A fault prediction method and system for a pine nut full-automatic flow processing equipment

Abstract

The application provides a fault prediction method and system for pine nut full-automatic flow processing equipment, and relates to the technical field of data analysis, comprising: collecting pine nut images in a plurality of screening areas through a camera; obtaining an actual quantity set and calculating an actual weight set; taking a standard weight threshold as a benchmark, performing consistency analysis on the actual weight set, obtaining a plurality of screening consistency coefficients; continuously monitoring a plurality of screening consistency coefficient sequences, and performing mean value calculation, fluctuation analysis and trend fitting; if any consistency coefficient mean value, fluctuation coefficient or trend coefficient does not satisfy a preset coefficient threshold, fault early warning is performed. Through the application, the technical problem that the existing technology is difficult to identify comprehensive faults due to complex and diverse fault types, resulting in poor accuracy of equipment fault prediction, can be solved. The application predicts whether the equipment has failed from the screening results of pine nuts, comprehensively detects the fault state of the equipment, and improves the accuracy of equipment fault prediction.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis, in particular to a fault prediction method and system of a pine nut full-automatic flow processing equipment. BACKGROUND

[0002] The pine nut full-automatic flow processing equipment is an equipment for realizing automatic processing of pine nuts from raw material feeding, sorting, shelling, kernel-shell separation, winnowing, screening, weighing, packaging and the like, and is used for nut processing to improve processing efficiency, ensure product quality and reduce labor cost. The existing method can predict and judge the running state of the equipment to a certain extent by trend analysis based on vibration, temperature, current and other sensor data, or analysis based on equipment running logs and historical fault data. However, the equipment structure is complex, and the fault types that may occur are often complex and diverse. The existing fault prediction method cannot comprehensively identify all types of faults, especially the initial and minor abnormalities that are often difficult to detect.

[0003] In summary, the prior art has the technical problem that the fault types are complex and diverse, it is difficult to identify comprehensive faults, especially the initial and minor abnormalities that are often difficult to detect, resulting in poor accuracy of equipment fault prediction. SUMMARY

[0004] The purpose of the present application is to provide a fault prediction method and system of a pine nut full-automatic flow processing equipment to solve the technical problem in the prior art that the fault types are complex and diverse, it is difficult to identify comprehensive faults, especially the initial and minor abnormalities that are often difficult to detect, resulting in poor accuracy of equipment fault prediction.

[0005] In order to achieve the above purpose, the present application provides a fault prediction method and system of a pine nut full-automatic flow processing equipment.

[0006] In a first aspect, the application provides a fault prediction method for a pine nut full-automatic flow processing equipment, which is realized by a fault prediction system for the pine nut full-automatic flow processing equipment. The fault prediction method comprises: collecting pine nut images in a plurality of screening areas by a camera to obtain a pine nut image set, wherein the pine nuts in each screening area have different specifications, and the specifications include a standard weight threshold; performing pine nut feature recognition based on the pine nut image set to obtain an actual quantity set and calculate an actual weight set; performing consistency analysis on the actual weight set based on the standard weight threshold to obtain a plurality of screening consistency coefficients of the plurality of screening areas; continuously monitoring a plurality of screening consistency coefficient sequences within a preset time range, and performing mean value calculation, fluctuation analysis and trend fitting to obtain a plurality of consistency coefficient means, a plurality of fluctuation coefficients and a plurality of trend coefficients; and performing fault early warning on a pine nut screening device if any of the consistency coefficient means, the fluctuation coefficients or the trend coefficients does not satisfy a preset coefficient threshold.

[0007] Optionally, according to historical screening records of pine nut screening devices of the same type, a sample pine nut image set is collected, and actual pine nut quantities of different sample pine nut images are obtained as sample pine nut quantities to obtain a sample quantity set; the sample pine nut image set and the sample quantity set are used to train and test a convolutional neural network until convergence, and a pine nut quantity recognition plug-in is obtained; the pine nut quantity recognition plug-in is used to perform pine nut feature recognition on a plurality of pine nut images of the pine nut image set respectively, and an actual quantity set is output.

[0008] Optionally, a plurality of total weights of pine nuts in a plurality of screening areas are obtained by weighing to construct a pine nut total weight set; mapping calculation is performed based on the pine nut total weight set and the actual quantity set to output an actual weight set, wherein the actual weight is the average weight of a single pine nut.

[0009] Optionally, a first standard weight threshold of a first screening area is randomly selected, and a first actual weight is obtained; a median of the first standard weight threshold is taken as a first standard weight mean value, a deviation ratio of the first standard weight mean value and the first actual weight is calculated, and an inverse of the deviation ratio is taken as a first screening consistency coefficient and added to the plurality of screening consistency coefficients.

[0010] Optionally, within a preset time range, image collection and consistency analysis are continuously performed to obtain a plurality of screening consistency coefficient sequences; mean value calculation is performed on the plurality of screening consistency coefficient sequences respectively to obtain a plurality of consistency coefficient means; fluctuation analysis and trend fitting are performed on the plurality of screening consistency coefficient sequences respectively to output a plurality of fluctuation coefficients and a plurality of trend coefficients.

[0011] Optionally, standard deviation of each of the plurality of screening consistency coefficient sequences is calculated to obtain a plurality of consistency coefficient standard deviations; and a plurality of fluctuation coefficients is calculated according to the plurality of consistency coefficient standard deviations and a plurality of consistency coefficient means, wherein the fluctuation coefficient is a ratio of the consistency coefficient standard deviation to the consistency coefficient mean.

[0012] Optionally, a two-dimensional coordinate system is constructed with a monitoring time node as an X-axis and a screening consistency coefficient as a Y-axis; distribution and curve fitting of the plurality of screening consistency coefficient sequences are performed in the two-dimensional coordinate system to obtain a plurality of trend curves; and the plurality of trend curves is identified to obtain a plurality of trend coefficients, wherein the trend coefficient is an overall rising or falling proportion.

[0013] Optionally, the preset coefficient threshold includes a consistency coefficient threshold, a fluctuation coefficient threshold and a trend falling proportion threshold, wherein the preset coefficient threshold is set based on the pine nut specification.

[0014] Optionally, if the consistency coefficient mean of any screening area is less than the consistency coefficient threshold and / or the fluctuation coefficient is greater than the fluctuation coefficient threshold and / or the falling proportion of the trend coefficient is greater than the trend falling proportion threshold, a fault warning is given to the pine nut screening device.

[0015] In a second aspect, the present application also provides a fault prediction system of a pine nut full-automatic flow processing equipment, which is used to execute the fault prediction method of the pine nut full-automatic flow processing equipment as described in the first aspect. The fault prediction system of the pine nut full-automatic flow processing equipment includes: an image acquisition module, which is used to acquire pine nut images in a plurality of screening areas through a camera to obtain a pine nut image set, wherein the pine nut specification of each screening area is different, and the pine nut specification includes a standard weight threshold; a weight calculation module, which is used to perform pine nut feature recognition according to the pine nut image set to obtain an actual quantity set and calculate an actual weight set; a consistency analysis module, which is used to take the standard weight threshold as a reference to perform consistency analysis on the actual weight set to obtain a plurality of screening consistency coefficients of a plurality of screening areas; a dynamic monitoring analysis module, which is used to continuously monitor a plurality of screening consistency coefficient sequences within a preset time range and perform mean calculation, fluctuation analysis and trend fitting to obtain a plurality of consistency coefficient means, a plurality of fluctuation coefficients and a plurality of trend coefficients; and a threshold judgment module, which is used to give a fault warning to the pine nut screening device if any consistency coefficient mean, fluctuation coefficient or trend coefficient does not meet a preset coefficient threshold.

[0016] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0017] The pine nut image set is acquired by collecting pine nut images in a plurality of screening areas through a camera, wherein the pine nut specifications of each screening area are different, and the pine nut specifications include a standard weight threshold; pine nut feature recognition is performed according to the pine nut image set, an actual quantity set is acquired, and an actual weight set is calculated; the actual weight set is analyzed for consistency with the standard weight threshold, a plurality of screening consistency coefficients of the plurality of screening areas are acquired; a plurality of screening consistency coefficient sequences within a preset time range are continuously monitored and acquired, and mean value calculation, fluctuation analysis and trend fitting are performed to obtain a plurality of consistency coefficient means, a plurality of fluctuation coefficients and a plurality of trend coefficients; if any consistency coefficient mean, fluctuation coefficient or trend coefficient does not satisfy a preset coefficient threshold, a fault warning is given to the pine nut screening device. That is, by analyzing the running state of the equipment according to the fluctuation of the weight of the pine nuts in the screening process, the consistency coefficients of the plurality of screening areas are monitored, slight but continuous deviations are found in advance in the screening process in different areas and different specifications, complex implicit faults such as equipment degradation, aging and imbalance are captured, and early warning is given, the accuracy and efficiency of the fault prediction of the full-automatic pine nut processing equipment are improved.

[0018] The above description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application, the specific embodiments of the present application can be implemented according to the content of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments or the prior art description. Obviously, the drawings in the following description are only exemplary, and other drawings can be obtained by those skilled in the art without creating additional labor.

[0020] Figure 1 The flowchart of the fault prediction method of the full-automatic pine nut processing equipment of the present application;

[0021] Figure 2 The structural diagram of the fault prediction system of the full-automatic pine nut processing equipment of the present application.

[0022] Explanation of reference signs: image acquisition module 11, weight calculation module 12, consistency analysis module 13, dynamic monitoring and analysis module 14, threshold judgment module 15. DETAILED DESCRIPTION

[0023] The application provides a pine nut full-automatic flow processing equipment fault prediction method and system, which solves the technical problem that the existing technology is difficult to identify comprehensive faults, especially the initial and slight abnormalities are difficult to be detected, resulting in poor accuracy of equipment fault prediction. By predicting whether a fault occurs from the screening results of pine nuts, the running state of the equipment is comprehensively analyzed, the image of each screening area and each specification of pine nuts is collected and analyzed, the consistency coefficient of multiple screening areas is monitored, and slight but continuous deviations are found in the screening process of different areas and different specifications. The complex implicit faults such as equipment degradation, aging and imbalance of calibration are captured, and early warning is given. The accuracy and efficiency of the pine nut full-automatic flow processing equipment fault prediction are improved.

[0024] The technical solutions in the application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are only some of the embodiments of the application, not all embodiments of the application. It should be understood that the application is not limited by the example embodiments described herein. Based on the embodiments of the application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the application. In addition, it should be noted that, for convenience of description, only parts related to the application are shown in the drawings, not all.

[0025] Embodiment one, please refer to the attached Figure 1 The application provides a pine nut full-automatic flow processing equipment fault prediction method, which is applied to a pine nut full-automatic flow processing equipment fault prediction system. The method specifically includes the following steps:

[0026] S100: Collect pine nut images in a plurality of screening areas through a camera to obtain a pine nut image set, wherein the pine nut specifications of each screening area are different, and the pine nut specifications include a standard weight threshold.

[0027] Specifically, by installing an industrial camera above each screening area, real-time image acquisition of pine nuts in the area is carried out, a pine nut image set of different areas is generated, and a complete pine nut image set is formed, including the images of several pine nuts in several screening areas. Because the specifications of pine nuts screened by each screening area are different, the characteristics (such as weight and size) of pine nuts corresponding to each image set are also different. The screening area is a separate screening unit preset for pine nuts of different specifications. Each screening area is specially configured according to the size, weight and other characteristics of the product to ensure that the same group of pine nuts meets the predetermined standard. Each area has a corresponding standard weight threshold and size threshold. For example, for large-size pine nuts, the standard size may be specified as a diameter of more than 10 mm and a weight of between 1.6 g and 1.8 g. By using a high-resolution camera to capture pine nut images without contact, the actual number of pine nuts in each screening area is obtained, and the weight data is calculated to quickly determine whether there is a deviation from the standard, thereby providing a basis for fault warning.

[0028] S200: According to the pine nut image set, the pine nut feature recognition is carried out, the actual number set is obtained, and the actual weight set is calculated.

[0029] Further, the S200 of the present application comprises:

[0030] According to the historical screening records of the same type of pine nut screening device, a sample pine nut image set is collected, and the actual number of different sample pine nut images is obtained as the sample pine nut number, to obtain a sample number set; taking the sample pine nut image as input and the sample number as supervision, the convolutional neural network is trained and tested using the sample pine nut image set and the sample number set until convergence, to obtain a pine nut number recognition plug-in; the pine nut number recognition plug-in is used to recognize the pine nut characteristics of several pine nut images in the pine nut image set, and an actual number set is output.

[0031] Further, the present application further comprises the following steps:

[0032] The total weight of several pine nuts in several screening areas is obtained by weighing, and a pine nut total weight set is constructed; according to the mapping calculation of the pine nut total weight set and the actual number set, an actual weight set is output, wherein the actual weight is the average weight of a single pine nut.

[0033] Specifically, all screening data records collected by the same type of pine nut screening device in actual operation are obtained, including the image, quantity and other data of pine nuts in each area. According to the past screening records of the same pine nut screening device, a large number of labeled sample images are extracted, each image is attached with the corresponding actual pine nut quantity data, and a sample pine nut image set and a sample pine nut quantity set are obtained. For example, assuming that the sample image set contains 1000 images, the actual pine nut quantities of multiple images can be determined as follows: image 1 is 18, image 2 is 20, image 3 is 21, and so on to form the sample quantity set.

[0034] The sample pine nut image is taken as input, and the sample quantity is taken as supervision to train the pine nut quantity recognition plug-in. The sample pine nut image set and the corresponding sample quantity set obtained from the historical screening records are preprocessed, and all images are uniformly adjusted to a fixed size (for example, 256x256 pixels) to ensure consistent input size. The pixel values are normalized to map [0, 255] to the range [0, 1]. In order to enhance the generalization ability of the model, data augmentation is performed on the training set, such as random rotation (±15°), translation, scaling, and flipping. The data set is divided into a training set and a validation set in the ratio of 80 / 20 to monitor the real-time effect during the training process. A regression model architecture based on convolutional layer-pooling layer-full connection layer is selected, including input layer, convolutional layer, pooling layer, full connection layer and output layer. For a batch, the preprocessed image is input into the model architecture for forward propagation to obtain the corresponding predicted quantity, the loss is calculated by the mean square error loss function, the Adam optimizer is used for back propagation, and the gradient is transmitted and the network weights of each layer are updated. After each epoch, the average loss on the entire training set is calculated as the training error indicator. After each epoch, the model performance is evaluated using the validation set, and the MSE loss value of the validation set is calculated. Observe whether the validation set loss decreases, if the decrease is less than the preset threshold (for example, the decrease is less than 0.005), it is close to convergence. For example, the training loss of the initial epoch (epoch1) is 0.18, and the validation loss is 0.17; after 50 epochs, the training loss decreases to 0.06, and the validation loss is 0.065; at the 150th epoch: the training loss is 0.032, and the validation loss remains at 0.034; from the 160th epoch, the validation loss fluctuates between 0.033 and 0.035, and there is no significant decrease, which meets the early stopping condition.

[0035] When the validation set loss is stable at a low value (for example, 0.033-0.035) for multiple epochs and does not decrease significantly, it is considered that the model parameters have been fully converged. At this time, the model has learned the mapping relationship between the input image and the actual pine nut quantity. The trained model is saved as a standard interface and deployed to the production environment.

[0036] The trained pine nut quantity recognition plug-in is embedded into the data processing module of the pine nut screening device, a plurality of pine nut images in the pine nut image set are input into the pine nut quantity recognition plug-in for feature recognition, pine nut quantity in each region actually detected is output by plug-in analysis of pine nut features in the image, an actual quantity set is formed, and automatic target counting is realized.

[0037] In each screening region, the sum of all pine nut weights measured by the weighing device (such as a pressure sensor) is obtained, and the total weight of pine nuts in each region is obtained, thereby obtaining a total weight set. The total weight set is a total weight data set constructed according to the region after weighing the pine nuts in the plurality of screening regions. For example, the weighing result of region 1 is 500 grams, the weighing result of region 2 is 520 grams, and so on.

[0038] For each region, the actual number of pine nuts in the region has been obtained by image recognition, forming an actual quantity set. Through mapping calculation, for each region, the average weight of a single pine nut is obtained by dividing the total weight of the region by the actual quantity. For example, assuming that the actual quantity set (unit: grams) in a plurality of screening regions is {25, 26, 24, 25, 25}, and the total weight set (unit: grams) of pine nuts corresponding to the actual quantity set is {500, 520, 480, 510, 495}, then the actual weight set (unit: grams / pine nut) is {20, 20, 20, 20.4, 19.8}. By weighing the total weight of pine nuts, combined with actual quantity mapping calculation, the average weight of a single pine nut in each screening region is accurately determined, and when the data of a certain region is obviously abnormal, it can be regarded as an indication signal of device failure or error, which is helpful for timely maintenance.

[0039] S300: Based on the standard weight threshold, consistency analysis is performed on the actual weight set to obtain a plurality of screening consistency coefficients of a plurality of screening regions.

[0040] Further, the S300 of the present application comprises:

[0041] A first standard weight threshold of a first screening region is randomly selected, and a first actual weight is obtained; the median of the first standard weight threshold is taken as a first standard weight average, the deviation proportion of the first standard weight average and the first actual weight is calculated, and the reciprocal of the deviation proportion is taken as a first screening consistency coefficient and added to the plurality of screening consistency coefficients.

[0042] Specifically, the standard weight threshold predefines the required range of pine nut weight for each screening region, such as stipulating that the weight of pine nuts in a certain region should be between 19 grams and 21 grams. The first standard weight threshold corresponding to the first screening region is randomly selected within the standard weight threshold, which defines an important requirement that pine nuts in the first screening region should meet. At the same time, the first actual weight corresponding to the actual weight set is obtained.

[0043] For the first screening area, the median of the first standard weight threshold is taken as the first standard weight average, such as the median between 19 grams and 21 grams is 20 grams. The deviation ratio of the first standard weight average and the first actual weight is calculated, and the ratio of the absolute value of the first standard weight average and the first actual weight is calculated, and the relative deviation ratio is obtained. According to the calculated deviation ratio, take its reciprocal as the first screening consistency coefficient, and add it to the screening consistency coefficients. The screening consistency coefficient is used to reflect the consistency between the actual weight and the standard weight.

[0044] For other screening areas, the same consistency analysis is performed as described above, and a plurality of screening consistency coefficients of a plurality of screening areas are obtained. The consistency analysis is a quantitative comparison analysis method, which compares the actual average weight of each screening area with the corresponding standard weight threshold, calculates the deviation, and measures the matching degree between the actual situation of each screening area and the preset standard. It is usually defined as the reciprocal of the deviation ratio, that is, the smaller the deviation, the larger the consistency coefficient, and the more consistent, and vice versa, the consistency coefficient is smaller.

[0045] Through the deviation ratio and the consistency coefficient calculation, the deviation between the actual weight of the pine nut in each screening area and the preset standard is objectively quantified. The higher the value of the consistency coefficient indicates that the actual weight in the area is more consistent with the standard, and the higher the screening accuracy of the equipment; if the value is low, it may indicate that there is an abnormal or fault risk.

[0046] S400: Continuously monitor to obtain a plurality of screening consistency coefficient sequences within a preset time range, and perform mean value calculation, fluctuation analysis and trend fitting to obtain a plurality of consistency coefficient means, a plurality of fluctuation coefficients and a plurality of trend coefficients.

[0047] Further, the S400 of the present application comprises:

[0048] Within a preset time range, continuously perform image acquisition and consistency analysis to obtain a plurality of screening consistency coefficient sequences; respectively calculate the mean value of the plurality of screening consistency coefficient sequences to obtain a plurality of consistency coefficient means; respectively perform fluctuation analysis and trend fitting on the plurality of screening consistency coefficient sequences, and output a plurality of fluctuation coefficients and a plurality of trend coefficients.

[0049] Specifically, the preset time range is a continuous time set in advance in the pine nut screening process, such as a monitoring period of 1 hour, 8 hours or a day. Through the industrial camera installed in each screening area, the pine nut image data of several areas is continuously or regularly collected within the preset time range. At the same time, a consistency analysis method is used for each collected image, that is, the actual number of pine nuts is determined through image collection, the total weight of pine nuts is obtained through weighing, and the actual weight is determined by mapping calculation according to the total weight of pine nuts and the actual number of pine nuts. It is compared and calculated with the standard weight threshold of the area to obtain the corresponding screening consistency coefficient. A screening consistency coefficient is obtained after each collection, forming a time series data, obtaining several screening consistency coefficient sequences, including the screening consistency coefficient sequences corresponding to several screening areas.

[0050] The mean value of each screening consistency coefficient sequence is calculated to obtain the mean value of each consistency coefficient, that is, the arithmetic mean of all consistency coefficients of each area is calculated to obtain the stable level of the overall screening consistency of each area. According to the mean value of the consistency coefficient, the standard deviation of the screening consistency coefficient sequence of each area is calculated, and according to the ratio of the standard deviation and the mean value, several fluctuation coefficients of several screening areas are obtained.

[0051] A two-dimensional coordinate system is constructed according to each monitoring time node and its corresponding screening consistency coefficient, curve fitting is performed, and several trend curves of several screening areas are obtained, and corresponding coefficients are obtained to reflect whether the trend is rising or falling. The trend coefficient is a quantitative index for describing the trend of data over time.

[0052] Through fluctuation analysis and trend analysis, the running state of the pine nut screening equipment in the preset time range is comprehensively evaluated, and the consistency and long-term change trend of the equipment operation are revealed respectively. If the fluctuation coefficient is too high or the trend coefficient shows a large downward trend, it means that the equipment needs to be maintained or adjusted. By responding to these early warning signals in a timely manner, equipment failures can be reduced and production efficiency can be improved.

[0053] Further, the application further includes the following steps:

[0054] The standard deviation of each screening consistency coefficient sequence is calculated to obtain the standard deviation of each consistency coefficient. According to the standard deviation of the consistency coefficient and the mean value of the consistency coefficient, a fluctuation coefficient is calculated, wherein the fluctuation coefficient is the ratio of the standard deviation of the consistency coefficient to the mean value of the consistency coefficient.

[0055] Specifically, the consistency coefficient standard deviation is the dispersion degree between each data value and the sequence average in the consistency coefficient sequence obtained by a certain region. The smaller the standard deviation, the smaller the data fluctuation and the more stable the result; the larger the standard deviation, the larger the fluctuation. For each screening region, take the consistency coefficient sequence obtained in the preset time range, and calculate the standard deviation of the sequence by statistical method. Calculate the square of the difference between each data and the average, take the average, and take the square root to get the standard deviation.

[0056] After obtaining the standard deviation and average of the consistency coefficient sequence, the fluctuation coefficient corresponding to each screening region is calculated, and the fluctuation coefficient is obtained by the ratio of the consistency coefficient standard deviation to the consistency coefficient average, and a plurality of fluctuation coefficients are obtained. Each screening region will obtain a fluctuation coefficient, which reflects the relative amplitude of the screening consistency coefficient fluctuation with time in the region. The plurality of fluctuation coefficients includes the fluctuation coefficients corresponding to the plurality of screening regions.

[0057] Exemplarily, assuming that a certain screening region has collected 10 consistency coefficients in a preset time, the specific values are [50, 52, 49, 51, 50, 50, 53, 48, 50, 51], the average is 50.4, the standard deviation (the square of the difference between each value and the average is calculated, and then the average is taken, and then the square root is taken) is 1.84, and the fluctuation coefficient is 0.0365. The stability of the consistency index of each screening region is determined by the ratio of the standard deviation and the average (i.e. the fluctuation coefficient). Similarly, the corresponding fluctuation coefficients of other screening regions are also calculated respectively, and a plurality of fluctuation coefficient sets are formed.

[0058] By calculating the standard deviation of the consistency coefficient sequence collected continuously, the consistency coefficient standard deviation of each screening region is obtained, and then the fluctuation coefficient is calculated in combination with the consistency coefficient average, the fluctuation degree of each screening region is quantified, and it is determined whether the equipment is running stably. If the fluctuation coefficient is high, it means that the running state of the equipment is not stable enough, and needs to be adjusted or maintained.

[0059] Further, the present application further comprises the following steps:

[0060] A two-dimensional coordinate system is constructed with the monitoring time node as the X-axis and the screening consistency coefficient as the Y-axis; in the two-dimensional coordinate system, the plurality of screening consistency coefficient sequences are distributed and curve-fitted to obtain a plurality of trend curves; the plurality of trend curves are identified to obtain a plurality of trend coefficients, wherein the trend coefficient is the overall rising or falling proportion.

[0061] Specifically, the preset monitoring time node (i.e., the time of collecting images) is set as the X axis, and the screening consistency coefficient calculated at each time node is set as the Y axis, a two-dimensional plane data graph is constructed, and the data trend over time is displayed. Using statistical methods, trend analysis is performed on the discrete data points in the two-dimensional coordinate graph, and one or more curves are obtained through fitting methods (such as linear regression or polynomial fitting), which describe the overall trend of data points over time. The consistency coefficient sequence in the constructed two-dimensional coordinate system is fitted using a curve fitting method, such as a quadratic or polynomial curve fitting method, and a fitting function is selected to minimize the error.

[0062] Each group of time and corresponding consistency coefficient is mapped to a point on a two-dimensional plane, and the data of all screening regions can form their own coordinate system independently or be plotted in the same graph for comparison. The trend relationship between these discrete points is found, and a trend curve is obtained using a fitting algorithm. Each group of consistency coefficient sequences will generate a trend curve, and multiple screening regions will obtain several trend curves to reflect the data trend.

[0063] The slope (or other parameters reflecting the overall change rate) extracted from the trend curve obtained by fitting is defined as the trend coefficient: if the slope k > 0, the trend coefficient is positive, indicating an overall increase; if the slope k < 0, the trend coefficient is negative, indicating an overall decrease; the absolute value of the trend coefficient reflects the change rate. Each group of consistency coefficient sequences will generate a trend curve, and multiple screening regions will obtain several trend curves to reflect whether there is an anomaly in each region data.

[0064] For example, assume that the screening consistency coefficients collected at 10 consecutive time nodes in a screening region are as follows: [50, 50.5, 51, 50.8, 51.2, 51.5, 51.3, 51.7, 52, 52.2], and the corresponding time nodes (in minutes) are T = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10], a straight line y = 0.22x + 50 is fitted using the least squares method, where x is the time and y is the screening consistency coefficient.

[0065] Within a preset time range, the screening consistency coefficient of each time node collected is taken as the Y axis data, and the monitoring time node is taken as the X axis data to construct a two-dimensional coordinate system. Through distribution analysis and curve fitting (such as linear regression) of these data points, a trend curve is obtained, and a trend coefficient is extracted therefrom, revealing the change direction of the consistency coefficient of each screening region over time, helping to analyze whether the equipment is running stably or whether there is a gradual deterioration problem. The trend coefficient (such as the slope of the fitted straight line) reflects the overall increase or decrease proportion of the screening consistency coefficient, which is an important indicator for judging the change of equipment running state and warning of fault risk.

[0066] S500: If any of the consistency coefficient mean, fluctuation coefficient, or trend coefficient does not meet the preset coefficient threshold, a fault warning is given for the pine nut screening device.

[0067] Further, the present application S500 comprises:

[0068] The preset coefficient threshold includes a consistency coefficient threshold, a fluctuation coefficient threshold, and a trend decline ratio threshold, wherein the preset coefficient threshold is set based on the pine nut specification.

[0069] Further, the present application further comprises the following steps:

[0070] If the consistency coefficient mean of any screening area is less than the consistency coefficient threshold and / or the fluctuation coefficient is greater than the fluctuation coefficient threshold and / or the decline ratio of the trend coefficient is greater than the trend decline ratio threshold, a fault warning is given for the pine nut screening device.

[0071] Specifically, for each screening area, the consistency coefficient mean, consistency fluctuation coefficient, and consistency trend coefficient are calculated and compared with the respective preset threshold. As long as any one of the trigger conditions is met, that is, any one of the consistency coefficient mean, fluctuation coefficient, and trend coefficient does not meet the preset coefficient threshold, a fault warning signal is issued.

[0072] According to the pine nut specification, the preset coefficient threshold is determined to judge whether the equipment is running normally. The preset coefficient threshold includes a consistency coefficient threshold, a fluctuation coefficient threshold, and a trend decline ratio threshold, which are used to judge whether the consistency coefficient mean, fluctuation coefficient, or trend coefficient is within an acceptable range.

[0073] If the consistency coefficient mean of any screening area is less than the consistency coefficient threshold, or the fluctuation coefficient of any screening area exceeds the fluctuation coefficient threshold, or the decline ratio of the trend coefficient of any screening area exceeds the trend decline ratio threshold, as long as any one of the conditions is met, a fault warning is triggered immediately to notify the maintenance personnel to check the equipment. The preset coefficient threshold provides a benchmark for judging whether the equipment is running normally, helping to discover potential problems in a timely manner, thereby improving production efficiency and product quality.

[0074] The screening area that triggered the warning is checked in detail to identify possible causes of failure. Based on the investigation results, necessary maintenance and adjustment are carried out to restore the normal operation of the screening device. The warning information and processing results are recorded for subsequent analysis and improvement. By responding to the fault warning in a timely manner, maintenance personnel can take appropriate measures to avoid greater losses caused by equipment failure.

[0075] In summary, the fault prediction method for a pine nut full-automatic flow processing equipment provided by the present application has the following technical effects:

[0076] The pine nut image set is acquired by collecting pine nut images in a plurality of screening areas through a camera, wherein the pine nut specifications of each screening area are different, and the pine nut specifications include a standard weight threshold; pine nut feature recognition is performed according to the pine nut image set to acquire an actual quantity set, and an actual weight set is calculated; the actual weight set is analyzed for consistency with the standard weight threshold as a reference to acquire a plurality of screening consistency coefficients of the plurality of screening areas; a plurality of screening consistency coefficient sequences within a preset time range are continuously monitored and acquired, and mean value calculation, fluctuation analysis and trend fitting are performed to obtain a plurality of consistency coefficient means, a plurality of fluctuation coefficients and a plurality of trend coefficients; if any consistency coefficient mean, fluctuation coefficient or trend coefficient does not satisfy a preset coefficient threshold, a fault warning is given to the pine nut screening device. That is, by analyzing the running state of the equipment according to the fluctuation of the weight of the pine nuts in the screening process, the consistency coefficients of the plurality of screening areas are monitored, and in the screening process in different areas and different specifications, slight but continuous deviations are found in advance, complex implicit faults such as equipment degradation, aging and imbalance of calibration are captured, and early warning is given, so that the accuracy and efficiency of fault prediction of the pine nut full-automatic flow processing equipment are improved.

[0077] In the same inventive concept as the fault prediction method of the pine nut full-automatic flow processing equipment in the foregoing embodiment one, the present application also provides a fault prediction system of a pine nut full-automatic flow processing equipment. Please refer to the accompanying drawings Figure 2 The fault prediction system of the pine nut full-automatic flow processing equipment comprises:

[0078] The image acquisition module 11 is configured to acquire a pine nut image set by collecting pine nut images in a plurality of screening areas through a camera, wherein the pine nut specifications of each screening area are different, and the pine nut specifications include a standard weight threshold; the weight calculation module 12 is configured to perform pine nut feature recognition according to the pine nut image set to acquire an actual quantity set, and calculate an actual weight set; the consistency analysis module 13 is configured to analyze the actual weight set for consistency with the standard weight threshold as a reference to acquire a plurality of screening consistency coefficients of the plurality of screening areas; the dynamic monitoring and analysis module 14 is configured to continuously monitor and acquire a plurality of screening consistency coefficient sequences within a preset time range, and perform mean value calculation, fluctuation analysis and trend fitting to obtain a plurality of consistency coefficient means, a plurality of fluctuation coefficients and a plurality of trend coefficients; and the threshold judgment module 15 is configured to give a fault warning to the pine nut screening device if any consistency coefficient mean, fluctuation coefficient or trend coefficient does not satisfy a preset coefficient threshold.

[0079] Further, the weight calculation module 12 in the fault prediction system of the pine nut full-automatic flow processing equipment is further configured to:

[0080] According to the historical screening record of the same kind of pine nut screening device, a sample pine nut image set is collected, and the actual number of different sample pine nut images is obtained as the sample pine nut quantity, to obtain a sample quantity set; taking the sample pine nut image as input and the sample quantity as supervision, the sample pine nut image set and the sample quantity set are used to train and test the convolutional neural network until convergence, to obtain a pine nut quantity recognition plug-in; the pine nut quantity recognition plug-in is used to respectively identify the pine nut characteristics of a plurality of pine nut images in the pine nut image set, and an actual quantity set is output.

[0081] Further, the weight calculation module 12 in the fault prediction system of the pine nut full-automatic flow processing equipment is further used for:

[0082] The total weight of a plurality of pine nuts in a plurality of screening areas is obtained by weighing, and a pine nut total weight set is constructed; the actual weight set is output by mapping calculation according to the pine nut total weight set and the actual quantity set, wherein the actual weight is the average weight of a single pine nut.

[0083] Further, the consistency analysis module 13 in the fault prediction system of the pine nut full-automatic flow processing equipment is further used for:

[0084] A first standard weight threshold of a first screening area is randomly selected, and a first actual weight is obtained; the median of the first standard weight threshold is taken as a first standard weight average, the deviation proportion of the first standard weight average and the first actual weight is calculated, and the reciprocal of the deviation proportion is taken as a first screening consistency coefficient, which is added to a plurality of screening consistency coefficients.

[0085] Further, the dynamic monitoring analysis module 14 in the fault prediction system of the pine nut full-automatic flow processing equipment is further used for:

[0086] Within a preset time range, image acquisition and consistency analysis are continuously performed to obtain a plurality of screening consistency coefficient sequences; a plurality of consistency coefficient averages are obtained by respectively performing mean value calculation on the plurality of screening consistency coefficient sequences; a plurality of fluctuation coefficients and a plurality of trend coefficients are output by respectively performing fluctuation analysis and trend fitting on the plurality of screening consistency coefficient sequences.

[0087] Further, the dynamic monitoring analysis module 14 in the fault prediction system of the pine nut full-automatic flow processing equipment is further used for:

[0088] A plurality of consistency coefficient standard deviations are obtained by respectively performing standard deviation calculation on the plurality of screening consistency coefficient sequences; a plurality of fluctuation coefficients are calculated according to the plurality of consistency coefficient standard deviations and the plurality of consistency coefficient averages, wherein the fluctuation coefficient is the ratio of the consistency coefficient standard deviation to the consistency coefficient average.

[0089] Further, the dynamic monitoring analysis module 14 in the pine nut full-automatic flow processing equipment fault prediction system is further used for:

[0090] Taking the monitoring time node as the X-axis and the screening consistency coefficient as the Y-axis, a two-dimensional coordinate system is constructed; in the two-dimensional coordinate system, the distribution and curve fitting of the plurality of screening consistency coefficient sequences are performed to obtain a plurality of trend curves; the plurality of trend curves are identified to obtain a plurality of trend coefficients, wherein the trend coefficient is the overall rising or falling proportion.

[0091] Further, the threshold judgment module 15 in the pine nut full-automatic flow processing equipment fault prediction system is further used for:

[0092] The preset coefficient threshold includes a consistency coefficient threshold, a fluctuation coefficient threshold, and a trend falling proportion threshold, wherein the preset coefficient threshold is set based on the pine nut specification.

[0093] Further, the threshold judgment module 15 in the pine nut full-automatic flow processing equipment fault prediction system is further used for:

[0094] If the consistency coefficient mean of any screening area is less than the consistency coefficient threshold and / or the fluctuation coefficient is greater than the fluctuation coefficient threshold and / or the falling proportion of the trend coefficient is greater than the trend falling proportion threshold, the pine nut screening device is faulted.

[0095] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. Figure 1 The pine nut full-automatic flow processing equipment fault prediction method and specific examples in embodiment one are also applicable to the pine nut full-automatic flow processing equipment fault prediction system in this embodiment. Through the foregoing detailed description of the pine nut full-automatic flow processing equipment fault prediction method, those skilled in the art can clearly understand the pine nut full-automatic flow processing equipment fault prediction system in this embodiment. Therefore, in order to make the specification simple, it will not be described in detail here. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant part is referred to the method part description.

[0096] The above description of the disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0097] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the application, the application can be practiced otherwise than as specifically set forth herein.

Claims

1. A fault prediction method for a pine nut full-automatic flow processing equipment, characterized in that, The method comprises the following steps: Collecting pine nut images in several screening areas through a camera to obtain a pine nut image set, wherein the pine nut specifications in each screening area are different, and the pine nut specifications include a standard weight threshold; Performing pine nut feature recognition according to the pine nut image set to obtain an actual quantity set and calculate an actual weight set; Taking the standard weight threshold as a reference, performing consistency analysis on the actual weight set to obtain several screening consistency coefficients of the several screening areas; Continuously monitoring several screening consistency coefficient sequences within a preset time range, and performing mean value calculation, fluctuation analysis and trend fitting to obtain several consistency coefficient means, several fluctuation coefficients and several trend coefficients; If any consistency coefficient mean, fluctuation coefficient or trend coefficient does not satisfy a preset coefficient threshold, performing fault warning on a pine nut screening device; Taking the standard weight threshold as a reference, performing consistency analysis on the actual weight set, comprising: Randomly selecting a first standard weight threshold of a first screening area and obtaining a first actual weight; Taking the median of the first standard weight threshold as a first standard weight mean, calculating the deviation proportion of the first standard weight mean and the first actual weight, and setting the reciprocal of the deviation proportion as a first screening consistency coefficient and adding it to the several screening consistency coefficients; Continuously monitoring several screening consistency coefficient sequences within a preset time range, and performing mean value calculation, fluctuation analysis and trend fitting, comprising: Continuously performing image acquisition and consistency analysis within the preset time range to obtain several screening consistency coefficient sequences; Respectively performing mean value calculation on the several screening consistency coefficient sequences to obtain several consistency coefficient means; Respectively performing fluctuation analysis and trend fitting on the several screening consistency coefficient sequences to output several fluctuation coefficients and several trend coefficients; Respectively performing fluctuation analysis on the several screening consistency coefficient sequences, comprising: Respectively performing standard deviation calculation on the several screening consistency coefficient sequences to obtain several consistency coefficient standard deviations; According to the several consistency coefficient standard deviations and the several consistency coefficient means, calculating to obtain several fluctuation coefficients, wherein the fluctuation coefficient is the ratio of the consistency coefficient standard deviation to the consistency coefficient mean; Respectively performing trend fitting on the several screening consistency coefficient sequences, comprising: Building a two-dimensional coordinate system with a monitoring time node as the X-axis and a screening consistency coefficient as the Y-axis; In the two-dimensional coordinate system, performing distribution and curve fitting on the several screening consistency coefficient sequences to obtain several trend curves; Identifying the several trend curves to obtain several trend coefficients, wherein the trend coefficient is the overall rising or falling proportion.

2. The fault prediction method of a pine nut full-automatic flow processing equipment according to claim 1, characterized in that, Performing pine nut feature recognition according to the pine nut image set to obtain an actual quantity set, comprising: According to the historical screening records of the same type of pine nut screening device, collecting a sample pine nut image set and obtaining the actual pine nut quantity of different sample pine nut images as sample pine nut quantities to obtain a sample quantity set; The sample pine nut image is input, and the sample quantity is supervised. The sample pine nut image set and the sample quantity set are used to train and test the convolutional neural network until convergence, and a pine nut quantity recognition plug-in is obtained; The pine nut quantity recognition plug-in is used to perform pine nut feature recognition on each pine nut image in the pine nut image set, and an actual quantity set is output.

3. The fault prediction method of a pine nut full-automatic flow processing equipment according to claim 1, characterized in that, An actual weight set is calculated, including: The total weight of the pine nuts in each screening area is weighed to obtain a pine nut total weight set; The pine nut total weight set and the actual quantity set are mapped and calculated to output an actual weight set, wherein the actual weight is the average weight of a single pine nut.

4. The fault prediction method of a pine nut full-automatic flow processing equipment according to claim 1, characterized in that, The preset coefficient threshold includes a consistency coefficient threshold, a fluctuation coefficient threshold, and a trend decline ratio threshold, wherein the preset coefficient threshold is set based on the pine nut specification.

5. The method of claim 4, wherein the method further comprises: determining a failure of the pine nut full-automatic flow processing apparatus based on the at least one of the first data and the second data. If the average consistency coefficient of any screening area is less than the consistency coefficient threshold and / or the fluctuation coefficient is greater than the fluctuation coefficient threshold and / or the decline ratio of the trend coefficient is greater than the trend decline ratio threshold, a fault warning is given to the pine nut screening device.

6. A fault prediction system for a pine nut full-automatic flow processing apparatus, characterized by, The steps of the fault prediction method for the pine nut full-automatic flow processing equipment according to any one of claims 1-5, wherein the fault prediction system of the pine nut full-automatic flow processing equipment includes: An image acquisition module is configured to acquire pine nut images in each screening area through a camera to obtain a pine nut image set, wherein the pine nut specification of each screening area is different, and the pine nut specification includes a standard weight threshold; A weight calculation module is configured to perform pine nut feature recognition based on the pine nut image set to obtain an actual quantity set and calculate an actual weight set; A consistency analysis module is configured to analyze the consistency of the actual weight set based on the standard weight threshold to obtain a plurality of screening consistency coefficients for each screening area; A dynamic monitoring and analysis module is configured to continuously monitor a plurality of screening consistency coefficient sequences within a preset time range, and perform mean value calculation, fluctuation analysis, and trend fitting to obtain a plurality of average consistency coefficients, a plurality of fluctuation coefficients, and a plurality of trend coefficients; A threshold judgment module is configured to perform a fault warning on the pine nut screening device if any of the average consistency coefficient, the fluctuation coefficient, or the trend coefficient does not meet the preset coefficient threshold.

Images