial window length; obtaining a pressure data denoising sequence through a moving average algorithm according to data characteristics within the optimal window length range of the data points;
Obtaining a change characteristic value and a target data point according to the change characteristic of the data in the pressure data denoising sequence; obtaining a data change concentration degree according to the distance characteristic and the quantity characteristic of the target data points; adjusting a preset pressure threshold according to the data change concentration degree and the change characteristic value to obtain an adaptive threshold;
early warning is carried out on the pressure of the electric half ball valve according to the self-adaptive threshold value;
The step of obtaining window fitness of the data point according to the data fluctuation characteristics in the initial window length range of the data point and the preset adjacent data point comprises the following steps:
Calculating the variance of the data value in the initial window length of the data point to obtain the fluctuation characteristic value of the data point; calculating the variance of the fluctuation characteristic value of the preset adjacent data point corresponding to the data point to obtain the adjacent fluctuation degree; calculating variances of fluctuation characteristic values of the data points and preset adjacent data points to obtain overall fluctuation degrees;
calculating the absolute value of the difference between the integral fluctuation degree and the adjacent fluctuation degree and carrying out negative correlation mapping to obtain an adjacent fluctuation difference value; mapping the negative correlation of the overall fluctuation degree to obtain an overall fluctuation characteristic value; calculating the product of the adjacent fluctuation difference value and the integral fluctuation characteristic value to obtain the window fit degree of the data point;
The step of obtaining an optimal window length for a data point based on the window fit and the initial window length comprises:
when the window fit degree of the data point exceeds a preset proper threshold value, taking the initial window length of the data point as the optimal window length of the data point;
When the window fit degree of the data point does not exceed a preset proper threshold value, taking the window adjustment value as the optimal window length of the data point;
the step of obtaining the window adjustment value includes:
Calculating the ratio of the window fit degree of a data point to the preset proper threshold value to obtain a window adjustment coefficient of the data point;
When the initial window length of the data point exceeds the length of a preset window, calculating the difference between the initial window length and the length of the preset window to obtain a window adjustment reference; calculating the product of the window adjustment reference and the window adjustment coefficient to obtain the adjustment length of the data point, calculating the sum of the adjustment length and the length of a preset window, and rounding down to obtain the window adjustment value of the data point;
When the initial window length of the data point does not exceed the length of a preset window, calculating the product of the initial window length and the window adjustment coefficient and rounding downwards to obtain a window adjustment value of the data point;
the step of obtaining a change characteristic value and a target data point according to the change characteristic of the data in the pressure data denoising sequence comprises the following steps:
Calculating a differential sequence of the pressure data denoising sequence to obtain a change characteristic sequence; taking the mode with the numerical value in the positive number in the change characteristic sequence as a change characteristic value; taking data points of positive numbers in the change characteristic sequence at corresponding positions in the pressure data denoising sequence as target data points;
The step of adjusting the preset pressure threshold according to the data change concentration degree and the change characteristic value to obtain an adaptive threshold comprises the following steps:
Calculating the ratio of the variation characteristic value to the maximum value in the variation characteristic sequence and carrying out negative correlation mapping to obtain a threshold adjustment weight; obtaining an adaptive adjustment value according to the product of the data change concentration, the threshold adjustment weight and a preset threshold adjustment amount; and calculating the difference value between the preset pressure threshold value and the self-adaptive adjustment value to obtain the self-adaptive threshold value.
2. The method for early warning of pressure data abnormality of a direct current variable frequency electric half ball valve according to claim 1, wherein the step of obtaining an initial window length of a data point according to a data difference characteristic in a preset window of the data point in the pressure data sequence comprises the steps of:
Calculating a differential sequence of data in a preset window of the data points in the pressure data sequence to obtain a neighborhood fluctuation sequence of the data points; calculating variance of the neighborhood fluctuation sequence and carrying out negative correlation mapping to obtain a stability index of a data point;
When the stability index exceeds a preset first stability threshold value, calculating and normalizing the difference value between the stability index and the preset first stability threshold value to obtain a first adjustment coefficient; calculating the product of the first adjustment coefficient and the length of a preset adjustment window to obtain the adjustment length of the first window; calculating the sum of the first window adjustment length and the length of a preset window and rounding down to obtain the initial window length of a data point;
When the stability index does not exceed a preset second stability threshold, calculating and normalizing the difference value between the preset second stability threshold and the stability index to obtain a second adjustment coefficient; calculating the product of the second adjustment coefficient and the length of a preset adjustment window to obtain the adjustment length of the second window; calculating the difference between the length of the preset window and the adjustment length of the second window, and rounding downwards to obtain the initial window length of the data point;
And when the stability index is in a closed interval of a preset second stability threshold value and a preset first stability threshold value, the length of the preset window is the initial window length of the data point.
3. The pressure data anomaly early warning method of a direct current variable frequency electric half ball valve according to claim 1, wherein the step of obtaining a pressure data denoising sequence through a moving average algorithm according to data characteristics within an optimal window length range of data points comprises the following steps:
obtaining denoising data points of the data points through a moving average algorithm according to the data in the optimal window length range of the data points; and constructing denoising data points corresponding to each data point in the pressure data sequence according to the sequence to obtain the pressure data denoising sequence.
4. The method for early warning of pressure data abnormality of a direct current variable frequency electric half ball valve according to claim 1, wherein the step of obtaining the data change concentration according to the distance characteristic and the number characteristic of the target data points comprises the following steps:
Calculating the quantity ratio of the data points in the pressure data denoising sequence to the target data points, and obtaining a fluctuation average distance value; calculating the average value of the time distances between any adjacent target data points to obtain a fluctuation distance characteristic value; and calculating and normalizing the difference value between the fluctuation average distance value and the fluctuation distance characteristic value to obtain the data change concentration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410268889.5A CN117874685B (en) | 2024-03-11 | 2024-03-11 | Pressure data abnormality early warning method for direct-current variable-frequency electric half ball valve |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410268889.5A CN117874685B (en) | 2024-03-11 | 2024-03-11 | Pressure data abnormality early warning method for direct-current variable-frequency electric half ball valve |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117874685A CN117874685A (en) | 2024-04-12 |
| CN117874685B true CN117874685B (en) | 2024-05-24 |
Family
ID=90579693
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410268889.5A Active CN117874685B (en) | 2024-03-11 | 2024-03-11 | Pressure data abnormality early warning method for direct-current variable-frequency electric half ball valve |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117874685B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116702081A (en) * | 2023-08-07 | 2023-09-05 | 西安格蒂电力有限公司 | Intelligent inspection method for power distribution equipment based on artificial intelligence |
| CN117272479A (en) * | 2023-10-08 | 2023-12-22 | 山东鑫之源新材料科技有限公司 | High-strength geomembrane bursting strength prediction method based on load time course analysis |
| CN117368920A (en) * | 2023-12-06 | 2024-01-09 | 山东三矿地质勘查有限公司 | D-insar-based coal mining area subsidence monitoring method and system |
| CN117390573A (en) * | 2023-12-12 | 2024-01-12 | 深圳前海慧联科技发展有限公司 | Wind turbine generator operation abnormality early warning method based on time sequence prediction |
| CN117411189A (en) * | 2023-12-14 | 2024-01-16 | 山东德源电力科技股份有限公司 | Monitoring data enhancement method of micro-grid coordination controller |
| CN117407700A (en) * | 2023-12-14 | 2024-01-16 | 国网山东省电力公司莱芜供电公司 | Method for monitoring working environment in live working process |
-
2024
- 2024-03-11 CN CN202410268889.5A patent/CN117874685B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116702081A (en) * | 2023-08-07 | 2023-09-05 | 西安格蒂电力有限公司 | Intelligent inspection method for power distribution equipment based on artificial intelligence |
| CN117272479A (en) * | 2023-10-08 | 2023-12-22 | 山东鑫之源新材料科技有限公司 | High-strength geomembrane bursting strength prediction method based on load time course analysis |
| CN117368920A (en) * | 2023-12-06 | 2024-01-09 | 山东三矿地质勘查有限公司 | D-insar-based coal mining area subsidence monitoring method and system |
| CN117390573A (en) * | 2023-12-12 | 2024-01-12 | 深圳前海慧联科技发展有限公司 | Wind turbine generator operation abnormality early warning method based on time sequence prediction |
| CN117411189A (en) * | 2023-12-14 | 2024-01-16 | 山东德源电力科技股份有限公司 | Monitoring data enhancement method of micro-grid coordination controller |
| CN117407700A (en) * | 2023-12-14 | 2024-01-16 | 国网山东省电力公司莱芜供电公司 | Method for monitoring working environment in live working process |
Non-Patent Citations (1)
| Title |
|---|
| Ladislav Stanke,et al..Towards to Optimal Wavelet Denoising Scheme—ANovelSpatial and Volumetric Mapping of Wavelet-Based Biomedical Data Smoothing.sensors.2020,全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117874685A (en) | 2024-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117278643B (en) | Vehicle-mounted cloud calibration data transmission system based on cloud edge cooperation | |
| CN116243097B (en) | Electric energy quality detection method based on big data | |
| CN112629585A (en) | Equipment on-line monitoring method and device based on multi-dimensional parameter estimation | |
| CN117407700B (en) | Method for monitoring working environment in live working process | |
| CN117421610B (en) | Data anomaly analysis method for electric energy meter running state early warning | |
| CN116304956B (en) | Chip temperature anomaly online detection method | |
| CN117708748B (en) | Operation monitoring system and method for centrifugal fan | |
| CN117786325B (en) | Ambient temperature wisdom early warning system of heavy-calibre thing networking water gauge | |
| CN116717734B (en) | Data monitoring method for steady-state operation of natural gas pipeline | |
| CN116451014B (en) | Real-time monitoring method for temperature of movable part in wireless passive equipment | |
| CN117874685B (en) | Pressure data abnormality early warning method for direct-current variable-frequency electric half ball valve | |
| CN114638435A (en) | Diesel engine security parameter prediction method based on data driving | |
| EP2522855B1 (en) | Displacement detection device for variable displacement compressor, and variable capacity compressor equipped with same | |
| CN117092980B (en) | Electrical fault detection control system based on big data | |
| CN116204805B (en) | Micro-pressure oxygen cabin and data management system | |
| CN116662111B (en) | Intelligent network management control system for train | |
| CN117404283A (en) | Control method and device for nitrogen-oxygen sensor pump unit | |
| JP6116398B2 (en) | Waveform estimation apparatus and waveform estimation method | |
| CN116448263A (en) | Method for detecting running state of boehmite production equipment | |
| CN113704323A (en) | Timing sequence anomaly detection algorithm based on HTM improvement | |
| CN107909677A (en) | A kind of marine mechanical and electrical equipment operation data stability extraction and method of discrimination | |
| CN118090078B (en) | Leakage online monitoring method for closed circulation water cooling system | |
| CN118211082A (en) | Oil filter element residual life prediction method and system based on data analysis | |
| CN117874445B (en) | Enzyme preparation production monitoring method for real-time online monitoring data analysis | |
| CN118091325B (en) | Intelligent cable detection method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |