Fine aggregate moisture content online detection method and system based on multi-source data fusion

Abstract

The present application belongs to the technical field of concrete mixing special equipment, and relates to a fine aggregate moisture content online detection method and a detection system based on multi-source data fusion. The fine aggregate moisture content online detection method based on multi-source data fusion comprises the following steps: calculating the moisture content of fine aggregate before being fed into a mixing host based on a microwave moisture model, and calculating a slump prediction value of concrete in the mixing host based on a slump prediction model; when the slump prediction value of the concrete is outside a preset range, it is determined that the moisture content of the fine aggregate deviates, the slump prediction value is input into a mix proportion prediction model based on a neural network, and the overall moisture content of the fine aggregate in the mixing host is calculated based on the mix proportion prediction model. The present application does not need to manually take a large number of samples for drying and measurement, thereby saving manpower and resources; online measurement is performed during the mixing process, so that large deviations existing in the detection system can be found in time and corrected; calibration through a large amount of manual experimental data is not needed, and the present application has strong adaptability to different production conditions.

Description

Technical Field

[0001] The present invention belongs to the technical field of special equipment for concrete mixing, and relates to an online detection method and detection system for fine aggregate moisture content based on multi-source data fusion. Background Art

[0002] During concrete production, fluctuations in fine aggregate moisture content can affect the actual concrete mix ratio. Accurate, real-time measurement of fine aggregate moisture content ensures consistent concrete quality. Existing methods commonly used domestically and internationally for calculating material moisture content include: drying methods, which are time-consuming; resistance and capacitance methods, which are subject to numerous interference factors; neutron methods, which pose potential radiation hazards; and infrared methods, which have limited penetration and can only measure surface moisture.

[0003] To address the above issues, the invention patent application with application number CN202011543379.2 discloses a microwave-based aggregate moisture content detection method. The microwave method, which measures moisture content, is a sampling detection method. Existing sampling detection methods suffer from delayed measurement results, multiple interference factors, and low accuracy for the actual production of concrete mixing plants. In concrete mixing plant production operations, the use of existing microwave measurement methods has the following problems: First, calibration is required before use for different production conditions in different mixing plants; second, the microwave moisture model corresponding to different material sources requires a large amount of sampled and dried measurement data to generate; third, due to changes in the fine aggregate source, frequent changes in weather conditions, and changes in silo conditions, the pre-measurement method cannot accurately reflect the moisture content of the fine aggregate, and the moisture content of the fine aggregate will change frequently. Therefore, it cannot meet the needs of the concrete mixing plant to automatically adjust the water consumption. Summary of the Invention

[0004] In order to overcome the deficiencies in the prior art, the present invention provides an online detection method and detection system for fine aggregate moisture content based on multi-source data fusion.

[0005] To achieve the above objectives, the online detection method of fine aggregate moisture content based on multi-source data fusion is developed. The specific steps are as follows:

[0006] Transmit multi-segment spectrum signals to the fine aggregate transported into the mixing main unit and collect microwave signals;

[0007] The collected microwave signals are input into the microwave moisture model, and the moisture content of the fine aggregate before being fed into the mixing main unit is calculated based on the microwave moisture model;

[0008] Acquiring characteristic information of a current curve of the mixing main unit during the mixing process, inputting the current curve characteristics into a slump prediction model based on a neural network, and calculating a predicted slump value of concrete in the mixing main unit based on the slump prediction model;

[0009] When the predicted slump value of concrete is within the preset range, it is determined that the moisture content of the fine aggregate meets the requirements;

[0010] When the predicted slump value of concrete is outside the preset range, the moisture content deviation of the fine aggregate is determined, and the slump prediction value is input into the mix ratio prediction model based on the neural network. The overall moisture content of the fine aggregate in the mixing main unit is calculated based on the mix ratio prediction model.

[0011] Furthermore, image information of the fine aggregate being transported into the mixing main unit is acquired, and the source type of the fine aggregate is determined based on the acquired image information.

[0012] Furthermore, the microwave moisture model is classified into a plurality of models according to the different sources of the fine aggregate, and the collected microwave signal is input into the microwave moisture model of the corresponding source according to the type of the source of the fine aggregate.

[0013] Furthermore, the current curve of the mixing main machine during the mixing process obtained from the historical production data is used to extract the current curve features. Combined with the corresponding slump label, the slump prediction model is trained using the training data to obtain the slump prediction model.

[0014] Furthermore, the preset range of the concrete slump prediction value is compared with the concrete slump of the construction standard, when the deviation is ±30mm.

[0015] Furthermore, the mass of raw materials per cubic meter of concrete, the amount of fine aggregate used, and the water consumption for slump estimation extracted from historical production data are used as model input features, and the moisture content of fine aggregate is used as output to construct a mix proportion prediction model based on a neural network model.

[0016] Furthermore, when the predicted slump value of concrete is outside the preset range, the characteristics of the effective microwave signal of fine aggregate are extracted, combined with the moisture content value calculated based on the slump prediction value, to calculate the correction parameters of the microwave moisture model, and the microwave moisture model is trained again.

[0017] The fine aggregate moisture content calculation system in the mixing host based on multi-source data fusion is characterized by comprising a microwave measurement module, a current extraction module, a database, and a data processing module;

[0018] The microwave measurement module is used to transmit multi-segment spectrum signals to the fine aggregate conveyed into the mixing main unit, and input the collected microwave signals into the data processing module;

[0019] A current extraction module is used to obtain the current curve characteristic information of the stirring host during the stirring process, and input the current curve characteristics into the data processing module;

[0020] Database module, used to store microwave moisture model, slump prediction model and mix ratio prediction model;

[0021] The data processing module is used to receive the microwave signal input by the microwave measurement module and call the microwave moisture model in the database module to calculate the moisture content of the fine aggregate before being sent to the mixing host. It receives the current curve characteristics input by the current extraction module and calls the slump prediction model to calculate the slump prediction value of the concrete in the mixing host. According to the calculated slump prediction value, it calls the mix ratio prediction model in the database to calculate the overall moisture content of the fine aggregate in the mixing host.

[0022] Furthermore, it also includes a material source identification module for collecting images of fine aggregates being transported and extracting image information related to identifying the type of fine aggregates in the images. The material source identification module sends the image information to the data processing module.

[0023] Furthermore, it also includes a correction and update module for extracting the characteristics of the effective microwave signal of fine aggregate, combining the moisture content value calculated based on the slump prediction value, calculating the correction parameters of the microwave moisture model, and retraining the microwave moisture model.

[0024] As can be seen from the above description of the present invention, the present invention provides an online detection system and method for the moisture content of fine aggregate. Based on multi-source data such as image information, microwave signals, and slump prediction values, the moisture content of fine aggregate in the mixing main unit is calculated through a microwave moisture model, a slump prediction model, and a mix ratio prediction model. There is no need for manual sampling, drying, and measurement, thus saving manpower and material resources. Online measurement is performed during the mixing process of the mixing equipment. When the source of the transported fine aggregate changes, the moisture content of the fine aggregate before being fed into the mixing main unit is verified by the slump prediction value in the mixing main unit. Large deviations in the detection system can be discovered and corrected in a timely manner to ensure the quality of concrete production. When the online detection system for the moisture content of fine aggregate provided by this patent is applied to different mixing stations, it does not require calibration through a large amount of manual experimental data. It only needs to adjust the moisture content data of the mix ratio prediction model to perform a simple correction on the microwave moisture model of the same type of fine aggregate to achieve moisture content detection. It has strong adaptability to different production conditions and is easy to migrate and apply in different mixing stations. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 Schematic diagram of the structure of the online detection system for fine aggregate moisture content based on multi-source data fusion. DETAILED DESCRIPTION

[0026] The present invention is further described below through specific embodiments.

[0027] The online detection method of fine aggregate moisture content based on multi-source data fusion has the following specific steps:

[0028] Fine aggregate is fed into the mixing main unit through the conveying system;

[0029] The system emits multi-segment spectrum signals to the fine aggregate being transported into the mixing main unit, collects microwave signals, and uses visual inspection technology to obtain image information of the fine aggregate being transported. The source type of the fine aggregate is determined based on the collected image information.

[0030] The microwave moisture model is divided into multiple models according to the different sources of fine aggregate. The collected microwave signal is input into the microwave moisture model of the corresponding source according to the type of fine aggregate source. The moisture content of the fine aggregate before being fed into the mixing main unit is calculated based on the microwave moisture model.

[0031] Obtaining characteristic information of the current curve of the mixer during the mixing process, inputting the current curve characteristics into a slump prediction model based on a neural network, and calculating a predicted slump value of the concrete in the mixer based on the slump prediction model. The slump prediction model is obtained by using the current curve of the mixer during the mixing process obtained from historical production data, extracting the current curve characteristics, combining the corresponding slump labels, and training the slump prediction model using training data to obtain a slump prediction model;

[0032] When the predicted slump value of concrete is within the preset range, the preset range of the predicted slump value of concrete is compared with the slump of concrete of the construction standard. If the deviation is ±30mm, it is determined that the moisture content of the fine aggregate meets the requirements;

[0033] When the predicted slump value of concrete is outside the preset range, the moisture content of the fine aggregate is determined to be deviation, indicating that the actual mix ratio of the concrete mixing raw materials does not meet the planned mix ratio, indicating that there is a large error between the actual value of the fine aggregate in the mixer and the moisture content of the fine aggregate measured before it is fed into the mixer.

[0034] At this time, the slump prediction value is input into a mix ratio prediction model based on a neural network. The overall moisture content of the fine aggregate in the mixing main unit is calculated based on the mix ratio prediction model. The mix ratio prediction model uses the mass of concrete raw materials per cubic meter (unit: kg / m3) extracted from historical production data, the amount of fine aggregate used, and the water consumption estimated by slump as model input features, and uses the moisture content of the fine aggregate as output to construct a mix ratio prediction model based on a neural network model.

[0035] When the predicted slump value of concrete is outside the preset range, the characteristics of the effective microwave signal of the fine aggregate are extracted, and the moisture content value calculated based on the slump prediction value is combined to calculate the correction parameters of the microwave moisture model. The microwave moisture model is retrained, which can effectively improve the parts of the microwave moisture model with large errors, complete the model correction, and store the retrained model update in the current model to realize the update of the moisture model.

[0036] Reference Figure 1 The online detection system for fine aggregate moisture content based on multi-source data fusion shown includes a microwave measurement module 1, a material source identification module 2, a current extraction module 3, a database 4, a data processing module 5, and a correction and update module 6;

[0037] The microwave measurement module 1 is used to transmit multi-segment spectrum signals to the fine aggregate transported into the mixing main unit, and input the collected microwave signals into the data processing module 5;

[0038] The material source identification module 2 is used to collect images of the fine aggregate being transported, extract image information related to identifying the type of the fine aggregate in the image, and send the image information to the data processing module 5;

[0039] The current extraction module 3 is used to obtain the current curve characteristic information of the stirring host during the stirring process, and input the current curve characteristics into the data processing module 5;

[0040] Database module 4, used for storing microwave moisture model, slump prediction model and mix ratio prediction model;

[0041] The data processing module 5 includes a calculation unit 51 and a comparison unit 52. The calculation unit 51 is used to receive the microwave signal input by the microwave measurement module 1 and call the microwave moisture model in the database module to calculate the moisture content of the fine aggregate before it is fed into the mixing main unit; receive the image information collected by the material source identification module 2 to determine the source type of the fine aggregate; receive the current curve characteristics input by the current extraction module 3 and call the slump prediction model to calculate the slump prediction value of the concrete in the mixing main unit;

[0042] The comparison unit 52 compares the calculated slump prediction value with the concrete slump of the construction standard. If the deviation is within the range of ±30 mm, it is determined that the moisture content of the fine aggregate meets the requirements. If the deviation is outside the range of ±30 mm, it is determined that the moisture content of the fine aggregate has a large deviation and correction is made.

[0043] The calculation unit 51 uses the mix ratio prediction model in the database to calculate the overall moisture content of the fine aggregate in the mixing main machine according to the calculated slump prediction value;

[0044] The correction and update module 6 is used to extract the characteristics of the effective microwave signal of the fine aggregate when the moisture content of the fine aggregate deviates, and calculate the correction parameters of the microwave moisture model through the partial least squares algorithm based on the moisture content value calculated based on the slump prediction value, and send the correction parameters to the database module 4 to re-train the microwave moisture model.

[0045] Compared with the existing technology, the present invention has the following advantages: based on multi-source data such as image information, microwave signals, and slump prediction values, the moisture content of fine aggregate in the mixing main unit is calculated through a microwave moisture model, a slump prediction model, and a mix ratio prediction model, eliminating the need for manual large-scale sampling, drying, and measurement, thus saving manpower and material resources;

[0046] Online measurement is performed during the mixing process of the mixing equipment. When the source of the transported fine aggregate changes, the slump prediction value in the mixing unit is used to verify the moisture content of the fine aggregate before it is fed into the mixing unit. This allows for timely detection and correction of any significant deviations in the detection system, ensuring the quality of concrete production.

[0047] When the fine aggregate moisture content online detection system provided by this patent is applied in different mixing plants, it does not require calibration through a large amount of manual experimental data. It only needs to adjust the moisture content data of the mix ratio prediction model and perform a simple correction on the microwave moisture model of the same category of fine aggregate to realize moisture content detection. It has strong adaptability to different production conditions and is easy to migrate and apply in different mixing plants.

[0048] The above are only some specific implementation methods of the present invention, but the design concept of the present invention is not limited to this. Any non-substantial changes to the present invention using this concept shall be deemed as an infringement of the protection scope of the present invention.

Claims

1. The online detection method of fine aggregate moisture content based on multi-source data fusion is characterized by: The specific steps are as follows: The system emits multi-segment spectrum signals to the fine aggregate entering the mixing main unit on the conveyor belt, collects microwave signals, and simultaneously uses visual inspection technology to obtain image information of the fine aggregate being conveyed. The source type of the fine aggregate is determined based on the collected image information. The microwave moisture model is divided into multiple models according to the different sources of fine aggregate. The collected microwave signal is input into the microwave moisture model of the corresponding source according to the type of fine aggregate source. The moisture content of the fine aggregate before being fed into the mixing main unit is calculated based on the microwave moisture model. Acquiring characteristic information of a current curve of the mixing main unit during the mixing process, inputting the current curve characteristics into a slump prediction model based on a neural network, and calculating a predicted slump value of concrete in the mixing main unit based on the slump prediction model; When the predicted slump value of concrete is within the preset range, it is determined that the moisture content of the fine aggregate meets the requirements; When the predicted slump value of concrete is outside the preset range, the moisture content deviation of the fine aggregate is determined, the slump prediction value is input into the mix ratio prediction model based on the neural network, and the overall moisture content of the fine aggregate in the mixing main unit is calculated based on the mix ratio prediction model; When the predicted slump value of concrete is outside the preset range, the characteristics of the effective microwave signal of the fine aggregate are extracted. Combined with the overall moisture content calculated based on the slump prediction value, the correction parameters of the microwave moisture model are calculated. The microwave moisture model is retrained and the retrained model update is stored in the current model to achieve the update of the microwave moisture model.

2. The online detection method for fine aggregate moisture content based on multi-source data fusion according to claim 1, characterized in that: The current curve of the mixing main unit during the mixing process obtained from historical production data is used to extract the current curve features. Combined with its corresponding slump label, the slump prediction model is trained using the training data to obtain a slump prediction model.

3. The online detection method for fine aggregate moisture content based on multi-source data fusion according to claim 1, characterized in that: The mass of raw materials per cubic meter of concrete, the amount of fine aggregate used, and the water consumption estimated by slump extracted from historical production data are used as model input features, and the overall moisture content of fine aggregate is used as output to construct a mix proportion prediction model based on a neural network model.

4. The fine aggregate moisture content calculation system in the mixing main unit based on multi-source data fusion is characterized by: Including microwave measurement module, current extraction module, database, and data processing module; The microwave measurement module is used to transmit multi-segment spectrum signals to the fine aggregate entering the mixing main unit on the conveyor belt, and input the collected microwave signals into the data processing module; A current extraction module is used to obtain the current curve characteristic information of the stirring host during the stirring process, and input the current curve characteristics into the data processing module; A database for storing microwave moisture models, slump prediction models, and mix ratio prediction models; a data processing module configured to receive the microwave signal inputted by the microwave measurement module and call the microwave moisture model in the database to calculate the moisture content of the fine aggregate before being fed into the mixing main unit; receive the current curve characteristic inputted by the current extraction module and call the slump prediction model to calculate the slump prediction value of the concrete in the mixing main unit; and call the mix ratio prediction model in the database based on the calculated slump prediction value to calculate the overall moisture content of the fine aggregate in the mixing main unit; A material source identification module is used to collect images of fine aggregate being transported, extract image information related to identifying the type of fine aggregate in the image, and send the image information to the data processing module; The correction and update module is used to extract the characteristics of the effective microwave signal of fine aggregate, combine it with the overall moisture content calculated based on the slump prediction value, calculate the correction parameters of the microwave moisture model, retrain the microwave moisture model, and update the retrained model and store it in the current model to achieve the update of the microwave moisture model.

Images