Control management system and method for pavement aggregate and asphalt

Abstract

The invention discloses a pavement aggregate and asphalt control management system and method, and belongs to the technical field of production control, and the method comprises the steps: carrying out the correlation analysis of production process parameters, and obtaining a conversion function; correlation analysis is conducted on the production process parameters and the product quality parameters, the influence weight of each production process parameter is obtained, and the production process parameters are divided into key parameters and non-key parameters based on the influence weights; sampling key parameters and non-key parameters to obtain first time sequence data and second time sequence data, and supplementing missing values based on a transfer function to obtain complete data; and if each production process parameter in the complete data is within the corresponding standard numerical range, processing the complete data based on the first prediction model and outputting a product quality parameter, otherwise, processing the complete data based on the second prediction model and outputting the product quality parameter. According to the invention, the production management system quickly processes a large amount of equipment data, thereby ensuring the production quality.

Description

Technical Field

[0001] The invention belongs to the technical field of production management, and in particular relates to a control management system and method for road aggregate and asphalt. Background Art

[0002] In the traditional asphalt production process, the quality of raw materials is greatly affected by many factors such as production equipment, production process and production environment. Due to the lack of real-time monitoring of production parameters, quality management usually relies on manual tests after production is completed to evaluate. This method has the problems of many types of parameters and complex processes. Once an abnormality is found during the test, the relevant batch of raw materials has often been mass-produced and needs to be reworked, resulting in economic losses.

[0003] With the advancement of production technology, some asphalt production companies have begun to introduce automated systems to monitor multiple parameters of production equipment and environment in real time, and combine these parameters with production processes for instant calculation. At the same time, the quality management system is gradually linked with the production system. When the test and calculation results show that the parameters do not meet the process requirements, the system can automatically feedback and adjust the equipment operation status to ensure that the production parameters are always within the compliance range.

[0004] Although the application of automation and real-time monitoring technologies has improved the management efficiency and quality control of asphalt production, the production data management system may face efficiency issues when processing large quantities of data, which will affect the further improvement of production quality. Summary of the invention

[0005] In order to solve the above problems, the present invention provides a control management system and method for road aggregate and asphalt to solve the problems in the prior art.

[0006] In order to achieve the above-mentioned purpose of the invention, the present invention proposes a control and management method for road aggregate and asphalt, comprising:

[0007] Collecting production process parameters and product quality parameters in historical production processes, performing correlation analysis on the production process parameters, and obtaining conversion functions for converting the production process parameters to each other;

[0008] Performing a correlation analysis on the production process parameters and the product quality parameters to obtain the influence weight of each production process parameter on the product quality parameter, and dividing the production process parameters into key parameters and non-key parameters based on the influence weight;

[0009] Establishing a first prediction model and a second prediction model based on the key parameters, the non-key parameters and the product quality parameters;

[0010] In the actual production process, the key parameters and the non-key parameters are sampled based on the first frequency and the second frequency respectively to obtain a plurality of first time series data and a second time series data;

[0011] Time-aligning the first time series data and the second time series data to obtain aligned data, and supplementing the missing values ​​of the aligned data based on the conversion function to obtain complete data;

[0012] If each of the production process parameters in the complete data is within the corresponding standard numerical range, the complete data is processed based on the first prediction model and the product quality parameters are output; otherwise, the complete data is processed based on the second prediction model and the product quality parameters are output.

[0013] Furthermore, the first prediction model is a linear regression model, which includes a long-term prediction model and a medium-term prediction model, and the second prediction model is a neural network model, and a first time interval, a second time interval and a third time interval are set, the second time interval is smaller than the first time interval, and the third time interval is smaller than the second time interval;

[0014] The long-term prediction model, the medium-term prediction model and the second prediction model respectively predict the product quality parameters in the first time period, the second time period and the third time period in the future based on the production process parameters collected corresponding to the first time interval, the second time interval and the third time interval.

[0015] Further, the first frequency and the second frequency are adjusted based on the following steps:

[0016] A monitoring interval is set, and the monitoring interval includes multiple sub-time intervals. Whenever the length of the sub-time interval is reached, the intermediate quality parameters of the semi-finished products output by each production equipment are obtained. After the number of the sub-time intervals reaches a predetermined number, the equipment state of the corresponding production equipment is divided into a stable state, an unstable state and a deteriorated state based on all the intermediate product quality parameters collected in the monitoring interval. If the equipment state of the production equipment changes from the stable state to the unstable state or the deteriorated state in an adjacent monitoring interval, the first frequency and the second frequency are increased, and the lengths of the first time interval, the second time interval and the third time interval are shortened.

[0017] Furthermore, dividing the equipment status of the production equipment includes the following steps:

[0018] Preset multiple data features and set a collection time, wherein the collection time includes multiple sub-time intervals, and when the collection time is reached, calculate the data features of the intermediate quality parameter in each collection time to form multiple data sets corresponding to the data features, and calculate the fluctuation value of the intermediate quality parameter in the monitoring interval based on the data sets;

[0019] The process capability is calculated based on the optimal quality value, the lowest quality value, the average quality value and the standard deviation of the quality value of the intermediate quality parameter in the monitoring interval; if the fluctuation value is less than the first threshold, the process capability is greater than the second threshold, and the production equipment is classified as the stable state; if the fluctuation value is greater than or equal to the first threshold, the production equipment is classified as the unstable state; if the fluctuation value is less than the first threshold, the process capability is less than or equal to the second threshold, and the production equipment is classified as the deteriorated state.

[0020] Furthermore, after the equipment status of the monitoring interval is generated, the corresponding intermediate quality parameters are retained and accumulated as previous data, and after completing the equipment status evaluation of the subsequent monitoring interval, the previous data are added to the intermediate quality parameters of the subsequent monitoring interval to obtain the overall fluctuation value and the process capability in multiple consecutive monitoring intervals.

[0021] Further, dividing the key parameters and the non-key parameters comprises the following steps:

[0022] Obtain the coefficient corresponding to each of the production process parameters in the first prediction model, calculate the impact weight of each of the production process parameters based on the proportion of the coefficient in the sum of all the coefficients, set the production process parameters whose impact weight is greater than a critical threshold as the key parameters, and set the production process parameters whose impact weight is less than or equal to the critical threshold as the non-key parameters.

[0023] Furthermore, the second prediction model is an LSTM model.

[0024] Furthermore, the data characteristics include median, kurtosis, skewness, and standard deviation.

[0025] The present invention also provides a control and management system for road aggregate and asphalt, which is used to implement the control and management method for road aggregate and asphalt described above, and the system includes:

[0026] An analysis module, used to collect production process parameters and product quality parameters in historical production processes, perform correlation analysis on the production process parameters, obtain conversion functions for mutually converting the production process parameters, perform correlation analysis on the production process parameters and the product quality parameters, obtain the influence weight of each of the production process parameters on the product quality parameters, and divide the production process parameters into key parameters and non-key parameters based on the influence weight;

[0027] A collection module, used for sampling the key parameters and the non-key parameters based on the first frequency and the second frequency respectively in an actual production process to obtain a plurality of first time series data and a second time series data;

[0028] A preprocessing module, used for performing time alignment on the first time series data and the second time series data to obtain aligned data, and supplementing missing values ​​on the aligned data based on the conversion function to obtain complete data;

[0029] A judgment module is used to establish a first prediction model and a second prediction model based on the key parameters, the non-key parameters and the product quality parameters. If each of the production process parameters in the complete data is within the corresponding standard numerical range, the complete data is processed based on the first prediction model and the product quality parameters are output; otherwise, the complete data is processed based on the second prediction model and the product quality parameters are output.

[0030] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0031] The present invention ensures the integrity and accuracy of production records by combining multiple sensors to collect production process parameters and product quality parameters. In actual production, the use of conversion functions to fill in missing data not only improves the continuity of data, but also reduces quality problems that may be caused by equipment failure or data loss. By introducing a combined prediction method of a linear regression model and a neural network model, the system can quickly process large quantities of equipment data through a linear model, thereby improving data processing efficiency. Within the normal range of production parameters, a linear regression model is used for rapid prediction, saving computing resources; when the parameters exceed the normal range, a neural network model is used for more accurate prediction. The frequency-divided sampling design of key parameters and non-key parameters ensures high-frequency capture of key data while reducing the overall burden of the system. In summary, the present invention constructs a set of efficient, flexible and intelligent production process management system. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A flow chart of the steps of the control and management method of pavement aggregate and asphalt of the present invention;

[0033] Figure 2 It is a schematic diagram of the structure of the control and management system of pavement aggregate and asphalt of the present invention;

[0034] Figure 3 This is a schematic diagram of the system interface of the asphalt management system of the present invention;

[0035] Figure 4 It is a schematic diagram of the system interface of the pavement aggregate system of the present invention. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0037] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0038] like Figure 1 As shown, a control and management method for road aggregate and asphalt includes:

[0039] S1: Collect production process parameters and product quality parameters in the historical production process, perform correlation analysis on the production process parameters, and obtain conversion functions for converting production process parameters to each other.

[0040] The production process parameters include but are not limited to the setting parameters of the production equipment. The production equipment includes belt conveyors, motors of key equipment, crushers and various electrical equipment. The production process parameters include pressure, crushed particle size, mixing time, current, etc., as well as the environmental parameters of the equipment during the production process, such as ambient temperature and humidity. Product quality parameters include the hardness, strength, density, etc. of asphalt or other finished products. The production process parameters are obtained by arranging current acquisition sensors, voltage acquisition sensors, pressure sensing sensors, temperature and humidity acquisition sensors, high-definition ball cameras, etc. in the production equipment, and the collected digital signals are transmitted to the edge computing device through the standard TTL level interface. In addition, the real-time power of the production equipment can be monitored and calculated through current and voltage.

[0041] The correlation analysis of the production process parameters is performed to obtain the conversion function. The conversion function converts at least one type of production process parameter into another type by analyzing the correlation between different parameters in the production process, so as to supplement the missing data. For example, in the high-temperature heating link, some sensors may cause the missing temperature data due to failure. By establishing the conversion function of pressure, time and temperature, at the time point when the temperature data is missing, the temperature is estimated through the conversion function based on the collected pressure and time data, and supplemented into the production record, so as to ensure the integrity of the production parameters.

[0042] S2: Conduct correlation analysis on production process parameters and product quality parameters, obtain the influence weight of each production process parameter on product quality parameters, and divide the production process parameters into key parameters and non-key parameters based on the influence weight.

[0043] Specifically, a linear correlation analysis is performed on the production process parameters and the product quality parameters to obtain a linear regression equation for calculating various product quality parameters. The corresponding influence weight is calculated according to the coefficient of the production process parameter in the linear regression equation. The specific calculation method is introduced later. The influence weight reflects the degree of influence of each production process parameter on the quality of the final product. Based on the influence weight, the production process parameters are divided into key parameters and non-key parameters. Key parameters are parameters that have a significant impact on product quality, while non-key parameters have a relatively small impact. This embodiment is provided with a critical threshold, which is divided according to the size relationship between the influence weight and the critical threshold.

[0044] S3: Establish a first prediction model and a second prediction model based on key parameters, non-key parameters and product quality parameters.

[0045] The first prediction model is a linear regression model. Due to its small amount of calculation, linear regression requires less time and resources for training, and the training speed is very fast. When all production process parameters are within the normal range, the first prediction model is used for prediction, which can reduce the amount of calculation of the system and quickly obtain the calculation results. The second prediction model is a neural network model, which is used to process complex nonlinear relationships. When there are production process parameters that exceed the standard, the neural network model is used for more accurate prediction, so that accurate prediction results can be obtained.

[0046] S4: In the actual production process, the key parameters and the non-key parameters are sampled based on the first frequency and the second frequency respectively to obtain a plurality of first time series data and a second time series data.

[0047] The first frequency is greater than the second frequency. The key parameters are collected at the first frequency, and the sampling frequency is higher, so as to ensure the acquisition of the real-time change trend of the key parameters in the production process; the non-key parameters are collected at the second frequency, and the sampling frequency is lower, so as to reduce the occupation of system resources. Through the above sampling, the first time series data corresponding to multiple key parameters and the second time series data corresponding to multiple non-key parameters can be obtained.

[0048] S5: Time-align the first time series data and the second time series data to obtain aligned data, and supplement the missing values ​​of the aligned data based on the conversion function to obtain complete data.

[0049] The first time series data and the second time series data are time-aligned to ensure that the sampling data of different types of parameters have the same time reference, which is convenient for subsequent input into the model for calculation; then, the missing data in the aligned data is identified, and the value of the missing data is inferred based on the previously calculated conversion function, so as to supplement the missing data, and the data obtained after supplementation is defined as complete data.

[0050] S6: If each production process parameter in the complete data is within the corresponding standard value range, the complete data is processed based on the first prediction model and the product quality parameters are output; otherwise, the complete data is processed based on the second prediction model and the product quality parameters are output.

[0051] The standard numerical range can be determined based on experience or the product manual. For example, for a certain asphalt raw material, the heating temperature is between 160°C and 180°C. After the production process parameters are collected, they are first sent to the edge computing device. The edge computing device determines whether the collected real-time production data are all within the standard numerical range. If they are all within the standard numerical range, the first prediction model is used to predict the product quality parameters, so as to quickly obtain the results. If there are parameters that are not within the standard numerical range, the production process parameters will continue to be sent to the server and calculated using the second prediction model. The edge computing module has a certain computing power, and the edge computing module is closer to the on-site production equipment, so the prediction results can be quickly obtained through the edge computing module.

[0052] The present invention combines a variety of sensors to collect production process parameters and product quality parameters, and uses an edge computing module to process and supplement missing data in real time, thereby ensuring the integrity and accuracy of production records. In actual production, the use of conversion functions to fill in missing data not only improves the continuity of data, but also reduces quality problems that may be caused by equipment failure or data loss. By introducing a combined prediction method of a linear regression model and a neural network model, the system can quickly process large quantities of equipment data through a linear model, thereby improving data processing efficiency. Within the normal range of production parameters, a linear regression model is used for rapid prediction, saving computing resources; when the parameters exceed the normal range, a neural network model is used for more accurate prediction. The frequency-divided sampling design of key parameters and non-key parameters ensures high-frequency capture of key data while reducing the overall burden of the system. In summary, the present invention constructs a set of efficient, flexible and intelligent production process management system.

[0053] In this embodiment, the first prediction model is a linear regression model, which includes a long-term prediction model and a medium-term prediction model. The second prediction model is a neural network model. A first time interval, a second time interval and a third time interval are set. The second time interval is smaller than the first time interval, and the third time interval is smaller than the second time interval.

[0054] The long-term prediction model, the medium-term prediction model and the second prediction model predict the product quality parameters in the first time period, the second time period and the third time period in the future based on the production process parameters collected corresponding to the first time period, the second time period and the third time period respectively.

[0055] The first time interval, the second time interval and the third time interval can be set to 24 hours, 6 hours and 30 minutes. In particular, in this embodiment, the first time interval is equal to the first time period, the second time interval is equal to the second time period, and the third time interval is equal to the third time period. For the first time interval and the second time interval, the time length span is relatively long, and more production data will be obtained. It is suitable to use linear regression models (long-term prediction models and medium-term prediction models) for rapid prediction and obtain long-term quality change trends. For example, the linear regression model is based on the production records of the previous 24 hours, analyzes the average effects of pressure and temperature on hardness, and predicts the hardness range of the product in the next 24 hours. The predicted hardness value is 80±4MPa. The following linear regression model can be established, y=β1x1+β2x2+…+β n x n+ε, where y is a product quality parameter, β1~βn are coefficients, x1~xn are input production process parameters, and ε is a bias term. In particular, one type of production process parameter can correspond to multiple coefficients, such as β1~βk corresponds to the production process parameter temperature, βk+1~βn corresponds to the production process parameter pressure, x1~xk are temperature values ​​measured at consecutive time points, and xk+1~xn are pressure values ​​measured at consecutive time points.

[0056] When there are production process parameters that exceed the standard value range, a more accurate prediction of product quality in the short term is required. The second prediction model uses complex multi-layer neural networks, such as LSTM networks, to achieve accurate predictions. For example, in actual production, the temperature of the equipment rises to 180°C due to a fault. The second prediction model predicts that the hardness in the next 30 minutes may exceed the safe range based on the production process parameters for 30 minutes. In this way, the real-time impact of parameter fluctuations on the current product quality is determined.

[0057] In this embodiment, the first frequency and the second frequency are further adjusted based on the following steps:

[0058] A monitoring interval is set, and the monitoring interval includes multiple sub-time intervals. Whenever the length of a sub-time interval is reached, the intermediate quality parameters of the semi-finished products output by each production equipment are obtained. After the number of sub-time intervals reaches a predetermined number, the equipment state of the corresponding production equipment is divided into a stable state, an unstable state and a deteriorated state based on all the intermediate product quality parameters collected in the monitoring interval. If the equipment state of the production equipment changes from a stable state to an unstable state or a deteriorated state in an adjacent monitoring interval, the first frequency and the second frequency are increased, and the duration of the first time interval, the second time interval and the third time interval are shortened.

[0059] When the production equipment is in an unstable state, it is more likely to fail, thereby causing abnormal parameters. Therefore, the present invention also monitors the state of the production equipment. When the equipment is in an unstable state, the frequency of collecting key parameters and non-key parameters is increased, so that abnormal parameters are detected in time. Specifically, a monitoring interval is first set, for example, 3 days, and the sub-time interval is 2 hours. After every 2 hours, the intermediate quality parameters of the semi-finished products produced by the production equipment in the past 2 hours are obtained, and the semi-finished products are used to continue production in the next link. Then, according to all the intermediate quality parameters collected within 3 days, the corresponding production equipment is divided into a stable state, an unstable state, and a deteriorated state. If the production equipment in the current monitoring interval is in a stable state, and the production equipment in the current monitoring interval is in an unstable state, the corresponding first frequency and the second frequency are increased. At the same time, in order to ensure that the collected data matches the input dimensions of the first prediction model and the second prediction model, the duration of the first time interval, the second time interval, and the third time interval is appropriately reduced according to the increased frequency. For example, the first frequency is adjusted from 60 times / hour to 120 times / hour, and the third time interval is shortened from 30 minutes to 15 minutes. Through this step, the equipment status is detected in reverse from the perspective of semi-finished product quality, so as to avoid the situation where the sensor still outputs normal data when the equipment is abnormal due to sensor damage, which affects the product quality. In particular, after the production equipment returns to normal, the increased first frequency and second frequency are adjusted back to the original values.

[0060] In this embodiment, dividing the equipment status of the production equipment includes the following steps:

[0061] Preset multiple data features and set the collection time, which includes multiple sub-time intervals. Whenever the collection time is reached, calculate the data features of the intermediate quality parameters within each collection time to form multiple data sets corresponding to the data features. Calculate the fluctuation value of the intermediate quality parameters within the monitoring interval based on the data sets.

[0062] The process capability is calculated based on the optimal quality value, the lowest quality value, the average quality value and the standard deviation of the quality values ​​of the intermediate quality parameters within the monitoring interval. If the fluctuation value is less than the first threshold, the process capability is greater than the second threshold, and the production equipment is classified as a stable state. If the fluctuation value is greater than or equal to the first threshold, the production equipment is classified as an unstable state. If the fluctuation value is less than the first threshold, the process capability is less than or equal to the second threshold, and the production equipment is classified as a deteriorated state.

[0063] In this embodiment, the data features include median, peak, skewness, and standard deviation.

[0064] As before, the duration of the sub-time interval is 2 hours, and 3 days include a total of 36 sub-time intervals, that is, the preset number is 36, and the collection time is set to 12 hours. The monitoring interval will include 6 collection time periods, and each collection time period includes 6 sub-time intervals. There will be 6 intermediate quality parameters in a data set. When the number of sub-time intervals reaches 36, the median, kurtosis, skewness and standard deviation of the 6 intermediate quality parameters are calculated. Since there are 6 collection time periods in the monitoring interval, 6 medians, 6 kurtosis, 6 skewness and 6 standard deviations will be obtained in the end. The six medians are combined into a median set, the six kurtosis are combined into a kurtosis set, and the other data features are similar; in this embodiment, the Mahalanobis distance is calculated for each set, such as for the median, the Mahalanobis distance between each median and the median set is calculated, each median will correspond to a Mahalanobis distance, all Mahalanobis distances are summed and averaged, and a value is obtained as the comprehensive distance corresponding to the median, and finally, the comprehensive distance calculated by the median, kurtosis, skewness and standard deviation is summed and averaged, and the obtained value is used as the fluctuation value. The larger the Mahalanobis distance, the lower the similarity between the data and the entire data set, and it can be deduced that the larger the fluctuation value, the greater the parameter fluctuation of the device, indicating that the device is more unstable. In other embodiments, the value of the set variance can also be used as the fluctuation value.

[0065] Afterwards, the process capability is calculated based on the optimal quality value, the lowest quality value, the average quality value and the standard deviation of the quality of the intermediate quality parameters. The specific calculation formula is: C = min[(U-μ) / 3s, (μ-L) / 3s], where C is the process capability, min[a,b] is the smaller value of a and b, U is the optimal quality value of the intermediate quality parameter, L is the lowest quality value of the intermediate quality parameter, μ is the average quality value of the intermediate quality parameter, and s is the standard deviation of the intermediate quality parameter. It can be seen from the formula that the greater the process capability, the better the state level of the equipment. Therefore, when the fluctuation value is less than the first threshold and the process capability is greater than the second threshold, it indicates that the production equipment is in a good state and stable, so it is classified as a stable state. If the fluctuation value is greater than or equal to the first threshold, whether the process capability is greater than the second threshold or less than the second threshold, it indicates that the current volatility of the production equipment is large, so it is classified as an unstable state. If the fluctuation value is less than the first threshold, the process capability is less than or equal to the second threshold, indicating that the equipment is in a stable and poor state, so the production equipment is classified as a deteriorated state.

[0066] In this embodiment, after the equipment status of the monitoring interval is generated, the corresponding intermediate quality parameters are retained and accumulated as previous data. After the equipment status evaluation of the subsequent monitoring interval is completed, the previous data is added to the intermediate quality parameters of the subsequent monitoring interval to obtain the overall fluctuation value and process capability in multiple consecutive monitoring intervals.

[0067] For example, after determining the equipment status of monitoring interval 1 and monitoring interval 2, the intermediate quality parameters in monitoring interval 1 are fused with the intermediate quality parameters in monitoring interval 2 to obtain intermediate quality parameters with a longer time span. Finally, the overall fluctuation value and process capability since the start of operation of the production equipment are evaluated based on the fused intermediate quality parameters, that is, the overall equipment status of monitoring interval 1 and monitoring interval 2 is evaluated, thereby realizing the evaluation of the operating status of the production equipment throughout the entire cycle.

[0068] In this embodiment, after the fluctuation value and process capability are obtained, the first frequency and the second frequency are adjusted based on the following method. For the first frequency, the first formula is used to adjust the first frequency, and the first formula is: F1=(f1+α·W)·(1+γ / C); for the second frequency, the second formula is used to adjust the second frequency, and the second formula is, F2=(f2+0.5α·W)·(1+γ / 2C), wherein F1 and F2 are the adjusted first frequency and the second frequency, respectively, f1 and f2 are the first frequency and the second frequency before adjustment, respectively, α and γ are preset adjustment coefficients, which are set to 10 and 20 in this embodiment, W is the fluctuation value, and C is the process capability. The values ​​of the fluctuation value and the process capability are usually between 0-4; it can be seen that in the first formula and the second formula, since the fluctuation value is directly added to the original frequency and then multiplied by the calculated result of the process capability, the influence of the fluctuation value on the calculated result is less than that of the process capability, and the process capability has a greater influence on the adjustment result of the frequency.

[0069] In this embodiment, dividing the key parameters and the non-key parameters includes the following steps:

[0070] The coefficient corresponding to each production process parameter in the first prediction model is obtained, and the influence weight of each production process parameter is calculated based on the proportion of the coefficient in the sum of all coefficients. The production process parameters with influence weights greater than the critical threshold are set as key parameters, and the production process parameters less than or equal to the critical threshold are set as non-key parameters.

[0071] Assume that in the first prediction model, the linear regression equation of the long-term prediction model is y = β1x1 + β2x2 + ... + β n x n +ε, where β1~βk all correspond to the temperature parameter of the production process, then the influence weight of temperature is Among them, p is the impact weight of temperature, n is the number of coefficients in the linear regression equation, βm is the mth coefficient in the long-term prediction model, and the calculation process of the remaining coefficients is similar. The critical threshold is set to 0.6. When the impact weight is greater than 0.6, the corresponding production process parameter is determined as a key parameter, otherwise it is determined as a non-key parameter.

[0072] like Figure 2As shown, the present invention also provides a control and management system for road aggregate and asphalt, which is used to implement the above-mentioned control and management method for road aggregate and asphalt, and the system includes:

[0073] An analysis module is used to collect production process parameters and product quality parameters in historical production processes, perform correlation analysis on the production process parameters, obtain conversion functions for mutually converting production process parameters, perform correlation analysis on the production process parameters and product quality parameters, obtain the influence weight of each production process parameter on the product quality parameter, and divide the production process parameters into key parameters and non-key parameters based on the influence weight;

[0074] An acquisition module is used to sample key parameters and non-key parameters based on the first frequency and the second frequency respectively in an actual production process to obtain a plurality of first time series data and second time series data;

[0075] A preprocessing module, used for time-aligning the first time series data and the second time series data to obtain aligned data, and supplementing the missing values ​​of the aligned data based on a conversion function to obtain complete data;

[0076] A judgment module is used to establish a first prediction model and a second prediction model based on key parameters, non-key parameters and product quality parameters. If each production process parameter in the complete data is within the corresponding standard numerical range, the complete data is processed based on the first prediction model and the product quality parameters are output; otherwise, the complete data is processed based on the second prediction model and the product quality parameters are output.

[0077] The asphalt production management system and aggregate production management system of the present invention are as follows Figure 3 and Figure 4 As shown, the system can monitor asphalt production in real time, for example Figure 3 In the asphalt real-time production column, the production status of the default company will be displayed first. You can select the company through the drop-down box and view the unprocessed warning information. You can display the details of the asphalt production process by clicking on the details page. You can view the historical status of the equipment by clicking on the "Time" drop-down filter box and clicking on the query. In addition, both the asphalt and aggregate management systems include a mix ratio column. The mix ratio module can adjust the product ratio parameters of each subordinate branch. The test management column can view the asphalt test results of each processing plant. The warning management column arranges all warning information in chronological order, including the cause of the warning, time, whether it is processed, and the name of the relevant person in charge. The transportation track column is used to display transportation data in a table format, including transportation batches, license plate information, cargo information, and transportation time; the track query column can display the complete transportation track, and the video monitoring column is used to connect to the camera in the factory area, and can be filtered according to the location of the monitoring screen.

[0078] It should be understood that the various technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the various technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0079] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for controlling and managing road aggregate and asphalt, characterized in that: include: Collecting production process parameters and product quality parameters in historical production processes, performing correlation analysis on the production process parameters, and obtaining conversion functions for converting the production process parameters to each other; Performing a correlation analysis on the production process parameters and the product quality parameters to obtain the influence weight of each production process parameter on the product quality parameter, and dividing the production process parameters into key parameters and non-key parameters based on the influence weight; Establishing a first prediction model and a second prediction model based on the key parameters, the non-key parameters and the product quality parameters; In the actual production process, the key parameters and the non-key parameters are sampled based on the first frequency and the second frequency respectively to obtain a plurality of first time series data and a second time series data; Time-aligning the first time series data and the second time series data to obtain aligned data, and supplementing the missing values ​​of the aligned data based on the conversion function to obtain complete data; If each of the production process parameters in the complete data is within the corresponding standard numerical range, the complete data is processed based on the first prediction model and the product quality parameters are output; otherwise, the complete data is processed based on the second prediction model and the product quality parameters are output.

2. The method according to claim 1, characterized in that The first prediction model is a linear regression model, which includes a long-term prediction model and a medium-term prediction model; the second prediction model is a neural network model; a first time interval, a second time interval and a third time interval are set; the second time interval is smaller than the first time interval, and the third time interval is smaller than the second time interval; The long-term prediction model, the medium-term prediction model and the second prediction model respectively predict the product quality parameters in the first time period, the second time period and the third time period in the future based on the production process parameters collected corresponding to the first time interval, the second time interval and the third time interval.

3. The method according to claim 2, characterized in that The first frequency and the second frequency are adjusted based on the following steps: A monitoring interval is set, and the monitoring interval includes multiple sub-time intervals. Whenever the length of the sub-time interval is reached, the intermediate quality parameters of the semi-finished products output by each production equipment are obtained. After the number of the sub-time intervals reaches a predetermined number, the equipment state of the corresponding production equipment is divided into a stable state, an unstable state and a deteriorated state based on all the intermediate product quality parameters collected in the monitoring interval. If the equipment state of the production equipment changes from the stable state to the unstable state or the deteriorated state in an adjacent monitoring interval, the first frequency and the second frequency are increased, and the lengths of the first time interval, the second time interval and the third time interval are shortened.

4. The method according to claim 3, characterized in that Classifying the equipment status of production equipment includes the following steps: Preset multiple data features and set a collection time, wherein the collection time includes multiple sub-time intervals, and when the collection time is reached, calculate the data features of the intermediate quality parameter in each collection time to form multiple data sets corresponding to the data features, and calculate the fluctuation value of the intermediate quality parameter in the monitoring interval based on the data sets; The process capability is calculated based on the optimal quality value, the lowest quality value, the average quality value and the standard deviation of the quality value of the intermediate quality parameter in the monitoring interval; if the fluctuation value is less than the first threshold, the process capability is greater than the second threshold, and the production equipment is classified as the stable state; if the fluctuation value is greater than or equal to the first threshold, the production equipment is classified as the unstable state; if the fluctuation value is less than the first threshold, the process capability is less than or equal to the second threshold, and the production equipment is classified as the deteriorated state.

5. The method according to claim 4, characterized in that After generating the equipment status of the monitoring interval, the corresponding intermediate quality parameters are retained and accumulated as preceding data; after completing the equipment status evaluation of the subsequent monitoring interval, the preceding data are added to the intermediate quality parameters of the subsequent monitoring interval to obtain the overall fluctuation value and the process capability in multiple consecutive monitoring intervals.

6. The method according to claim 1, characterized in that The division of the key parameters and the non-key parameters comprises the following steps: Obtain the coefficient corresponding to each of the production process parameters in the first prediction model, calculate the impact weight of each of the production process parameters based on the proportion of the coefficient in the sum of all the coefficients, set the production process parameters whose impact weight is greater than a critical threshold as the key parameters, and set the production process parameters whose impact weight is less than or equal to the critical threshold as the non-key parameters.

7. The method according to claim 2, characterized in that The second prediction model is an LSTM model.

8. The method according to claim 4, characterized in that The data characteristics include median, kurtosis, skewness, and standard deviation.

9. A control and management system for road aggregate and asphalt, used to implement the method according to any one of claims 1 to 8, characterized in that: include: An analysis module, used to collect production process parameters and product quality parameters in historical production processes, perform correlation analysis on the production process parameters, obtain conversion functions for mutually converting the production process parameters, perform correlation analysis on the production process parameters and the product quality parameters, obtain the influence weight of each of the production process parameters on the product quality parameters, and divide the production process parameters into key parameters and non-key parameters based on the influence weight; A collection module, used for sampling the key parameters and the non-key parameters based on the first frequency and the second frequency respectively in an actual production process to obtain a plurality of first time series data and a second time series data; A preprocessing module, used for performing time alignment on the first time series data and the second time series data to obtain aligned data, and supplementing missing values ​​on the aligned data based on the conversion function to obtain complete data; A judgment module is used to establish a first prediction model and a second prediction model based on the key parameters, the non-key parameters and the product quality parameters. If each of the production process parameters in the complete data is within the corresponding standard numerical range, the complete data is processed based on the first prediction model and the product quality parameters are output; otherwise, the complete data is processed based on the second prediction model and the product quality parameters are output.

Images