A method for predicting evaporator failures for evaporating materials

Abstract

The present application relates to the technical field of evaporator operation data processing, and specifically to an evaporator fault prediction method for evaporating materials, the method comprising: collecting sound signals, pressure sequences and flow sequences at various positions of the evaporator valve; segmenting the sound signal, and performing feature extraction to obtain a feature sequence, performing modal decomposition on the sound signal according to the numerical distribution of the feature sequence to obtain each modal component; screening abnormal sound points according to the frequency domain feature differences of all modal components of the sound signal at different positions; obtaining valve abnormality coefficients according to the fluctuations of the pressure sequence at the abnormal sound point and the trend changes and numerical fluctuations of the flow sequence; screening abnormal valve positions according to the valve abnormality coefficients to obtain the fault prediction results of the evaporator at the current moment. The present application can improve the accuracy of evaporator fault prediction.

Description

Technical Field

[0001] The present application relates to the technical field of evaporator operation data processing, and in particular to an evaporator fault prediction method for evaporating materials. Background Art

[0002] The evaporator is an industrial equipment mainly used to evaporate the solvent in the liquid material by heating, so as to increase the concentration of the solid components of the material. When the material processed by the evaporator contains high concentrations of calcium and magnesium ions, the increase in temperature during the evaporation process will cause these ions to combine with anions such as carbonate to form insoluble inorganic salts, which will gradually deposit on valves and pipes to form hard scale, causing valve blockage, resulting in interruption of the production process and affecting production efficiency.

[0003] At present, the prediction of evaporator valve failure usually relies on the abnormal detection of valve monitoring data, but ignores the dynamic behavior of the fluid in the evaporator when the valve is blocked. This causes the evaporator equipment to have a delayed response to the failure and cannot accurately predict the valve failure. Summary of the invention

[0004] In order to solve the above technical problems, the present application provides an evaporator fault prediction method for evaporating materials to solve the existing problems.

[0005] The present invention discloses a method for predicting evaporator failures for evaporating materials using the following technical solutions:

[0006] An embodiment of the present application provides a method for predicting evaporator failures for evaporating materials, the method comprising the following steps:

[0007] S1, collecting sound signals, pressure sequences and flow sequences at various positions of the evaporator valve within a preset time period before the current moment;

[0008] S2, segment the sound signal and perform feature extraction to obtain a feature sequence, determine the stopping threshold in the modal decomposition algorithm according to the noise content and information complexity of the feature sequence, perform modal decomposition on the sound signal according to the stopping threshold to obtain each modal component; filter out abnormal sound points according to the frequency domain feature differences of all modal components of the sound signal at different positions;

[0009] S3, according to the fluctuation of the pressure sequence at the abnormal sound point, the pressure fluctuation change value is obtained; according to the trend change and value fluctuation of the flow sequence at the abnormal sound point, the flow trend fluctuation index is obtained;

[0010] S4, comprehensively analyzing the pressure fluctuation change value and flow trend fluctuation index at each abnormal sound point, and obtaining the valve abnormality coefficient of each abnormal sound point; according to the valve abnormality coefficient of all abnormal sound points, the abnormal valve position is screened to obtain the fault prediction result of the evaporator at the current moment.

[0011] Furthermore, the method for obtaining each modal component includes:

[0012] The sound signal is evenly divided into several frames; the feature extraction of the sound signal contained in each frame is performed to obtain a feature vector; the feature vectors of all frames are arranged in chronological order to obtain a feature sequence of the sound signal; the signal-to-noise ratio and information entropy of the feature sequence are calculated;

[0013] The preset stop threshold is adjusted according to the signal-to-noise ratio and information entropy of the characteristic sequence of the sound signal to obtain an adaptive stop threshold;

[0014] The adaptive stopping threshold is used as the stopping threshold in the modal decomposition algorithm to perform modal decomposition on the sound signal to obtain each modal component.

[0015] Further, the method for obtaining the adaptive stop threshold includes: recording the adaptive stop threshold as , ;in, is the preset stop threshold; is the preset first influence coefficient; SNR is the signal-to-noise ratio of the characteristic sequence; is the preset second influence coefficient; is the information entropy of the feature sequence.

[0016] Furthermore, the method of screening abnormal sound points according to the frequency domain feature differences of all modal components of the sound signal at different positions includes:

[0017] The modal components are converted into the frequency domain to obtain the modal frequencies; the frequency domain eigenvalues ​​of the modal components are obtained according to the numerical distribution characteristics of the amplitudes corresponding to all the modal frequencies of the modal components;

[0018] The mean of the frequency domain eigenvalues ​​of all modal components of the sound signal is taken as the abnormality discrimination value of the sound signal; the mean of the abnormality discrimination values ​​of the sound signal at all positions is taken as the sound abnormality threshold; the position corresponding to the sound signal with an abnormality discrimination value greater than the sound abnormality threshold is recorded as the sound abnormal point.

[0019] Furthermore, the frequency domain eigenvalues ​​of the modal components include:

[0020] The maximum value of the amplitude corresponding to all modal frequencies is taken as the frequency peak value of the modal component; the standard deviation of the frequency peak values ​​of all modal frequencies corresponding to each modal component is calculated as the frequency domain eigenvalue of the modal component.

[0021] Furthermore, the method for obtaining the pressure fluctuation change value includes:

[0022] For each abnormal sound point, a first-order difference sequence of the pressure sequence at the abnormal sound point is obtained; the standard deviation of all elements in the first-order difference sequence is calculated and recorded as the pressure fluctuation change value at the abnormal sound point.

[0023] Furthermore, the method for obtaining the flow trend fluctuation index includes:

[0024] Perform a straight line fitting on the flow sequence at the abnormal sound point to obtain the slope of the fitted line; record the flow trend fluctuation index as , ; In the formula, For flow sequence The standard deviation of For flow sequence The slope of the fitted line; is the flow sequence is the average value; || is the absolute value symbol.

[0025] Furthermore, the method for obtaining the valve abnormality coefficient of each abnormal sound point includes: recording the valve abnormality coefficient of the abnormal sound point as , ; Wherein, P is the pressure fluctuation change value; F is the flow trend fluctuation index.

[0026] Furthermore, the abnormal valve position is screened according to the valve abnormality coefficients of all abnormal sound points to obtain the fault prediction result of the evaporator at the current moment, including:

[0027] The valve abnormality threshold is calculated based on the numerical distribution of the valve abnormality coefficients of all sound abnormality points; the sound abnormality points whose valve abnormality coefficients are greater than the valve abnormality threshold are recorded as valve abnormality positions; if the number of valve abnormality positions is greater than the preset value, a valve blockage fault occurs in the evaporator.

[0028] Furthermore, the method for calculating the valve abnormality threshold includes: calculating the sum of the mean and standard deviation of the valve abnormality coefficients of all sound abnormal points as the valve abnormality threshold.

[0029] This application has at least the following beneficial effects:

[0030] This application analyzes the change in total energy of the sound signal after wavelet transformation and the change in the peak value of each inherent mode function after modal decomposition when the valve is blocked, effectively extracts the characteristics of the sound signal at different frequencies, screens the abnormal sound points, avoids the traditional method of extracting some characteristics of the sound signal and ignoring other information that may reflect the fault state, and constructs the pressure fluctuation change value and flow trend fluctuation index based on the change characteristics of pressure and flow in the valve area of ​​the abnormal sound point, and obtains the valve abnormality coefficient of each abnormal sound point, comprehensively considering the fluctuation degree of pressure and flow of the fluid when the valve is blocked and the trend information of flow change over time. Finally, the fault prediction result is obtained based on the valve abnormality coefficient of all abnormal sound points; the key features are extracted based on the frequency domain analysis and time series data analysis method to more accurately reflect the dynamic behavior of the fluid when the valve is blocked, avoiding the problem of misjudgment caused by relying only on pressure or flow data in the traditional method, and further improving the accuracy of evaporator fault prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0032] Figure 1 A flowchart of the steps of a method for predicting evaporator failure for evaporating materials provided in the present application;

[0033] Figure 2 A block diagram for obtaining sound abnormal points provided for one embodiment of the present application. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0035] 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 application belongs.

[0036] The specific scheme of the evaporator fault prediction method for evaporating materials provided by the present application is described in detail below with reference to the accompanying drawings.

[0037] An embodiment of the present application provides a method for predicting evaporator failures for evaporating materials. Specifically, the following method for predicting evaporator failures for evaporating materials is provided. Figure 1 , the method comprises the following steps:

[0038] S1, collecting sound signals, pressure sequences and flow sequences at various positions of the evaporator valve within a preset time period before the current moment.

[0039] Install sound pressure sensors at various positions on the evaporator valve device, and use a data acquisition card to record sound signals in real time. In this embodiment, the sampling frequency is 44.1KHz; the collected sound signals are further filtered, amplified, and analog-to-digital converted to convert continuous signals into digital signals. The pressure and flow data are preprocessed using filtering algorithms, standardization and normalization algorithms to remove noise and irrelevant information, ensure the accuracy and completeness of the data, and facilitate further analysis. The preprocessing operation is a well-known technology and will not be repeated here. At the same time, in this embodiment, a pressure and flow sensor is installed 25cm downstream of the valve pipeline to collect pressure sequences and flow sequences. The length of each collected data sequence is n, and the time interval for collecting data is T. In this embodiment, n is 25 and T is 5s.

[0040] S2, segment the sound signal and extract features to obtain a feature sequence. Determine the stopping threshold in the modal decomposition algorithm based on the noise content and information complexity of the feature sequence. Perform modal decomposition on the sound signal based on the stopping threshold to obtain each modal component. Screen out sound abnormal points based on the frequency domain feature differences of all modal components of the sound signal at different positions.

[0041] During the operation of the evaporator, valve blockage is a common fault phenomenon. This fault usually causes a significant change in the sound signal. The fluid passing through the narrow channel will produce eddies and turbulence. These phenomena will produce more high-frequency noise. Due to the changes in fluid dynamics caused by blockage, the nonlinearity and non-stationarity of the signal will be more significant. Valve blockage is accompanied by an increase in the kinetic energy of the fluid because the fluid is forced to pass through a smaller opening. This increased kinetic energy may be converted into acoustic energy, resulting in an increase in the overall energy of the sound signal.

[0042] The empirical mode decomposition algorithm is usually used to obtain the modal components of the sound signal. By analyzing the changes in different frequency components, valve blockage events can be more accurately identified. The setting of the stop threshold in the stop criterion of the empirical mode decomposition has a significant impact on the results and quality of the signal decomposition. If the threshold is too small, it will waste computing resources for over-plastic screening, and if the threshold is too large, it will not be enough to accurately extract the modal components. Reasonable adjustment of the threshold can balance the consumption of computing resources and the accuracy of modal component extraction. In signals with strong nonlinearity, more modal components may be required to capture the local characteristics of the signal, so the threshold can be appropriately increased. However, when the noise level in the signal is high and the signal-to-noise ratio is low, the threshold needs to be lowered to avoid decomposing too many meaningless modal components.

[0043] Since sound is a non-stationary signal, it is usually divided into short time frames. In this embodiment, the sound signal is evenly divided into several frames. The time of each frame is 20 milliseconds, and the feature extraction is performed on the sound signal contained in each frame to obtain a feature vector.

[0044] Among them, the feature extraction of the sound signal can extract the frequency, time domain and acoustic features of the speech; in this embodiment, the feature vector is obtained from each frame by extracting the MFCC Mel-frequency cepstral coefficients, wherein the specific implementation process of the Mel-frequency cepstral coefficients is a well-known technology and will not be repeated. As other embodiments of the present application, the short-time Fourier transform method can be used to extract the features of the sound signal, and the present application does not limit it.

[0045] Further, the feature vectors of all frames are arranged in chronological order to obtain the feature sequence of the sound signal. The signal-to-noise ratio and information entropy of the feature sequence of the sound signal are calculated; the calculation of the signal-to-noise ratio and information entropy are both well-known technologies and will not be described in detail. The preset stop threshold is adjusted according to the signal-to-noise ratio and information entropy of the feature sequence of the sound signal to obtain an adaptive stop threshold. The calculation formula of the adaptive stop threshold is:

[0046]

[0047] In the formula, is the adaptive stopping threshold, is the preset stop threshold, which is 0.25 in this embodiment. It is the first influence coefficient, reflecting the influence of SNR on the threshold. Here it is taken as 0.01. SNR is the signal-to-noise ratio of the feature sequence. is the second influence coefficient, which reflects the influence of nonlinearity on the threshold. In this embodiment, the value is 0.02. is the information entropy of the feature sequence.

[0048] When the SNR is low, the threshold will be reduced accordingly to reduce the generation of meaningless modal components and prevent excessive decomposition of signals in a noisy environment. As the entropy value increases, the nonlinearity of the signal increases, and the threshold increases so that more intrinsic mode functions (IMFs) are decomposed, which can better analyze the complex structure of the signal. The first influence coefficient and the second influence coefficient serve to unify the magnitude of the stop threshold adjustment.

[0049] According to the set threshold, the sound signal is modally decomposed, and the adaptive stopping threshold is used as the stopping threshold of the stopping criterion in the EMD empirical mode decomposition algorithm to control the extraction of modal components and obtain each modal component. The specific implementation process of the EMD empirical mode decomposition algorithm is a well-known technology and will not be repeated here.

[0050] When the valve is blocked or has an abnormal fault, the vibration of certain frequency components in the sound signal increases, which in turn affects the increase of the vibration intensity of the sound signal at certain frequencies. Therefore, this embodiment analyzes the change characteristics of the sound signal according to the modal components of the sound signal, and determines the working state of the valve according to the change characteristics of the sound signal.

[0051] First, a Fourier transform is performed on each modal component to identify each modal frequency and its corresponding amplitude in the spectrum; these amplitudes represent the vibration intensity of different frequencies of each mode. Therefore, this embodiment obtains the frequency domain eigenvalue of the modal component based on the numerical distribution characteristics of the amplitude corresponding to all modal frequencies of the modal component.

[0052] Specifically, for the identified modal frequencies, the maximum value of the amplitudes corresponding to all modal frequencies is taken as the frequency peak value; the standard deviation of the frequency peak values ​​of all modal frequencies corresponding to each modal component is calculated as the frequency domain eigenvalue of the modal component;

[0053] Furthermore, the mean of the frequency domain eigenvalues ​​of all modal components of the sound signal is used as the abnormality discrimination value of the sound signal; the frequency peak value of some modal frequencies becomes larger, that is, when the valve is blocked, the vibration of certain frequency components in the sound signal increases, which in turn affects the increase in the vibration intensity of the sound signal at certain frequencies, making the standard deviation of the frequency peak value increase, and the valve blockage or abnormal failure is more likely to occur.

[0054] Therefore, the average of the abnormality discrimination values ​​of the sound signals at all positions is taken as the sound abnormality threshold; and the position corresponding to the sound signal whose abnormality discrimination value is greater than the sound abnormality threshold is recorded as the sound abnormal point.

[0055] The block diagram for obtaining abnormal sound points is as follows: Figure 2 shown.

[0056] S3, according to the fluctuation of the pressure sequence at the abnormal sound point, the pressure fluctuation change value is obtained; according to the trend change and value fluctuation of the flow sequence at the abnormal sound point, the flow trend fluctuation index is obtained.

[0057] As the valve begins to become clogged, the fluid requires a higher pressure to pass through a smaller opening, and the opening through which the fluid passes is reduced, resulting in an increase in the pressure difference before and after the valve. Especially when the valve is partially clogged, the flow of the fluid may become uneven, causing pressure fluctuations. The blockage causes the effective flow area to be reduced, so the flow through the valve is correspondingly reduced, and the flow of the fluid becomes irregular.

[0058] In order to analyze the pressure changes in continuous time, for each sound abnormal point, the first-order difference sequence of the pressure sequence at the sound abnormal point is obtained; the standard deviation of all elements in the first-order difference sequence is calculated and recorded as the pressure fluctuation change value at the sound abnormal point; the standard deviation of the first-order difference sequence measures the volatility of the difference sequence.

[0059] Since the blockage restricts the flow of fluid, the flow rate decreases over time, so the flow rate trend will show an obvious decline. Further considering the actual volatility of the flow data and its trend changes over time, the flow trend fluctuation index is constructed according to the fluctuation degree and trend changes of the flow sequence.

[0060] Specifically, the flow sequence at the abnormal sound point is fitted by the least square method to obtain the slope of the fitted line; the flow trend fluctuation index is recorded as , ; In the formula, For flow sequence The standard deviation of For flow sequence The slope of the fitted line; is the flow sequence is the average value; || is the absolute value symbol.

[0061] As the blockage occurs, the flow measurement value will decrease, the mean of the flow series will decrease, and the standard deviation of the flow data will increase. The slope obtained by linear regression When it is a negative number, it means that the flow measurement value shows a downward trend when blocked, and shows an upward trend after taking the absolute value. The ratio with the average flow value becomes larger, and the flow trend fluctuation index becomes larger.

[0062] S4, comprehensively analyzing the pressure fluctuation change value and flow trend fluctuation index at each abnormal sound point, and obtaining the valve abnormality coefficient of each abnormal sound point; according to the valve abnormality coefficient of all abnormal sound points, the abnormal valve position is screened to obtain the fault prediction result of the evaporator at the current moment.

[0063] When the valve starts to get clogged, the fluid passing through the valve is restricted. As the flow channel becomes narrower, the fluid velocity increases in the narrow part. According to Bernoulli's principle, the increased velocity leads to a decrease in pressure, and the pressure fluctuations before and after the valve become significant, resulting in a significant decrease in flow. Based on the above analysis, the valve abnormality coefficient of the abnormal sound point is obtained by combining the pressure fluctuation change value and the flow trend fluctuation index. The valve abnormality coefficient of the abnormal sound point is recorded as , ; Wherein, P is the pressure fluctuation change value; F is the flow trend fluctuation index.

[0064] As another embodiment of the present application, the product of the pressure fluctuation change value and the flow trend fluctuation index may be calculated as the valve abnormality coefficient.

[0065] The pressure increase caused by blockage will directly lead to a decrease in flow rate. Therefore, the pressure fluctuation change value and the flow trend fluctuation index will increase, so the valve abnormality coefficient becomes larger, indicating that the evaporator is affected by valve blockage.

[0066] Furthermore, according to the numerical distribution of valve abnormality coefficients of all sound abnormal points, the valve abnormality threshold is calculated; specifically, the sum of the mean and standard deviation of the valve abnormality coefficients of all sound abnormal points is calculated as the valve abnormality threshold; the sound abnormal point with a valve abnormality coefficient greater than the valve abnormality threshold is recorded as the valve abnormal position. If the number of valve abnormal positions is greater than the preset value a=3, the evaporator has a valve blockage fault.

[0067] Through the above description of the implementation method in combination with the accompanying drawings, technical personnel in the relevant field can understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0068] The above contents are only specific implementation methods of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be covered by the protection scope of the present application.

Claims

1. A method for predicting evaporator failures for evaporating materials, characterized in that: The method comprises the following steps: S1, collecting sound signals, pressure sequences and flow sequences at various positions of the evaporator valve within a preset time period before the current moment; S2, segment the sound signal and perform feature extraction to obtain a feature sequence, determine the stopping threshold in the modal decomposition algorithm according to the noise content and information complexity of the feature sequence, perform modal decomposition on the sound signal according to the stopping threshold to obtain each modal component; filter out abnormal sound points according to the frequency domain feature differences of all modal components of the sound signal at different positions; S3, according to the fluctuation of the pressure sequence at the abnormal sound point, the pressure fluctuation change value is obtained; according to the trend change and value fluctuation of the flow sequence at the abnormal sound point, the flow trend fluctuation index is obtained; S4, combining the pressure fluctuation change value and flow trend fluctuation index at each abnormal sound point to obtain the valve abnormality coefficient of each abnormal sound point; according to the valve abnormality coefficient of all abnormal sound points, the abnormal valve position is screened to obtain the fault prediction result of the evaporator at the current moment; The method for obtaining each modal component includes: The sound signal is evenly divided into several frames; the feature extraction of the sound signal contained in each frame is performed to obtain a feature vector; the feature vectors of all frames are arranged in chronological order to obtain a feature sequence of the sound signal; the signal-to-noise ratio and information entropy of the feature sequence are calculated; The preset stop threshold is adjusted according to the signal-to-noise ratio and information entropy of the characteristic sequence of the sound signal to obtain an adaptive stop threshold; Using the adaptive stopping threshold as the stopping threshold in the modal decomposition algorithm, the sound signal is modally decomposed to obtain each modal component; The method for obtaining the adaptive stop threshold comprises: recording the adaptive stop threshold as , ;in, is the preset stop threshold; is the preset first influence coefficient; SNR is the signal-to-noise ratio of the characteristic sequence; is the preset second influence coefficient; is the information entropy of the feature sequence; The method for obtaining the flow trend fluctuation index includes: Perform a straight line fitting on the flow sequence at the abnormal sound point to obtain the slope of the fitted line; record the flow trend fluctuation index as , ; In the formula, For flow sequence The standard deviation of For flow sequence The slope of the fitted line; is the flow sequence The average value of ;||is the absolute value symbol; The method for obtaining the valve abnormality coefficient of each abnormal sound point includes: recording the valve abnormality coefficient of the abnormal sound point as , ; Wherein, P is the pressure fluctuation change value; F is the flow trend fluctuation index.

2. The method for predicting evaporator failure for evaporating materials according to claim 1, characterized in that: The method of screening abnormal sound points according to the frequency domain feature differences of all modal components of the sound signal at different positions includes: The modal components are converted into the frequency domain to obtain the modal frequencies; the frequency domain eigenvalues ​​of the modal components are obtained according to the numerical distribution characteristics of the amplitudes corresponding to all the modal frequencies of the modal components; The mean of the frequency domain eigenvalues ​​of all modal components of the sound signal is taken as the abnormality discrimination value of the sound signal; the mean of the abnormality discrimination values ​​of the sound signal at all positions is taken as the sound abnormality threshold; the position corresponding to the sound signal with an abnormality discrimination value greater than the sound abnormality threshold is recorded as the sound abnormal point.

3. The method for predicting evaporator failure for evaporating materials according to claim 2, characterized in that: The frequency domain eigenvalues ​​of the modal components include: The maximum value of the amplitude corresponding to all modal frequencies is taken as the frequency peak value of the modal component; the standard deviation of the frequency peak values ​​of all modal frequencies corresponding to each modal component is calculated as the frequency domain eigenvalue of the modal component.

4. The method for predicting evaporator failure for evaporating materials according to claim 1, characterized in that: The method for obtaining the pressure fluctuation change value includes: For each abnormal sound point, a first-order difference sequence of the pressure sequence at the abnormal sound point is obtained; the standard deviation of all elements in the first-order difference sequence is calculated and recorded as the pressure fluctuation change value at the abnormal sound point.

5. The method for predicting evaporator failure for evaporating materials according to claim 1, characterized in that: The abnormal valve position is screened according to the valve abnormality coefficient of all abnormal sound points to obtain the fault prediction result of the evaporator at the current moment, including: The valve abnormality threshold is calculated based on the numerical distribution of the valve abnormality coefficients of all sound abnormality points; the sound abnormality points whose valve abnormality coefficients are greater than the valve abnormality threshold are recorded as valve abnormality positions; if the number of valve abnormality positions is greater than the preset value, a valve blockage fault occurs in the evaporator.

6. A method for predicting evaporator failures for evaporating materials as claimed in claim 5, characterized in that: The method for calculating the valve abnormality threshold comprises: calculating the sum of the mean and standard deviation of the valve abnormality coefficients of all sound abnormal points as the valve abnormality threshold.

Images