Fault detection method and system for plastic pipe extruder

Abstract

The present invention relates to the field of data processing technology, and more specifically, to a method and system for detecting faults in a plastic tube extruder. The method comprises: collecting the temperature, pressure, and extrusion rate of the extruder's extrusion port at a preset collection frequency to obtain a temperature sequence, a pressure sequence, and an extrusion rate sequence; calculating the variance of the temperature sequence; and preliminarily determining the presence of noise data in the temperature sequence if the variance exceeds a preset variance threshold; obtaining suspected noise data; obtaining a first Pearson correlation coefficient; and obtaining a fault prediction result for the extruder. The method of the present invention can effectively improve the accuracy of extruder monitoring results.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and more particularly to a fault detection method and system for a plastic pipe extruder. Background Art

[0002] With the advancement of technology, the use of plastic pipes in our daily lives and work is becoming increasingly widespread. Plastic pipes are particularly important in the fields of construction, agriculture, industry, and medicine, where they play an irreplaceable role. In water supply systems, plastic pipes are commonly used in buildings' tap water systems, ensuring long-term stable operation and resisting corrosion from chemicals in the water. In drainage systems, including drainage, exhaust, and sanitary pipes, plastic pipes can quickly and efficiently remove wastewater and exhaust gases from buildings, resisting corrosion from chemicals in the wastewater. Plastic pipes can also be used as conduit for electrical wiring installations, protecting wires and cables from damage.

[0003] Plastic pipes are typically made from synthetic resin (polyester) with additives such as stabilizers, lubricants, and plasticizers, and are processed through a plastic pipe extruder. To ensure the quality of the produced plastic pipes, the extruder must be operating normally during production. Therefore, it is crucial to monitor and predict the operating status of the extruder, allowing for prompt maintenance in the event of a malfunction.

[0004] With the popularization of deep learning, the trained neural network model can be combined with the collected operating parameters of the extruder to monitor the extruder and predict the fault status of the extruder, such as Figure 1 The method for monitoring an extruder using a neural network is as follows: first, the physical parameters of the extrusion port of the extruder are collected using sensors, and then the collected physical parameters are input into a neural network model to obtain a fault prediction result and display it on a display interface; however, since the sensors used to collect the operating parameters of the extruder may be interfered with and generate noise, the collected operating parameters of the extruder usually contain noise data, resulting in poor accuracy of the monitoring results of the extruder. Summary of the Invention

[0005] In order to solve the technical problem of poor accuracy of monitoring results when monitoring an extruder in the prior art, the present invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a fault detection method for a plastic pipe extruder, comprising: collecting the temperature, pressure, and extrusion rate of an extrusion port of the extruder at a preset collection frequency, thereby obtaining a temperature sequence, a pressure sequence, and an extrusion rate sequence; calculating the variance of the temperature sequence, and preliminarily determining that noise data exists in the temperature sequence if the variance is greater than a preset variance threshold;

[0007] Fitting the temperature data in the temperature sequence to obtain a first fitting curve, and calculating the residual corresponding to each temperature data; determining the temperature data whose residual is greater than a preset residual threshold as suspected noise data;

[0008] Fitting the pressure data corresponding to the neighborhood temperature data of the suspected noise data to obtain a second fitting curve, and calculating a first Pearson correlation coefficient between the neighborhood temperature data points of the suspected noise data and the pressure data points on the second fitting curve;

[0009] If the value of the first Pearson correlation coefficient is -1, the suspected noise data is determined to be non-noise data, and the temperature sequence, pressure sequence, and extrusion rate sequence are input into the preset neural network model to obtain the fault prediction result of the extruder; otherwise, the suspected noise data is determined to be real noise data, and median filtering is performed on it; and the filtered temperature sequence, pressure sequence, and extrusion rate sequence are input into the preset neural network model to obtain the fault prediction result of the extruder.

[0010] The fault detection method for a plastic pipe extruder of the present invention first preliminarily determines whether noise exists in the temperature sequence based on the variance corresponding to the temperature sequence after collecting the temperature sequence of the extrusion port. Under the condition that noise exists, suspected noise data is further screened out based on the fitting curve of the temperature data and the residual of the curve data points. After screening out the suspected noise data, the relationship between the temperature and pressure of the extrusion port during the plastic extrusion process is taken into consideration, and combined with the value of the first Pearson correlation coefficient, it is further determined whether the suspected noise data is real noise, and the real noise is filtered and denoised, thereby greatly improving the accuracy of the collected extrusion port temperature data.

[0011] In addition, since the temperature data is further processed only when the preliminary judgment result is that there is noise in the temperature sequence, and when the preliminary judgment result is that there is no noise in the temperature sequence, the collected temperature sequence, pressure sequence and extrusion rate sequence are directly input into the preset neural network model to obtain the fault prediction result, the neural network model can automatically learn the relationship between the collected physical parameters of the extrusion port and the fault detection results, thereby greatly improving the efficiency and accuracy of extruder monitoring; furthermore, when using the neural network model to detect the fault status of the extruder, the temperature, pressure and extrusion rate of the extrusion port are comprehensively considered, rather than considering a single physical parameter, thereby avoiding misjudgment and further improving the accuracy of the extruder monitoring results.

[0012] Preferably, the pressure collection method of the extrusion port includes:

[0013] A pressure sensor is placed at the extrusion port, and the pressure at the extrusion port is calculated based on the value collected by the pressure sensor. The calculation expression is:

[0014] ;

[0015] Where, Indicates the pressure at the extrusion port. It represents the value collected by the pressure sensor, and k represents the compensation coefficient, which ranges from 0.05 to 0.1.

[0016] The pressure collection method of the extrusion port of the present invention does not directly use the value collected by the pressure sensor as the pressure of the extrusion port when collecting the pressure of the extrusion port, but takes into account the friction between the plastic tube and the inner wall of the extrusion port, corrects the value collected by the pressure sensor, and uses the corrected value as the collected extrusion port pressure, thereby improving the accuracy of extrusion port pressure collection.

[0017] Preferably, the neighborhood range is a range consisting of five data points centered on the suspected noise data point.

[0018] Preferably, the method further includes: calculating the noise performance level of each pressure data point in the pressure sequence, and performing median filtering on the pressure data points whose noise performance level is greater than a noise performance level threshold. The calculation expression is:

[0019] ;

[0020] Where, Indicates the noise performance of the i-th pressure data point, represents the exponential function with base e, represents the mean of the pressure data points in the neighborhood of the i-th pressure data point, represents the mean value between the i-th pressure data point and its adjacent pressure data points. represents the second Pearson correlation coefficient between the pressure data point and the corresponding temperature data point within the neighborhood of the i-th pressure data point; Indicates the pressure data point with the largest value in the neighborhood of the i-th pressure data point.

[0021] The method for calculating the noise performance degree of pressure data points of the present invention comprehensively considers the volatility of the i-th pressure data point and the changing relationship between the temperature and pressure of the extrusion port when calculating the noise performance degree of the pressure data points, thereby more accurately calculating the noise performance degree of the pressure data points.

[0022] Preferably, the training method of the neural network model includes:

[0023] The temperature series, pressure series, and extrusion rate series of the extruder at historical moments are collected and labeled to form a training set.

[0024] Input the data in the training set into the neural network model to obtain the predicted value;

[0025] Compare the predicted value with the true value and calculate the loss value using the average error function;

[0026] Use the chain rule to calculate the gradient of the loss function with respect to each parameter, and then backpropagate these gradients back to each layer in the network; and update the weights and bias of each neuron according to the gradient value;

[0027] Iteratively updates the weights and biases of each neuron until the loss converges or the preset number of iterations is reached.

[0028] Preferably, the method further comprises: collecting the rotation speed of the extruder screw in real time;

[0029] Call the target extrusion rate according to the real-time extrusion rate and the thickness requirement of the plastic pipe, and calculate the difference between the target extrusion rate and the real-time extrusion rate;

[0030] Performing PID calculation on the difference to obtain the target speed of the extruder screw;

[0031] The speed of the extruder screw is controlled by PID according to the difference between the target speed and the real speed of the extruder screw.

[0032] In the process of monitoring the extrusion port, the present invention further takes into account the influence of the real-time extrusion rate on the thickness of the plastic pipe, automatically calls the target extrusion rate according to the thickness requirement of the plastic pipe, and performs PID control on the speed of the extruder screw according to the target extrusion rate, thereby helping to make the thickness of the plastic pipe meet the thickness requirement.

[0033] Preferably, the method further comprises: adjusting the target speed of the extruder screw according to the temperature of the extrusion port, and the target speed of the screw after adjustment is calculated as follows:

[0034] ;

[0035] Where, Indicates the target speed of the screw after adjustment. Indicates the target speed of the screw before adjustment. Indicates the real-time temperature of the extrusion port. Indicates the standard temperature of the extrusion nozzle.

[0036] When the real-time temperature of the extrusion port increases, the fluidity of the plastic will improve and the thickness of the extrudate will become thinner, while when the real-time temperature decreases, the fluidity of the plastic will deteriorate and the thickness of the extrudate will increase. Therefore, the present invention calculates the target speed of the extruder screw and further adjusts the target speed of the extruder screw according to the difference between the real-time temperature and the standard temperature, thereby helping to maintain a constant thickness of the produced plastic pipe and further improving the control accuracy of the thickness of the produced plastic pipe.

[0037] Preferably, the method further comprises: correcting the target speed of the screw after adjustment according to the pressure of the extrusion port, and the screw speed correction value is calculated as follows:

[0038] ;

[0039] Where, Indicates the screw speed correction value, Indicates the pressure value at the extrusion port. Indicates the standard value of the pressure at the extrusion port. Represents the exponential function with base e.

[0040] After adjusting the target speed of the screw according to the extrusion port temperature, the influence of the extrusion port pressure on the thickness of the plastic pipe is further taken into consideration, and the target speed of the screw is further corrected according to the extrusion port pressure, thereby further improving the control accuracy of the thickness of the produced plastic pipe.

[0041] In a second aspect, the present invention provides a fault detection system for a plastic tube extruder, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the fault detection method for the plastic tube extruder of the present invention is implemented.

[0042] The beneficial effect of the present invention is that the method of the present invention can effectively improve the efficiency and accuracy of extruder monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the following detailed description with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an illustrative and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0044] Figure 1 The present invention schematically shows a flow chart of a method for predicting failure of an extruder using a neural network model in the prior art;

[0045] Figure 2 is a flow chart schematically illustrating a fault detection method for a plastic tube extruder according to an embodiment of the present invention;

[0046] Figure 3 Schematic diagram of force analysis of an extrusion port according to an embodiment of the present invention;

[0047] Figure 4 FIG. 1 is a schematic diagram showing the structure of a fault detection system for a plastic pipe extruder according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work shall fall within the scope of protection of the present invention.

[0049] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0050] Fault detection method embodiment of plastic pipe extruder:

[0051] like Figure 2 As shown, the fault detection method of the plastic pipe extruder of the present invention comprises:

[0052] S101, acquiring a temperature sequence, a pressure sequence, and an extrusion rate sequence at an extrusion port of an extruder, specifically: acquiring the temperature, pressure, and extrusion rate at the extrusion port of the extruder at a preset acquisition frequency, thereby acquiring the temperature sequence, the pressure sequence, and the extrusion rate sequence;

[0053] The temperature of the extrusion port can be measured by installing a resistance temperature sensor, thermocouple temperature sensor, or radiation pyrometer at the extrusion port. The pressure of the extrusion port can be measured by installing a pressure sensor inside the extrusion port. A flow meter can be installed at the extrusion port to measure the plastic extrusion rate.

[0054] S102, calculating the variance of the temperature sequence, and if the variance is greater than a preset variance threshold, preliminarily determining that noise data exists in the temperature sequence;

[0055] The variance of the temperature series is too large, indicating that the temperature data fluctuates greatly, which is likely due to interference with the temperature sensor, resulting in noisy data.

[0056] S103, obtaining suspected noise data, specifically: fitting the temperature data in the temperature sequence to obtain a first fitting curve, and calculating the residual corresponding to each temperature data; determining the temperature data whose residual is greater than a preset residual threshold as suspected noise data;

[0057] The fitting method can adopt a polynomial fitting method. The value of the preset residual threshold can be determined based on experiments. In this embodiment, the residual threshold can be set to 3 degrees Celsius.

[0058] S104, obtaining a first Pearson correlation coefficient, specifically: fitting the pressure data corresponding to the temperature data in the neighborhood of the suspected noise data to obtain a second fitting curve, and calculating the first Pearson correlation coefficient between the temperature data points in the neighborhood of the suspected noise data and the pressure data points on the second fitting curve;

[0059] Since the temperature data and the pressure data are collected at the same collection frequency, the pressure data corresponding to the temperature data refers to the pressure data collected at the time of temperature data collection.

[0060] In this embodiment, the neighborhood range is a range consisting of five data points centered on the suspected noise data point. In other embodiments, the neighborhood range may also be other appropriate ranges.

[0061] During the extrusion process, temperature changes directly affect the plastic's melt state and fluidity. When the temperature rises, the plastic's meltability increases, improving its fluidity and leading to a decrease in pressure at the extrusion port. Conversely, when the temperature drops, the plastic's meltability decreases, impairing its fluidity and leading to an increase in pressure at the extrusion port. Therefore, there is a linear relationship between the temperature and pressure data at the extrusion port, and the temperature trend at the extrusion port is opposite to the pressure trend: an upward trend in the temperature data corresponds to a downward trend in the pressure data. The first Pearson correlation coefficient accurately reflects the correlation between the temperature data points in the neighborhood of the noise data and the pressure data points on the second fitted curve. A value of 0 indicates no linear relationship between the temperature and pressure data points. A value of 1 indicates that when one of the temperature and pressure data points increases, the other also increases. A value of -1 indicates that when one of the temperature and pressure data points increases, the other decreases.

[0062] S105. Obtain the fault prediction result of the extruder, specifically: if the value of the first Pearson correlation coefficient is -1, the suspected noise data is determined to be non-noise data, and the temperature sequence, pressure sequence and extrusion rate sequence are input into the preset neural network model to obtain the fault prediction result of the extruder; otherwise, the suspected noise data is determined to be real noise data, and median filtering is performed on it; and the filtered temperature sequence, pressure sequence and extrusion rate sequence are input into the preset neural network model to obtain the fault prediction result of the extruder.

[0063] In this embodiment, the neural network model includes a first feature extraction model, a second feature extraction model, a third feature extraction model, a feature splicing model and a classification model, wherein the first feature extraction model is used to perform time series feature extraction on the temperature sequence to obtain temperature features; the second feature extraction model is used to perform time series feature extraction on the pressure sequence to obtain pressure features; the third feature extraction model is used to perform time series feature extraction on the extrusion rate sequence to obtain extrusion rate features; the feature splicing model is used to splice the temperature features, pressure features and extrusion rate features together to obtain splicing features; the classification model is used to perform dimensionality transformation on the splicing features and output fault prediction results.

[0064] The fault detection method for the plastic pipe extruder of the present embodiment, after collecting the temperature sequence of the extrusion port, first preliminarily judges whether there is noise in the temperature sequence based on the variance corresponding to the temperature sequence, and under the condition that there is noise, further screens out suspected noise data based on the fitting curve of the temperature data and the residual of the curve data point; after screening out the suspected noise data, considering the relationship between the temperature and pressure of the extrusion port during the plastic extrusion process, combined with the value of the first Pearson correlation coefficient, further determines whether the suspected noise data is real noise, and filters and denoises the real noise, thereby greatly improving the accuracy of the collected extrusion port temperature data; in addition, since the preliminary judgment result is that the temperature sequence The temperature data is further processed only when there is noise. When the preliminary judgment result is that there is no noise in the temperature sequence, the collected temperature sequence, pressure sequence and extrusion rate sequence are directly input into the preset neural network model to obtain the fault prediction result. The neural network model can automatically learn the relationship between the collected physical parameters of the extrusion port and the fault detection results, thereby greatly improving the efficiency and accuracy of extruder monitoring; furthermore, when using the neural network model to detect the fault status of the extruder, the temperature, pressure and extrusion rate of the extrusion port are comprehensively considered, rather than considering a single physical parameter, thereby avoiding misjudgment and further improving the accuracy of the extruder monitoring results.

[0065] In one embodiment, the pressure collection method of the extrusion port includes:

[0066] A pressure sensor is placed at the extrusion port, and the pressure at the extrusion port is calculated based on the value collected by the pressure sensor. The calculation expression is:

[0067] ;

[0068] Where, Indicates the pressure at the extrusion port. Indicates the pressure value collected by the pressure sensor, k indicates the compensation coefficient, and its value range is 0.05~0.1. The value range of can be obtained through experiments.

[0069] like Figure 3 As shown, the pressure at the extrusion port includes the pressure F1 perpendicular to the extrusion port surface and the friction force f parallel to the extrusion port surface. The resultant force F of the two is the extrusion port pressure. However, the pressure sensor installed at the extrusion port can only collect the pressure F1 perpendicular to the extrusion port surface. The actual pressure at the extrusion port is slightly greater than the force collected by the pressure sensor. Therefore, the pressure value collected by the pressure sensor is multiplied by a correction factor to obtain the actual pressure at the extrusion port. The extrusion port pressure collection method of this embodiment can more accurately collect the pressure at the extrusion port, thereby further improving the accuracy of the monitoring results of the extruder.

[0070] In one embodiment, the method further includes: calculating the noise performance level of each pressure data point in the pressure sequence, and performing median filtering on the pressure data points whose noise performance level is greater than a noise performance level threshold. The calculation expression is:

[0071] ;

[0072] Where, Indicates the noise performance of the i-th pressure data point, represents the exponential function with base e, represents the mean of the pressure data points in the neighborhood of the i-th pressure data point, represents the mean value between the i-th pressure data point and its adjacent pressure data points. represents the second Pearson correlation coefficient between the pressure data point and the corresponding temperature data point within the neighborhood of the i-th pressure data point; Indicates the pressure data point with the largest value in the neighborhood of the i-th pressure data point.

[0073] The second Pearson correlation coefficient between the pressure data point and the corresponding temperature data point in the neighborhood of the i-th pressure data point can reflect the degree of difference between the i-th pressure data point and the pressure data points in its neighborhood. The larger the value, the greater the volatility of the i-th pressure data point and the more likely it is noise. The second Pearson correlation coefficient between the pressure data point and the corresponding temperature data point in the neighborhood of the i-th pressure data point can reflect the linear relationship between the pressure data point and the corresponding temperature data point in the neighborhood of the i-th pressure data point. Since the temperature of the extrusion port increases or causes the pressure to decrease under normal circumstances, if the second Pearson correlation coefficient is -1, it means that during the time period corresponding to the neighborhood of the i-th pressure data point, as the pressure data increases, the temperature data decreases, indicating that the i-th pressure data point is likely not noise data. Otherwise, it means that when the pressure data changes, the temperature data may increase or remain unchanged. The pressure value of the i-th pressure data point may be noise generated by interference with the pressure sensor, rather than the actual pressure value at the extrusion port. In calculating the noise performance of the pressure data point, this embodiment comprehensively considers the volatility of the i-th pressure data point and the changing relationship between the temperature and pressure at the extrusion port when calculating the noise performance of the pressure data point, thereby more accurately calculating the noise performance of the pressure data point. In one embodiment, the training method of the neural network model includes:

[0074] S201, collecting the temperature sequence, pressure sequence, and extrusion rate sequence of the extruder at historical moments, and labeling them to form a training set;

[0075] Multiple sets of extruder operation data sequences at historical moments can be collected, and one set of extruder operation data sequences includes a temperature sequence, a pressure sequence, and an extrusion rate sequence that correspond to each other; the collected extruder operation data sequences include both the operation data sequences of the extruder under normal conditions and the operation data sequences of the extruder under abnormal conditions.

[0076] S202, inputting the data in the training set into the neural network model to obtain a predicted value;

[0077] S203, comparing the predicted value with the true value, and calculating the loss value using the average error function;

[0078] S204. Calculate the gradient of the loss function with respect to each parameter using the chain rule, and then backpropagate these gradients back to each layer in the network; and update the weights and biases of each neuron based on the gradient values;

[0079] S205. Iteratively update the weight and bias of each neuron until the loss value converges or reaches a preset number of iterations.

[0080] In one embodiment, the method further comprises: collecting the rotation speed of the extruder screw in real time;

[0081] S301, calling a target extrusion rate according to the real-time extrusion rate and the thickness requirement of the plastic pipe, and calculating the difference between the target extrusion rate and the real-time extrusion rate;

[0082] Normally, under the action of the screw's rotation, the plastic is forced toward the die head by the screw threads. During the screw's pushing process, the plastic is heated by the barrel's heat source. Friction between the barrel, the screw, and the plastic molecules also generates a large amount of heat, causing the plastic to gradually plasticize. By controlling the rate at which the plastic is extruded from the extruder, the length of time the plastic is heated in the barrel and the degree of friction between the plastic molecules can be controlled, thereby indirectly controlling the plastic's temperature. Changes in the plastic's temperature directly affect the plastic's meltability and fluidity, and thus the thickness of the extrudate. Therefore, for the same extruder, there is a corresponding relationship between the thickness of the plastic tube and the extrusion rate. Through experiments, the extrusion rates corresponding to various plastic tube thicknesses can be obtained.

[0083] S302, performing PID calculation on the difference to obtain a target speed of the extruder screw;

[0084] S303 , performing PID control on the speed of the extruder screw according to the difference between the target speed and the real-time speed of the extruder screw.

[0085] In the process of monitoring the extrusion port, this embodiment further takes into account the impact of the real-time extrusion rate on the thickness of the plastic pipe, automatically calls the target extrusion rate according to the thickness requirement of the plastic pipe, and performs PID control on the speed of the extruder screw according to the target extrusion rate, thereby helping to make the thickness of the plastic pipe meet the thickness requirement.

[0086] In one embodiment, the method further includes adjusting the target speed of the extruder screw according to the temperature of the extrusion port, and the target speed of the screw after adjustment is calculated as follows:

[0087] ;

[0088] Where, Indicates the target speed of the screw after adjustment. Indicates the target speed of the screw before adjustment. Indicates the real-time temperature of the extrusion port. Indicates the standard temperature of the extrusion nozzle.

[0089] When the real-time temperature of the extrusion port increases, the fluidity of the plastic will improve and the thickness of the extrudate will become thinner, while when the real-time temperature decreases, the fluidity of the plastic will deteriorate and the thickness of the extrudate will increase. Therefore, in this embodiment, after calculating the target speed of the extruder screw, the target speed of the extruder screw is further adjusted according to the difference between the real-time temperature and the standard temperature, thereby helping to maintain a constant thickness of the produced plastic pipe and further improving the control accuracy of the thickness of the produced plastic pipe.

[0090] In one embodiment, the method further includes: correcting the target speed of the screw after adjustment according to the pressure at the extrusion port, and the screw speed correction value is calculated as follows:

[0091] ;

[0092] Where, Indicates the screw speed correction value, Indicates the pressure value at the extrusion port. Indicates the standard value of the pressure at the extrusion port. Represents the exponential function with base e.

[0093] Pressure during extrusion affects the fluidity of the plastic, and thus the thickness of the extrudate. Increasing pressure improves the fluidity of the plastic, potentially resulting in a thinner extrudate. Conversely, decreasing pressure can lead to a thicker extrudate. In this embodiment, after adjusting the target speed of the extruder screw, the effect of pressure at the extrusion port on the thickness of the produced plastic pipe is further considered. When the pressure at the extrusion port is greater than the standard pressure, the adjusted target speed of the screw is increased, thereby reducing the plastic temperature and increasing the thickness of the plastic pipe, thereby compensating for the effect of pressure at the extrusion port on the thickness of the produced plastic pipe. When the pressure at the extrusion port is less than the standard pressure, the adjusted target speed of the screw is decreased, thereby increasing the plastic temperature and reducing the thickness of the plastic pipe, thereby compensating for the effect of pressure at the extrusion port on the thickness of the produced plastic pipe. In this embodiment, after adjusting the target speed of the screw based on the extrusion port temperature, the effect of pressure at the extrusion port on the thickness of the plastic pipe is further considered. The target speed of the screw is further adjusted based on the pressure at the extrusion port, thereby further improving the control accuracy of the thickness of the produced plastic pipe.

[0094] Example of a fault detection system for a plastic pipe extruder:

[0095] The present invention also provides a fault detection system for a plastic pipe extruder. Figure 4 As shown, the fault detection system for the plastic tube extruder includes a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, the fault detection method for the plastic tube extruder described in the above embodiment is implemented.

[0096] The fault detection system of the plastic tube extruder also includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0097] In the present invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, the computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application, module, or both. Any such computer storage medium can be part of, accessible to, or connectable to a device. Any application or module described in the present invention can be implemented using computer-readable / executable instructions that can be stored or otherwise maintained by such a computer-readable medium.

[0098] In the description of this specification, "multiple" and "several" mean at least two, such as two, three or more, unless otherwise clearly defined.

[0099] While several embodiments of the present invention have been shown and described herein, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Numerous modifications, variations, and alternatives will occur to those skilled in the art without departing from the concept and spirit of the present invention. It should be understood that various alternatives to the embodiments of the present invention described herein may be employed in practicing the present invention.

Claims

1. A method for detecting a fault of a plastic tube extruder, characterized in that: include: The temperature, pressure and extrusion rate of the extrusion port of the extruder are collected according to a preset collection frequency, thereby obtaining a temperature sequence, a pressure sequence and an extrusion rate sequence; Calculating the variance of the temperature sequence, and if the variance is greater than a preset variance threshold, preliminarily determining that noise data exists in the temperature sequence; Fitting the temperature data in the temperature series to obtain a first fitting curve, and calculating the residual corresponding to each temperature data; The temperature data whose residual is greater than a preset residual threshold is determined as suspected noise data; Fitting the pressure data corresponding to the neighborhood temperature data of the suspected noise data to obtain a second fitting curve, and calculating a first Pearson correlation coefficient between the neighborhood temperature data points of the suspected noise data and the pressure data points on the second fitting curve; If the value of the first Pearson correlation coefficient is -1, the suspected noise data is determined to be non-noise data, and the temperature series, pressure series, and extrusion rate series are input into the preset neural network model to obtain the fault prediction result of the extruder; otherwise, the suspected noise data is determined to be real noise data and median filtering is performed on it; The filtered temperature sequence, pressure sequence and extrusion rate sequence are input into the preset neural network model to obtain the fault prediction results of the extruder.

2. The fault detection method for a plastic tube extruder according to claim 1, wherein: The pressure collection methods at the extrusion port include: A pressure sensor is placed at the extrusion port, and the pressure at the extrusion port is calculated based on the value collected by the pressure sensor. The calculation expression is: ； Where, Indicates the pressure at the extrusion port. It represents the value collected by the pressure sensor, and k represents the compensation coefficient, which ranges from 0.05 to 0.

1.

3. The fault detection method for a plastic tube extruder according to claim 2, wherein: The neighborhood range is a range consisting of five data points centered on the suspected noise data point.

4. The fault detection method for a plastic pipe extruder according to claim 1, wherein: Also includes: Calculate the noise performance of each pressure data point in the pressure sequence, and perform median filtering on the pressure data points whose noise performance is greater than the noise performance threshold. The calculation expression is: ； Where, Indicates the noise performance of the i-th pressure data point, represents the exponential function with base e, represents the mean of the pressure data points in the neighborhood of the i-th pressure data point, represents the mean value between the i-th pressure data point and its adjacent pressure data points. represents the second Pearson correlation coefficient between the pressure data point and the corresponding temperature data point within the neighborhood of the i-th pressure data point; Indicates the pressure data point with the largest value in the neighborhood of the i-th pressure data point.

5. The fault detection method for a plastic pipe extruder according to claim 1, wherein: The training method of the neural network model includes: The temperature series, pressure series, and extrusion rate series of the extruder at historical moments are collected and labeled to form a training set. Input the data in the training set into the neural network model to obtain the predicted value; Compare the predicted value with the true value and calculate the loss value using the mean square error function; Use the chain rule to calculate the gradient of the loss function with respect to each parameter, and then backpropagate these gradients back to each layer in the network; and update the weights and bias of each neuron according to the gradient value; Iteratively updates the weights and biases of each neuron until the loss converges or the preset number of iterations is reached.

6. The fault detection method for a plastic tube extruder according to claim 1, wherein: Also includes: Real-time acquisition of extruder screw speed; Call the target extrusion rate according to the real-time extrusion rate and the thickness requirement of the plastic pipe, and calculate the difference between the target extrusion rate and the real-time extrusion rate; Performing PID calculation on the difference to obtain the target speed of the extruder screw; The speed of the extruder screw is controlled by PID according to the difference between the target speed and the real speed of the extruder screw.

7. The fault detection method for a plastic tube extruder according to claim 1, wherein: Also includes: The target speed of the extruder screw is adjusted according to the temperature of the extrusion port. The target speed calculation expression of the adjusted screw is: ； Where, Indicates the target speed of the screw after adjustment. Indicates the target speed of the screw before adjustment. Indicates the real-time temperature of the extrusion port. Indicates the standard temperature of the extrusion nozzle.

8. The method for detecting a fault of a plastic tube extruder according to claim 7, wherein: Also includes: The target speed of the screw after adjustment is corrected according to the pressure at the extrusion port. The calculation expression of the screw speed correction value is: ； Where, Indicates the screw speed correction value, Indicates the pressure value at the extrusion port. Indicates the standard value of the pressure at the extrusion port. Represents the exponential function with base e.

9. A fault detection system for a plastic tube extruder, comprising a processor and a memory, wherein the memory stores computer program instructions, characterized in that: When the computer program instructions are executed by the processor, the fault detection method for the plastic tube extruder according to any one of claims 1 to 8 is implemented.

Images