A raw material conveying control system for candy production

Abstract

The present application relates to the raw material conveying control field, specifically to a kind of raw material conveying control system for confectionery production, the system includes data acquisition, prediction model establishment, prediction, control execution, image color analysis, feedback adjustment and so on module, prediction model establishment module is to acquisition data cleaning, normalization processing and trains LSTM model, constructs conveying prediction model;Prediction module exports optimal conveying parameter according to real-time data;Control execution module accurately controls conveying equipment according to predicted value;Image color analysis module acquires raw material image, judges whether raw material color is abnormal after multi-step processing;Feedback adjustment module processes abnormal raw material, records relevant information and stores to database;The present application can accurately control raw material conveying, guarantee product quality, and have abnormal raw material processing and data traceability function, effectively solve the problems existing in traditional raw material conveying control.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of raw material conveying control, in particular to a raw material conveying control system for candy production. BACKGROUND

[0002] With the vigorous development of the candy market, consumers' demand for the quality, variety and personalized customization of candies is increasing, which puts higher requirements on the raw material conveying link in the candy production process. However, the current raw material conveying technology for candy production has many shortcomings and cannot meet the needs of industry development.

[0003] The color of raw materials is one of the important indicators for judging the quality of raw materials. Abnormal color often indicates that the raw materials have quality problems such as deterioration and contamination. In the traditional raw material conveying process, there is a lack of real-time monitoring and analysis means for the color of raw materials, and it is impossible to discover and handle the abnormal color of raw materials in time. This makes raw materials with quality problems possibly enter the subsequent production link, affecting the quality and safety of candies and damaging the market reputation of enterprises. In the process of candy production, a large amount of historical data and real-time data related to raw material conveying are accumulated. However, the traditional conveying control system fails to fully exploit the value of these data. Due to the lack of effective data analysis and learning mechanism, it is impossible to establish a precise conveying prediction model and to intelligently control the raw material conveying process, so as to realize efficient and intelligent production management.

[0004] To solve the above problems, a technical solution is proposed. SUMMARY

[0005] In order to solve the technical problems proposed in the background, the present application provides a raw material conveying control system for candy production.

[0006] The purpose of the present application can be achieved by the following technical solutions: the present application is a raw material conveying control system for candy production, which comprises a data acquisition module, a prediction model establishment module, a prediction module, a control execution module, an image color analysis module, a feedback adjustment module and a database.

[0007] The data acquisition module is used to acquire the viscosity and density characteristic data of raw materials and the environmental temperature and air pressure condition data in the production workshop, and is connected with the order management program port of the enterprise to obtain the data of production orders. The specific process is as follows:

[0008] The viscometer and densimeter are installed in each raw material storage tank to acquire the viscosity and density values of raw materials in the raw material storage tank in real time. The environmental temperature sensors are installed in different areas of the production workshop, including the raw material storage area and the conveying pipeline, to acquire the environmental temperature values of raw materials in real time. The air pressure sensor is installed on the ceiling of the production workshop to acquire the air pressure value in the workshop in real time.

[0009] The conveying pipeline and the conveying equipment are provided with conveying parameter sensors to monitor the conveying speed of the raw materials in real time;

[0010] The port of the data acquisition module is connected to the port of the order management program of the enterprise, and the data acquisition frequency is set to be every hour to obtain the data of the production order, including product name, product specification, batch number and production quantity.

[0011] The prediction model establishment module cleans and normalizes the collected historical data and real-time data, and trains the model through the processed data, and the specific steps are as follows:

[0012] The preprocessing unit and the model training unit are set in the prediction model establishment module;

[0013] The preprocessing unit cleans the missing values and error values of the real-time data in the collection process, specifically: detecting missing data in real-time data and filling it through median, then sequentially obtaining abnormal values in real-time data, obtaining method is Z-score, deleting the obtained abnormal values, obtaining processed raw material data, environment data and order data;

[0014] The processed raw material data, environment data and order data are converted into standard normal distribution with mean value of 0 and standard deviation of 1, specifically: arranging the processed data into a data set, calculating the mean value of each original data point in the data set to obtain the original data mean value, then subtracting the original data mean value from each original data point to obtain the original data difference value, and calculating the standard deviation of the data set, dividing the original data difference value by the standard deviation of the data set to obtain the standard normal distribution data, thereby constructing the normalization process;

[0015] The normalized data is subjected to time series feature extraction, including seasonal characteristics and trend characteristics;

[0016] The raw material data, environment data and order data are decomposed into trend items, seasonal items and residual items using time series decomposition method, and the three items are used as seasonal feature points;

[0017] A linear regression model is used to fit the time series data, specifically: setting the time series as Yt, the time index as t, establishing a linear regression model Yt=β0+β1t+εt and obtaining the trend coefficient β1, wherein β0 represents the intercept and ε represents the error term;

[0018] The model training unit selects long short-term memory network (LSTM) as the deep learning model, including input layer, LSTM layer and output layer, specifically:

[0019] First, the dataset is divided into a training set and a temporary set with a ratio of 7:3. The temporary set is then divided into a validation set and a test set with a ratio of 2:1. The seasonal feature points and trend coefficients are sent as input vectors to the input layer. LSTM selects Tanh as the activation function and maps the input vector to the [-1, 1] interval. The output layer selects the mean square error as the loss function for training to construct a transportation prediction model.

[0020] The prediction module uses the trained delivery prediction model to output the optimal delivery parameter values ​​based on the real-time input of raw material characteristics, production order requirements, and environmental conditions. The specific steps are as follows:

[0021] The prediction module receives the viscosity value, density value, ambient temperature value, air pressure value and production demand, and loads the viscosity value, density value, ambient temperature value, air pressure value and production demand as input feature quantities into the delivery prediction model. The delivery value is calculated as the feature vector value of the output layer. Through continuous iterative calculation through forward propagation, the output layer obtains the optimal delivery value, and the optimal delivery value is sent to the control execution module;

[0022] The control execution module controls the operation of the conveying equipment according to the optimal conveying value predicted by the model, achieving precise control of the raw material conveying process. The specific steps are as follows:

[0023] Receive the optimal delivery value sent by the prediction module, extract the standard delivery value preset in the database, and make a difference between the optimal delivery value and the standard delivery value to obtain a delivery compensation value. If the delivery compensation value is a positive number, generate a control enhancement instruction and send it to the delivery pump control terminal, the valve adjustment terminal and the fresh air device. The delivery pump control terminal uses the standard delivery value as the initial delivery capacity, increases it according to the preset speed value of the motor, and increases it once per second. Then, the current transmission speed is continuously collected through the data acquisition module, and the preset speed value of the motor is adjusted until the delivery compensation value reaches 0. The valve adjustment terminal uses the current pressure value as the initial pressure point, and extracts the standard pressure value in the database, and reduces the initial pressure point until it reaches the standard pressure value. Similarly, the fresh air device uses the current temperature value as the initial temperature, extracts the standard temperature in the data, and performs cooling treatment on the pipeline until it reaches the standard temperature.

[0024] If the delivery compensation value is negative, the current speed is determined to be too high, and a control reduction instruction is generated and sent to the delivery pump control terminal, valve adjustment terminal and fresh air device. The standard delivery value is used as the initial delivery capacity, and the delivery speed is reduced according to the preset speed value, and so on, until the delivery compensation value reaches 0; the valve adjustment terminal uses the current pressure value as the initial pressure point and increases the pressure until it reaches the standard pressure value.

[0025] The fresh air device uses the current temperature value as the initial temperature and heats the pipe until it reaches the standard temperature.

[0026] The image color analysis module collects images of each raw material conveying process, identifies different raw materials, and obtains corresponding candidate color templates, and compares them with real-time raw material color images. The specific steps are as follows:

[0027] High-definition cameras are installed in each conveying pipeline area, and each conveying pipeline area is numbered. Real-time images of each initial raw material in each numbered pipeline are collected. First, each initial raw material image is converted into a grayscale image. The gradient amplitude and direction of each grayscale image are obtained by edge detection algorithm. The position of the local maximum of the gradient amplitude is taken as the boundary to obtain the raw material contour information of each initial raw material image. The pre-set raw material contour images in the database are extracted and matched in turn. The pre-set raw material contour images with a matching similarity exceeding a first similarity threshold are marked as raw materials to be detected. The image of the raw material to be detected is obtained. The gray values of the center pixel and the neighborhood pixels of the initial raw material image are compared by local binary pattern. The gray values of the neighborhood pixels are binary coded according to the size relationship with the center pixel to obtain a binary code. The pixels in the initial raw material image are counted and an initial histogram is generated. Similarly, the above steps are performed on the raw material to be detected to obtain a detection histogram. The sum of the ratio of the difference square of the corresponding interval of the initial histogram to the value of the corresponding interval of the detection histogram is calculated to obtain the chi-square distance.

[0028] The candidate color template of the raw material to be detected in the database is extracted. The sum of the squares of the differences between the color values of the candidate color template and the pixels in each local region of the initial raw material image is calculated. The size of the template T is m x n, and the size of the initial raw material image L is M x N. In the image L, the local region L x,y The square difference matching value SSD(x, y) of the template T is calculated. Where i and j are index variables for traversing each element in the template T.

[0029] By analogy, the template T is slid on the image L, and the value SSD(x, y) of each position (x, y) is calculated to obtain the square difference matching value of the image. The average matching value is calculated by averaging. The preset matching degree threshold is extracted. If the average matching value is greater than the preset matching degree threshold, it means that the color is abnormal. The raw material is marked as abnormal raw material, and the number of the position is marked. The adjustment instruction is generated and sent to the feedback adjustment module. If not, it means that the raw material color is consistent with the standard color.

[0030] The feedback adjustment module processes the marked raw material according to the received adjustment instruction. The specific process is as follows:

[0031] After receiving the adjustment instruction, the location number of the corresponding abnormal raw material is obtained, and then the control valve terminal with the location number is obtained, a blocking signal is generated and sent to the control valve. The control valve blocks the connection between the conveying pipeline where the abnormal raw material is located and the subsequent production links, and guides the abnormal raw material to a special isolation area or temporary storage container. The source of the abnormal raw material, the conveying pipeline number, the time of discovery and the type of abnormality are recorded. The abnormality type includes information such as color that is too dark, too light and has mixed colors, and the information is sent to the database for storage to facilitate subsequent tracing and analysis.

[0032] Compared with the existing technology, the present invention has the following advantages: the image color analysis module collects raw material images by installing a high-definition camera in the conveying pipeline area, identifies the raw materials through grayscale conversion, edge detection, contour matching and other operations, and determines whether the raw material color is abnormal through local binary pattern and square difference matching calculations. Once an abnormality is found, it is promptly marked and an adjustment instruction is sent to the feedback adjustment module. This function can effectively prevent raw materials with abnormal colors from entering subsequent production links, ensuring the color and quality of candy products and improving the market competitiveness of products.

[0033] The control execution module compares the optimal delivery value from the prediction module with the standard delivery values ​​preset in the database and calculates a delivery compensation value to precisely control the operation of the conveying equipment. When the delivery compensation value is positive, the delivery pump, valves, and fresh air device work together to increase delivery capacity and adjust pressure and temperature. When it is negative, the adjustment is reversed. This process achieves precise control of the raw material delivery speed, pressure, and temperature, ensuring that the raw materials are delivered under optimal conditions, improving the accuracy of raw material delivery, and ensuring the stable quality of candy production. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. The following drawings are not intentionally scaled to the actual size, and the focus is on illustrating the main purpose of the present invention.

[0035] Figure 1 Schematic diagram of module connection of the present invention. DETAILED DESCRIPTION

[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts also fall within the scope of protection of the present invention.

[0037] Please refer to Figure 1As shown, the present application is a raw material conveying control system for candy production, which comprises a data acquisition module, a prediction model establishment module, a prediction module, a control execution module, an image color analysis module, a feedback adjustment module and a database.

[0038] The data acquisition module is used to acquire the viscosity and density characteristic data of the raw materials and the environmental temperature and air pressure condition data in the production workshop in real time, and is connected with the order management program port of the enterprise to obtain the production order data, and the specific process is as follows:

[0039] The viscometer and the densimeter are installed in each raw material storage tank to acquire the viscosity value and the density value of the raw materials in the raw material storage tank in real time; the environmental temperature sensors are installed in different areas in the production workshop, including the raw material storage area and the conveying pipeline, to acquire the environmental temperature value of the raw materials in real time; the air pressure sensor is installed on the ceiling of the production workshop to acquire the air pressure value in the workshop in real time;

[0040] The conveying parameter sensors are installed in the conveying pipeline and the conveying equipment to monitor the conveying speed of the raw materials in real time;

[0041] The port of the data acquisition module is connected with the order management program port of the enterprise, the data acquisition frequency is set to be every hour, and the production order data is obtained, including the product name, the product specification, the batch number and the production quantity.

[0042] The prediction model establishment module cleans and normalizes the collected historical data and real-time data, and trains the model through the processed data, and the specific steps are as follows:

[0043] The prediction model establishment module cleans and normalizes the collected historical data and real-time data, and trains the model through the processed data, and the specific steps are as follows:

[0044] The preprocessing unit cleans the missing values and error values of the real-time data in the acquisition process, specifically: detecting the missing data in the real-time data and filling it through the median, then sequentially obtaining the abnormal values in the real-time data, obtaining the abnormal values by Z-score, deleting the obtained abnormal values, and obtaining the processed raw material data, environmental data and order data;

[0045] The processed raw material data, environmental data and order data are converted into standard normal distribution with mean value of 0 and standard deviation of 1, specifically: arranging the processed data into a data set, calculating the mean value of each original data point in the data set to obtain the original data mean value, then subtracting the original data mean value from each original data point to obtain the original data difference value, and calculating the standard deviation of the data set, dividing the original data difference value by the standard deviation of the data set to obtain the standard normal distribution data, thereby constructing the normalization process;

[0046] The normalized data is subjected to time series feature extraction, including seasonal features and trend features. It should be noted that the seasonal features refer to regular changes in raw material production and transportation with seasonal changes. For example, low temperatures in winter or high temperatures in summer cause fluctuations in the viscosity and density of raw materials. Seasonal fluctuations in order quantity and product type are periodically associated with raw material consumption. Trend features reflect the sustained upward, downward, or stable change trends in raw material transportation over a longer time span.

[0047] The raw material data, environmental data, and order data are decomposed into trend items, seasonal items, and residual items using time series decomposition. These three items are used as seasonal feature points.

[0048] A linear regression model is used to fit the time series data. Specifically, let the time series be Yt, and the time index be t. A linear regression model Yt = β0 + β1t + εt is established, and the trend coefficient β1 is obtained, where β0 represents the intercept, and ε represents the error term.

[0049] The model training unit selects a long short-term memory network (LSTM) as the deep learning model, including an input layer, an LSTM layer, and an output layer. Specifically:

[0050] First, the data set is divided into a training set and a temporary set, with a division ratio of 7:3. Then, the temporary set is divided into a validation set and a test set, with a division ratio of 2:1. The seasonal feature points and trend coefficients are sent to the input layer as input vectors. The LSTM selects Tanh as the activation function to map the input vectors to the [-1, 1] interval. The output layer selects the mean square error as the loss function for training. In this way, a transportation prediction model is constructed.

[0051] The prediction module uses the trained transportation prediction model to output the optimal transportation parameter value based on real-time input of raw material properties, production order demand, and environmental conditions. The specific steps are as follows:

[0052] The prediction module receives the viscosity value, density value, environmental temperature value, air pressure value, and production demand. The viscosity value, density value, environmental temperature value, air pressure value, and production demand are loaded as input feature quantities into the transportation prediction model. The transportation value is calculated as the feature vector value of the output layer. Through continuous iterative calculation of forward propagation, the optimal transportation value is obtained for the output layer. The optimal transportation value is sent to the control execution module.

[0053] The control execution module controls the operation of the transportation equipment based on the optimal transportation value predicted by the model, achieving precise control of the raw material transportation process. The specific steps are as follows:

[0054] The optimal conveying value sent by the receiving prediction module is subtracted from the preset standard conveying value in the database to obtain a conveying compensation value. If the conveying compensation value is positive, a control enhancement instruction is generated and sent to the conveying pump control terminal, the valve adjustment terminal and the fresh air device. The conveying pump control terminal takes the standard conveying value as the initial conveying capacity and increases the motor preset speed value every second. The current conveying speed is continuously collected by the data collection module. The motor preset speed value is adjusted until the conveying compensation value is 0. The valve adjustment terminal takes the current pressure value as the initial pressure point and extracts the standard pressure value in the database. The initial pressure point is reduced until the standard pressure value is reached. Similarly, the fresh air device takes the current temperature value as the initial temperature and extracts the standard temperature in the data. The pipeline is cooled until the standard temperature is reached.

[0055] If the conveying compensation value is negative, it is determined that the current speed is too high. A control reduction instruction is generated and sent to the conveying pump control terminal, the valve adjustment terminal and the fresh air device. The standard conveying value is taken as the initial conveying capacity. The conveying speed is reduced according to the preset speed value. In turn, until the conveying compensation value is 0. The valve adjustment terminal takes the current pressure value as the initial pressure point. The pressure is enhanced until the standard pressure value is reached.

[0056] The fresh air device takes the current temperature value as the initial temperature. The pipeline is warmed until the standard temperature is reached.

[0057] The image color analysis module collects images of the conveying process of each raw material, identifies different raw materials and obtains corresponding candidate color templates. The real-time raw material color image is compared. The specific steps are as follows:

[0058] High-definition cameras are installed in each conveying pipeline area, and each conveying pipeline area is numbered, and each initial raw material image of each numbered pipeline is collected in real time. First, each initial raw material image is converted into a gray image, the corresponding raw material gradient amplitude and direction of each gray image are obtained through an edge detection algorithm, the position of the local maximum of the gradient amplitude is taken as the boundary, the raw material contour information of each initial raw material image is obtained, and each preset raw material contour image in the database is sequentially matched. The preset raw material contour image with a matching similarity greater than a first similarity threshold is marked as a raw material to be detected. It is to be noted that the chi-square distance is used to measure the difference between two histograms, and the smaller the chi-square distance, the more similar the two histograms, and the more similar the image textures. A second chi-square distance threshold is extracted, and if the chi-square distance is less than the second chi-square distance threshold, it is determined that the raw material in the raw material image to be detected is matched to the corresponding initial raw material image.

[0059] A candidate color template of the raw material image to be detected in the database is extracted. It is to be noted that the candidate color template is an image with uniform raw material color and typical characteristics selected from a large number of collected images. The sum of squares of differences between the color values of the candidate color template and the corresponding pixels in each local region of the initial raw material image is calculated. The size of the template T is m x n, the size of the initial raw material image L is M x N, and m, n, M and N respectively represent rows and columns. The local region L x,y The sum of squares difference matching value SSD(x, y) of the template T and the image L is calculated, and the formula is as follows: Where i and j are index variables for traversing each element in the template T.

[0060] By analogy, the template T is slid on the initial raw material image L, the value SSD(x, y) of each position (x, y) is calculated, the sum of squares difference matching values of the image are obtained, and the average matching value is calculated by mean value calculation. A preset matching degree threshold is extracted, and if the average matching value is greater than the preset matching degree threshold, it is determined that the color is abnormal, the raw material is marked as an abnormal raw material, and the number of the position is marked. An adjustment instruction is generated and sent to the feedback adjustment module. If not, it is determined that the raw material color is consistent with the standard color.

[0061] The feedback adjustment module processes the marked raw material according to the received adjustment instruction, and the specific process is as follows:

[0062] Upon receiving the adjustment instruction, the position number of the corresponding abnormal raw material is obtained, the control valve terminal of the position number is obtained, a blocking signal is generated and sent to the control valve, the control valve blocks the connection between the conveying pipeline of the abnormal raw material and the subsequent production link, the abnormal raw material is guided to a special isolation area or a temporary storage container, the source of the abnormal raw material, the conveying pipeline number, the discovery time and the abnormal type are recorded, the abnormal type includes information such as too deep, too shallow and appearance of mixed color, and is sent to a database for storage, so as to facilitate subsequent tracing and analysis.

[0063] The above is a description of the present application and should not be considered as a limitation. Although several exemplary embodiments of the present application are described, those skilled in the art will readily understand that many modifications can be made to the exemplary embodiments without departing from the novel teachings and advantages of the present application. Therefore, all such modifications are intended to be included within the scope of the present application as defined by the claims. It should be understood that the above is a description of the present application and should not be considered as a limitation. The disclosed specific embodiments are not intended to be limiting, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present application is defined by the claims and their equivalents.

Claims

1. A raw material conveying control system for candy production, comprising a prediction module, a control execution module, a data acquisition module, an image color analysis module, a feedback adjustment module, and a database, characterized in that: The prediction module uses the trained delivery prediction model to output the optimal delivery parameter values ​​based on the real-time input of raw material characteristics, production order requirements and environmental conditions. Specifically, it receives viscosity values, density values, ambient temperature values, air pressure values ​​and output requirements, loads the viscosity values, density values, ambient temperature values, air pressure values ​​and output requirements as input feature quantities into the delivery prediction model, calculates the delivery value as the feature vector value of the output layer, and continuously iterates the calculation through forward propagation until the output layer obtains the optimal delivery value, which is then sent to the control execution module. The image color analysis module collects images of the raw material conveying process, identifies different raw materials, obtains corresponding candidate color templates, and compares them with the real-time raw material color images; converts each initial raw material image into a grayscale image, obtains the corresponding raw material gradient amplitude and direction of each grayscale image through an edge detection algorithm, uses the position of the local maximum gradient amplitude as the boundary, obtains the raw material contour information of each initial raw material image, performs overlap matching in sequence, and marks the preset raw material contour image with a matching similarity exceeding a first similarity threshold as the raw material to be detected; obtains the image of the raw material to be detected, compares the grayscale values ​​of the central pixel and the neighboring pixels of the initial raw material image through a local binary mode, binarizes the grayscale values ​​of the neighboring pixels according to their size relationship with the central pixel to obtain a binary code, counts the pixels in the initial raw material image and generates an initial histogram, processes the raw material image to obtain a histogram to be detected, calculates the sum of the ratios of the squares of the difference values ​​of the corresponding intervals of the initial histogram to the values ​​of the corresponding intervals of the histogram to be detected, and obtains the chi-square distance; if the chi-square distance is less than a second chi-square distance threshold, it is determined that the raw material in the raw material image to be detected matches the corresponding initial raw material image; Calculate the sum of the squares of the differences between the color values ​​of the corresponding pixels in each local area of ​​the candidate color template and the original raw material image, and match the squared difference between the local area with the upper left corner of the original raw material image and the template; slide the template on the original raw material image, calculate each position, obtain the squared difference matching value of the image, and calculate the average to obtain the average matching value. If the average matching value is greater than the preset matching threshold, the raw material is marked as abnormal raw material, and an adjustment instruction is generated and sent to the feedback adjustment module; After receiving the adjustment instruction, the feedback adjustment module processes the marked raw materials, specifically: obtains the location number of the corresponding abnormal raw material, then obtains the control valve terminal with the number, generates a blocking signal and sends it to the control valve, and the control valve blocks the connection between the conveying pipeline where the abnormal raw material is located and the subsequent production links, and guides the abnormal raw materials to a special isolation area or temporary storage container, records the source of the abnormal raw material, the conveying pipeline number, the time of discovery and the type of abnormality, which includes information data such as too dark, too light and the appearance of mixed colors, and sends it to the database for storage.

2. A raw material conveying control system for candy production according to claim 1, characterized in that: The system also includes a prediction model building module, which cleans and normalizes the collected historical data and real-time data and trains the model using the processed data. The specific steps are as follows: The prediction model establishment model is provided with a pre-processing unit and a model training unit; The preprocessing unit cleans missing and erroneous values ​​in the real-time data during the collection process. Specifically, it detects missing data in the real-time data and fills it with the median. It then sequentially obtains outliers in the real-time data using the Z-score method and deletes the obtained outliers to obtain processed raw material data, environmental data, and order data. The processed raw material data, environmental data, and order data are converted into a standard normal distribution with a mean of 0 and a standard deviation of 1. Specifically, the processed data are arranged into a data set, the values ​​of each original data point in the data set are averaged to obtain the original data mean, the value of each original data point is subtracted from the original data mean to obtain the original data difference, and the standard deviation of the data set is calculated. The original data difference is divided by the standard deviation of the data set to obtain the standard normal distribution data, thereby constructing a normalization process; Extract time series features from the normalized data. Time series features include seasonal features and trend features. The time series decomposition method is used to decompose the raw material data, environmental data and order data into trend items, seasonal items and residual items, and these three items are used as seasonal characteristic points; Use the linear regression model to fit the time series data. Specifically, let the time series be Yt, the time index be t, and establish a linear regression model. And get the trend coefficient ,in is represented as the intercept, is represented as an error term.

3. A raw material conveying control system for candy production according to claim 2, characterized in that: The model training unit selects a long short-term memory network LSTM as a deep learning model, including an input layer, an LSTM layer, and an output layer, specifically: First, the dataset is divided into a training set and a temporary set with a ratio of 7:

3. The temporary set is then divided into a validation set and a test set with a ratio of 2:

1. The seasonal feature points and trend coefficients are sent as input vectors to the input layer. LSTM selects Tanh as the activation function and maps the input vector to the [-1, 1] interval. The output layer selects the mean square error as the loss function for training to construct a transportation prediction model.

4. A raw material conveying control system for candy production according to claim 1, characterized in that: The control execution module controls the operation of the conveying equipment according to the optimal conveying value predicted by the model to achieve precise control of the raw material conveying process. The specific steps are as follows: Receive the optimal delivery value sent by the prediction module, extract the standard delivery value preset in the database, and make a difference between the optimal delivery value and the standard delivery value to obtain a delivery compensation value. If the delivery compensation value is a positive number, generate a control enhancement instruction and send it to the delivery pump control terminal, the valve adjustment terminal and the fresh air device. The delivery pump control terminal uses the standard delivery value as the initial delivery capacity, increases it according to the preset speed value of the motor, and increases it once per second. Then, the current transmission speed is continuously collected through the data acquisition module, and the preset speed value of the motor is adjusted until the delivery compensation value reaches 0. The valve adjustment terminal uses the current pressure value as the initial pressure point, and extracts the standard pressure value in the database, and reduces the initial pressure point until it reaches the standard pressure value. Similarly, the fresh air device uses the current temperature value as the initial temperature, extracts the standard temperature in the data, and performs cooling treatment on the pipeline until it reaches the standard temperature. If the delivery compensation value is negative, it is determined that the current speed is too high, and a control reduction instruction is generated and sent to the delivery pump control terminal, the valve adjustment terminal and the fresh air device. The standard delivery value is used as the initial delivery capacity, and the delivery speed is reduced according to the preset speed value, and so on, until the delivery compensation value reaches 0; the valve adjustment terminal uses the current pressure value as the initial pressure point, and increases the pressure until it reaches the standard pressure value. The fresh air device uses the current temperature value as the initial temperature and heats the pipeline until it reaches the standard temperature.

5. A raw material conveying control system for candy production according to claim 4, characterized in that: The data acquisition module is used to collect the viscosity and density characteristic data of the raw materials, as well as the ambient temperature and air pressure conditions in the production workshop in real time, and connects with the enterprise's order management program port to obtain production order data. The specific process is as follows: Install viscometers and densitometers in each raw material storage tank to collect the viscosity and density values ​​of the raw materials in the tank in real time; install various ambient temperature sensors in different areas of the production workshop, including the raw material storage area and the conveying pipeline, to collect the ambient temperature values ​​of the raw materials in real time; install an air pressure sensor on the ceiling of the production workshop to collect the air pressure value in the workshop in real time; Install conveying parameter sensors on conveying pipelines and conveying equipment to monitor the conveying speed of raw materials in real time; The port of the data acquisition module is connected to the port of the enterprise's order management program, and the data acquisition frequency is set to every hour to obtain the data of the production order, which includes product name, product specification, batch number and production quantity.

Images