An online adaptive control method and system for breaking sieves of dry ash agglomerates

Abstract

An online adaptive sieving control method and system for dry ash caking, which relates to the technical field of sieving control. By arranging sensors on the dry ash conveying path to monitor the dry ash in real time, a caking identification model is constructed. The target dry ash is identified for caking through the model, and caking classification identification is carried out based on the identification result. Multiple sieving control strategies are formulated, and then an optimization adjustment information set is obtained through the feedback adjustment unit to continuously optimize and adjust the multiple sieving control strategies, obtaining multiple sieving control adjustment strategies. Online sieving control is carried out according to the implemented adjustment strategies, solving the technical problems of inability to perform precise sieving control according to the actual situation, low sieving efficiency, and easy damage to equipment, achieving the technical effect of adaptively performing sieving control adjustment and efficiently and precisely performing dry ash sieving, improving the processing level and equipment life of dry ash sieving, and reducing the risk of equipment damage.

Description

Technical Field

[0001] The present invention relates to the technical field of broken sieve control, and particularly relates to an online adaptive broken sieve control method and system for dry ash caking. Background Art

[0002] The broken sieve of dry ash caking refers to the process in which during the transportation or storage of dry ash, due to factors such as humidity, temperature, and particle size, the dry ash forms lumps, and then these lumps are broken into smaller particles or powders through specific screening equipment or methods so that they can flow smoothly and continue the subsequent transportation or utilization process.

[0003] Existing broken sieve control methods for dry ash caking often lack real-time performance and self-adaptability, and cannot accurately adjust according to the actual situation of dry ash caking, resulting in low screening efficiency and even possible damage to the screening equipment. Summary of the Invention

[0004] This application provides an online adaptive broken sieve control method and system for dry ash caking, which is used to solve the technical problems existing in the prior art that the broken sieve control method often lacks real-time performance and self-adaptability, cannot accurately adjust according to the actual situation of dry ash caking, resulting in low screening efficiency and even possible damage to the screening equipment.

[0005] In view of the above problems, this application provides an online adaptive broken sieve control method and system for dry ash caking.

[0006] In a first aspect, this application provides an online adaptive broken sieve control method for dry ash caking. The method is applied to an online adaptive broken sieve control system for dry ash caking. The online adaptive broken sieve control system for dry ash caking is communicatively connected to a feedback adjustment unit and a remote terminal. The method includes:

[0007] Deploy a plurality of sensors along the dry ash transportation path, and through the plurality of sensors, perform real-time monitoring on the target dry ash to obtain real-time flow state information of the target dry ash;

[0008] Construct a caking recognition model based on the real-time flow state information;

[0009] Activate the caking recognition model to perform caking recognition on the target dry ash to obtain a caking recognition result;

[0010] Perform caking classification identification based on the caking recognition result, and formulate a plurality of broken sieve control strategies according to the caking classification identification result;

[0011] Obtain an optimization adjustment information set through the feedback adjustment unit, and continuously optimize and adjust the plurality of broken sieve control strategies according to the optimization adjustment information set to obtain a plurality of broken sieve control adjustment strategies;

[0012] Extract the real-time dry ash conveying status information and real-time dry ash caking status information, execute the multiple screening-breaking control adjustment strategies, and obtain the policy execution configuration parameters;

[0013] Connect to the remote terminal, send an execution instruction through the remote terminal, and based on the execution instruction, start the target screening-breaking device for online screening-breaking intelligent control according to the policy execution configuration parameters.

[0014] In a second aspect, the present application provides an online adaptive screening-breaking control system for dry ash caking. The system is communicatively connected to a feedback adjustment unit and a remote terminal. The system includes:

[0015] A real-time monitoring module, configured to deploy a plurality of sensors based on the dry ash conveying path, and through the plurality of sensors, perform real-time monitoring on the target dry ash to obtain the real-time flow status information of the target dry ash;

[0016] An identification model construction module, configured to construct a caking identification model based on the real-time flow status information;

[0017] An identification result acquisition module, configured to activate the caking identification model to perform caking identification on the target dry ash and obtain the caking identification result;

[0018] A screening-breaking control strategy formulation module, configured to perform caking classification identification based on the caking identification result, and formulate a plurality of screening-breaking control strategies according to the caking classification identification result;

[0019] A screening-breaking control adjustment module, configured to obtain an optimization adjustment information set through the feedback adjustment unit, and continuously optimize and adjust the plurality of screening-breaking control strategies according to the optimization adjustment information set to obtain a plurality of screening-breaking control adjustment strategies;

[0020] A configuration parameter acquisition module, configured to extract the real-time dry ash conveying status information and real-time dry ash caking status information, execute the plurality of screening-breaking control adjustment strategies, and obtain the policy execution configuration parameters;

[0021] An intelligent control module, configured to connect to the remote terminal, send an execution instruction through the remote terminal, and based on the execution instruction, start the target screening-breaking device for online screening-breaking intelligent control according to the policy execution configuration parameters.

[0022] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0023] An online adaptive screening control method for dry ash caking provided by the present application monitors dry ash in real time by arranging multiple sensors on the dry ash conveying path, constructs a caking recognition model based on real-time flow state information, identifies caking of target dry ash through the caking recognition model, and then classifies and labels the caking based on the caking recognition result, formulates multiple screening control strategies, and continuously optimizes and adjusts the multiple screening control strategies by obtaining an optimized adjustment information set through a feedback adjustment unit to obtain multiple screening control adjustment strategies, extracts real-time dry ash conveying state information and real-time dry ash caking state information, executes the multiple screening control adjustment strategies, obtains strategy execution configuration parameters, connects to a remote terminal, sends an execution instruction through the remote terminal, and starts the target screening device for online screening intelligent control based on the execution instruction according to the strategy execution configuration parameters, solving the technical problems in the prior art that the screening control method often lacks real-time performance and self-adaptability, cannot accurately adjust according to the actual situation of dry ash caking, resulting in low screening efficiency and even possible damage to the screening equipment, achieving the technical effect of performing real-time feedback based on the screening effect, adaptively adjusting the screening control according to the actual situation, and efficiently and accurately screening dry ash, not only improving the processing level of dry ash screening, but also increasing the service life of the screening equipment and reducing the risk of equipment damage. Description of the Drawings

[0024] Figure 1 It is a schematic flowchart of an online adaptive screening control method for dry ash caking provided by the present application;

[0025] Figure 2 It is a schematic structural diagram of an online adaptive screening control system for dry ash caking provided by the present application.

[0026] Description of the reference numerals: Real-time monitoring module 11, recognition model construction module 12, recognition result acquisition module 13, screening control strategy formulation module 14, screening control adjustment module 15, configuration parameter acquisition module 16, intelligent control module 17. Detailed Embodiments

[0027] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0028] Embodiment 1

[0029] As Figure 1As shown, the present application provides an online adaptive sieving control method for dry ash caking. The method is applied to an online adaptive sieving control system for dry ash caking. The online adaptive sieving control system for dry ash caking is communicatively connected to a feedback adjustment unit and a remote terminal. The method includes:

[0030] Step S100: Deploy multiple sensors along the dry ash conveying path, and use the multiple sensors to monitor the target dry ash in real time to obtain the real-time flow state information of the target dry ash;

[0031] Furthermore, step S100 of the present application further includes:

[0032] Step S110: Retrieve the historical conveying data record file of the target dry ash, and obtain multiple position information and multiple flow state information of the target dry ash at multiple positions in the dry ash conveying path. Among them, the multiple position information and the multiple flow state information are in a corresponding relationship;

[0033] Step S120: Determine multiple key positions of the dry ash conveying path based on the multiple position information and the multiple flow state information;

[0034] Step S130: Deploy the multiple sensors according to the multiple key positions;

[0035] Step S140: Monitor the real-time flow states of the multiple key positions in the dry ash conveying path according to the multiple sensors, and extract effective flow features according to the monitoring results;

[0036] Step S150: Form comprehensive flow state information of the dry ash conveying path based on the effective flow features, and output the real-time flow state information.

[0037] Specifically, to ensure the self - adaptability and precise regulation of the screening - breaking control, first, multiple sensors are arranged in multiple paths of the conveying pipeline to monitor the flow condition of the target dry ash in real - time. Among them, the historical conveying archive data containing multiple position information and flow state information during the dry - ash conveying path is retrieved first to obtain an information set in which the position information and the flow state information correspond one by one. By analyzing the information set, multiple key position points in the dry - ash conveying path can be determined. The key position points reflect the pipeline positions where the flow state of dry ash is likely to change during the conveying process. For example, the starting section of the conveying pipeline transitioning from a static state to a steady - flow state, the pipeline turning point where, due to the change in pipeline direction and uneven dry - ash particles, congestion is likely to occur under centrifugal force and wall friction, the pipeline branch point where the flow velocity changes or congestion is likely to occur due to the change in flow rate and flow direction, and the front and rear positions of the screening - breaking equipment related to the screening - breaking efficiency, etc. The pipeline position points where the influencing factors related to the screening - breaking control are traced are used as key position points for sensor layout, which can improve the accuracy of real - time monitoring and then globally control the conveying situation in the dry - ash pipeline.

[0038] Furthermore, the sensors include vibration sensors, temperature sensors, pressure sensors, load sensors, etc. By arranging multiple sensors corresponding to the factors affecting the dry - ash flow state at multiple key position points of the conveying pipeline, the real - time flow states of multiple key position points are monitored. Among them, multiple key position points have various types of real - time monitoring results, including flow velocity, pressure change, etc. The multiple types of monitoring data contain the monitoring information of multiple key position points. Then, the features related to the current flow state are extracted from the multiple monitoring results as effective flow features, that is, the effective flow features reflect the influencing factors affecting the current overall flow state. These factors can directly and effectively act on the overall flow state at that time, causing the flow state to change. Finally, based on the effective flow features, the comprehensive flow state information of the dry - ash conveying path is formed as the real - time flow state information. By forming the comprehensive flow state based on the effective features and then outputting the real - time flow state information, while ensuring the completeness of data acquisition and analysis, the effectiveness of subsequent analysis is improved, and thus the overall control efficiency is enhanced.

[0039] Step S200: Construct a caking recognition model based on the real - time flow state information;

[0040] Furthermore, step S200 of this application further includes:

[0041] Step S210: Construct a caking information library based on historical caking data, perform feature retrieval on the caking information library, and obtain a caking feature set;

[0042] Step S220: Traverse the real-time flow state information to match the caking feature set and generate a caking feature vector set;

[0043] Step S230: Determine a data training feature set, a data supervision feature set, and a data verification feature set based on the caking feature vector set. Use a deep learning algorithm to perform data training on the data training feature set and the data supervision feature set. Iteratively test the training results based on the data verification feature set until convergence, and obtain the caking recognition model.

[0044] Specifically, construct a caking information library based on historical caking data. The historical caking data includes archived caking degree data, data on factors causing caking, and identification information on the flow state, and there is a corresponding relationship. The caking degree data includes the caking rate, which reflects the proportion of caking; the caking particle size distribution, which reflects the particle size of the caking; the caking bulk density, which reflects the looseness and packing tightness of the caking; the caking hardness, which reflects the moisture content of the caking, etc. The data on factors causing caking includes the temperature and humidity of the storage environment, the storage time, etc. Store the above data in the same database to complete the construction of the caking information library, laying a foundation for the construction of the caking recognition model.

[0045] Furthermore, perform similarity feature retrieval on the content in the caking information library, classify and merge data elements with similar features to form a caking feature set. For example, combine caking-related data sets with similar hardness, particle size, and flow state identification as a caking feature set. The caking feature set includes multiple caking feature sets. Then, by traversing the real-time flow state information and sequentially matching it in the caking feature set, a caking feature vector set is generated. The caking feature vector set represents the contribution degree of each factor data corresponding to the real-time flow state information, that is, the weight value of each data information affecting the real-time flow state information. By generating the caking feature vector set, the flow condition of dry ash in the conveying pipeline can be analyzed in detail, thereby ensuring the effectiveness and accuracy of the subsequent caking recognition model.

[0046] Randomly divide the set of agglomeration feature vectors, and obtain a data training feature set, a data supervision feature set, and a data verification feature set according to the ratios of 60%, 25%, and 15% respectively. Build an agglomeration recognition model through a deep learning algorithm in combination with the above feature sets. First, select an appropriate neural network structure and model architecture, such as a fully connected neural network, a convolutional neural network (CNN), or a recurrent neural network (RNN), etc. Define the input layer, hidden layer, and output layer of the model, and set appropriate activation functions, loss functions, and optimization algorithms. Use the data training feature set to train the model, and use the data supervision feature set to guide the training process of the model, so that the model can identify and distinguish different categories of data, optimize the model structure, and use the data verification feature set to evaluate and iteratively optimize the model performance. When the training result tends to be stable and converges, obtain the agglomeration recognition model. Through the construction of the agglomeration recognition model based on the real-time flow state information, the accuracy of the subsequent obtained recognition result is improved, and thus the accuracy of controlling the dry ash agglomeration according to the actual situation is ensured.

[0047] Step S300: Activate the agglomeration recognition model to perform agglomeration recognition on the target dry ash, and obtain an agglomeration recognition result;

[0048] Furthermore, step S300 of the present application further includes:

[0049] Step S310: Activate the agglomeration recognition model to judge whether the target dry ash is less than a preset flow threshold in the dry ash conveying path. If so, it is regarded that there is an agglomeration phenomenon, traverse the set of agglomeration feature vectors for data transformation, and determine the deep agglomeration features;

[0050] Step S320: Analyze the change trend of the flow state of the target dry ash based on the deep agglomeration features;

[0051] Step S330: Perform agglomeration state recognition according to the change trend, and output the agglomeration recognition result based on the agglomeration state.

[0052] When performing screening, the caking recognition model is activated to judge the flow state of the target dry ash in the dry ash conveying path through the caking recognition model. Usually, it is understood by the flow velocity, but it is not limited to the flow velocity. It can also be other data elements that can characterize the flow state. Here, the flow velocity is taken as an example. When its flow velocity is less than the preset flow threshold, it indicates that there is a caking phenomenon of the dry ash in the conveying path. Based on this flow state information, traverse in the caking feature vector set, and then obtain the corresponding relevant factor information with this flow state identifier, such as the caking ratio, caking packing density, caking position, etc. under this flow state identifier. And the above relevant factor information includes the corresponding weight ratio. Therefore, the flow state information data can be transformed into relevant factor data in vector dimension, and then the deep caking features can be analyzed. Here, it is explained that the preset flow threshold can be set by those skilled in the art according to the actual working environment, selected data elements, etc.

[0053] Furthermore, the deep caking feature refers to the vector feature for analyzing and expressing the causes of each dimension of caking, which can be used to express the change direction and magnitude of data in different dimensions. Therefore, based on the deep caking feature, analyze and predict the change trend of the flow state of the target dry ash, and then identify the caking state of the dry ash according to the change of the overall dry ash flow state. Among them, the caking state identification refers to the judgment of the number of dry ash cakings, caking size, caking time in the dry ash conveying pipeline under the change trend of this flow state, and the caking hardness, etc. based on the temperature and humidity in the pipeline combined with time. Through the identified caking state, output the caking recognition result.

[0054] By activating the caking recognition model to perform caking recognition on the target dry ash and obtaining the caking recognition result, the cakings in the current flow state can be quickly classified, which lays a foundation for the formulation of subsequent screening control schemes, further improves the overall screening efficiency, and ensures the accuracy and self - adaptability of subsequent regulation according to the screening scheme.

[0055] Step S400: Based on the caking recognition result, conduct caking classification identification, and formulate multiple screening control strategies according to the caking classification identification result;

[0056] Specifically, based on the caking recognition result, conduct caking classification identification. For example, the recognition results within a certain threshold range of the caking growth rate and diffusion range size are classified and labeled under the same classification label, and the caking position is marked and divided based on the caking size, etc. According to the caking recognition result, multiple caking classification labels can be divided. Furthermore, each caking classification has corresponding multiple screening control strategy schemes. Therefore, perform global optimization among multiple screening control strategy schemes, and select the scheme with the highest caking treatment efficiency at each key position point as the final screening control strategy scheme to obtain multiple screening control strategies.

[0057] In the process of globally optimizing multiple screening control strategy schemes, multiple optimization strategies can be combined, such as a parallel global optimization algorithm based on interval mathematics, which reduces the computational amount and improves the optimization efficiency through parallelization and splitting methods.

[0058] The optimal screening control scheme is obtained by globally optimizing multiple screening control schemes, and then multiple control strategies based on the optimal scheme are obtained by decomposing the optimal scheme to each key position point, providing a basis for the optimization of subsequent multiple screening control adjustment strategies.

[0059] Step S500: Obtain an optimization adjustment information set through the feedback adjustment unit, and continuously optimize and adjust the multiple screening control strategies according to the optimization adjustment information set to obtain multiple screening control adjustment strategies;

[0060] Furthermore, by obtaining an optimization adjustment information set through the feedback adjustment unit, step S500 of this application further includes:

[0061] Step S510: Formulate a feedback target based on the optimization direction, and establish the feedback adjustment unit according to the feedback target;

[0062] Step S520: Evaluate the execution effects of the multiple screening control strategies through the feedback adjustment unit to obtain multiple strategy execution effects;

[0063] Step S530: Determine whether the multiple strategy execution effects meet the preset target. If not, generate a feedback result;

[0064] Step S540: Determine the optimization adjustment parameters, optimization adjustment direction, and optimization adjustment intensity according to the feedback result and the multiple strategy execution effects;

[0065] Step S550: Add the optimization adjustment parameters, the optimization adjustment direction, and the optimization adjustment intensity to the optimization adjustment information set.

[0066] Specifically, the screening effect of dry ash is monitored in real time, an optimization target is formulated based on the monitoring effect, the real-time screening effect is compared with the formulated optimization target, an optimization direction is obtained based on the comparison result, and then a feedback target is formulated based on the optimization direction. The feedback adjustment unit is established through the feedback target to evaluate the execution effects of the current multiple screening control strategies, and the evaluated execution effects are compared with the preset optimization target. If it still does not meet the expectation, a feedback result is generated to optimize and adjust the multiple screening control strategies, and the above process is iterated until the preset target effect is achieved.

[0067] Further, after generating the feedback result, deeply analyze the optimization and adjustment direction in combination with the execution effects of multiple current screening-breaking control strategies to obtain an optimization and adjustment information set. The optimization and adjustment information set includes optimization and adjustment parameters, optimization and adjustment directions, and optimization and adjustment intensities. Then, continuously optimize and adjust the multiple existing screening-breaking control strategies according to the optimization and adjustment information set to obtain multiple screening-breaking control adjustment strategies.

[0068] Through the in-depth analysis of the optimization and adjustment based on the execution effects of multiple strategies, the optimization objectives are clarified, and the optimization efficiency and the precise control of the screening-breaking are improved.

[0069] Furthermore, according to the optimization and adjustment information set, continuously optimize and adjust the multiple screening-breaking control strategies to obtain multiple screening-breaking control adjustment strategies. Step S500 of this application further includes:

[0070] Step S560: For each screening-breaking control strategy among the multiple screening-breaking control strategies, sequentially optimize and adjust according to the optimization and adjustment parameters, the optimization and adjustment directions, and the optimization and adjustment intensities in the optimization and adjustment information set to obtain multiple optimization and adjustment results;

[0071] Step S570: Conduct a screening-breaking simulation run according to the multiple optimization and adjustment results, and evaluate the adjustment effect based on the operation results according to the evaluation indicators to generate multiple adjustment evaluation results;

[0072] Step S580: Continuously optimize and adjust the multiple screening-breaking control strategies according to the multiple adjustment evaluation results, and output the multiple screening-breaking control adjustment strategies.

[0073] Specifically, in order to further improve the accuracy of the adjustment of the screening-breaking control strategy and the overall screening-breaking efficiency, after obtaining multiple screening-breaking control adjustment strategies, sequentially optimize and adjust each screening-breaking control strategy based on the optimization and adjustment parameters, the optimization and adjustment directions, and the optimization and adjustment intensities in the optimization and adjustment information set. Integrate the adjusted results and then conduct a screening-breaking simulation run based on the dry ash screening-breaking simulation software, and evaluate the adjustment effect of the simulation results according to the evaluation indicators. The evaluation indicators refer to the crushing rate of dry ash agglomerates, equipment energy consumption, equipment wear, etc. Through the evaluation based on the indicators, multiple adjustment evaluation results can be generated. Continuously optimize and adjust the multiple screening-breaking control strategies based on this result, and output multiple screening-breaking control adjustment strategies with the highest confidence level, and use this strategy to execute the subsequent online screening-breaking control.

[0074] Through the quasi - operation of parameters, with the execution effect as a reference, the execution effects of multiple screening - breaking control adjustment strategies are intuitively obtained in advance and optimized, further ensuring the credibility of the adjustment plan. By evaluating the adjustment results based on evaluation indicators, it not only reflects the execution effect after adjustment, but also incorporates equipment energy consumption, wear, etc. into the optimization scope, improving the service life of the equipment and reducing the loss risk of the screening equipment.

[0075] Step S600: Extract the real - time dry ash conveying status information and real - time dry ash caking status information, execute the multiple screening - breaking control adjustment strategies, and obtain the policy execution configuration parameters.

[0076] Step S700: Connect to the remote terminal, send an execution instruction through the remote terminal, and based on the execution instruction, start the target screening - breaking equipment for intelligent online screening - breaking control according to the policy execution configuration parameters.

[0077] Specifically, based on the real - time dry ash conveying status information and real - time dry ash caking status information, multiple screening - breaking control adjustment strategies are executed, and the execution parameters are configured according to the adjustment strategies. Among them, the execution configuration parameters include the running time, frequency, intensity, etc. of the screening - breaking equipment. Further, the screening - breaking control system communicates with the remote terminal, and an execution instruction is sent to the target screening - breaking equipment through the remote terminal to start and run the target screening - breaking equipment with the execution configuration parameters, realizing the intelligent control of online screening - breaking.

[0078] Furthermore, when starting the target screening - breaking equipment for intelligent online screening - breaking control, step S700 of this application further includes:

[0079] Step S710: Start the target screening - breaking equipment and conduct real - time operation monitoring on the target screening - breaking equipment to generate a screening - breaking operation control file.

[0080] Step S720: Based on the screening - breaking operation control file, conduct operation status identification according to the operation nodes to generate multiple operation status identification results.

[0081] Step S730: Extract the abnormal operation status identification data set according to the multiple operation status identification results.

[0082] Step S740: Trace the anomalies of the corresponding operation nodes according to the abnormal operation status identification data set to generate abnormal data sources.

[0083] Step S750: Conduct automatic anomaly correction based on the abnormal data sources to obtain an anomaly correction result, and adjust the operation deviation of the started target screening - breaking equipment according to the anomaly correction result.

[0084] Optionally, start the target screening device, and based on multiple sensors arranged at key positions, monitor the real-time operation data of the device, collect the operation data in real time and perform data analysis, identify the operation state of the device and form a screening operation control file. The data in the screening operation control file has a time sequence identifier, and then identify the operation state of the data in the file according to the operation nodes of the device to obtain multiple identification results. The operation nodes are key nodes in the operation process of the screening device, such as: device startup, stable operation, parameter adjustment, fault handling, shutdown maintenance, etc. If the device operates normally at a certain node, apply the corresponding normal state identifier. If a fault or abnormal situation occurs, apply the corresponding fault or abnormal state identifier. Extract the abnormal data in the screening operation control file through the identification results of the operation status to form a data set of abnormal states. Furthermore, based on the data set marked with abnormal operation states, the specific key position points, that is, the abnormal data sources, can be located. These positions are the key points for abnormal traceability analysis. Extract the data records at the key positions under this operation node, including environmental conditions, parameter adjustments, etc., and conduct a detailed analysis to find out the specific reasons for the abnormality. Automatically correct based on the abnormal data source according to the abnormal automatic correction strategy stored in the system in advance. These strategies can include adjusting device parameters, optimizing the operation environment, etc. Further, adjust the operation deviation of the target screening device through the parameters after abnormal correction to restrict the device to return to the normal operation level. If the correction result is not ideal, that is, the device still has an operation deviation, further adjust the operation parameters or operation strategies of the device through an optimization algorithm according to the correction result feedback to ensure that the device can operate stably in the predetermined working state. Therefore, through the above solution, the control adjustment adaptability and control accuracy of the screening device can be improved, and the operation level of dry ash screening is guaranteed.

[0085] Through the technical solution of the above embodiment, an online adaptive screening control method for dry ash caking provided by the present application solves the technical problems in the prior art that the screening control method often lacks real-time performance and adaptability, cannot accurately control according to the actual situation of dry ash caking, resulting in low screening efficiency and even possible damage to the screening equipment, and achieves the technical effect of real-time feedback based on the screening effect, adaptively performing screening control adjustment according to the actual situation, and efficiently and accurately screening dry ash. It not only improves the processing level of dry ash screening, but also improves the service life of the screening equipment and reduces the risk of equipment damage.

[0086] Embodiment 2

[0087] Based on the same inventive concept as the online adaptive screening control method for dry ash caking in the foregoing embodiment, as Figure 2As shown in the figure, the present application provides an online adaptive sieving-breaking control system for dry ash caking. The system is communicatively connected to a feedback regulation unit and a remote terminal. The system includes:

[0088] A real-time monitoring module 11, configured to arrange a plurality of sensors based on the dry ash conveying path, and perform real-time monitoring on the target dry ash through the plurality of sensors to obtain real-time flow state information of the target dry ash;

[0089] A caking identification model construction module 12, configured to construct a caking identification model based on the real-time flow state information;

[0090] An identification result acquisition module 13, configured to activate the caking identification model to perform caking identification on the target dry ash and obtain a caking identification result;

[0091] A sieving-breaking control strategy formulation module 14, configured to perform caking classification identification based on the caking identification result, and formulate a plurality of sieving-breaking control strategies according to the caking classification identification result;

[0092] A sieving-breaking control adjustment module 15, configured to obtain an optimization adjustment information set through the feedback regulation unit, and continuously optimize and adjust the plurality of sieving-breaking control strategies according to the optimization adjustment information set to obtain a plurality of sieving-breaking control adjustment strategies;

[0093] A configuration parameter acquisition module 16, configured to extract real-time dry ash conveying state information and real-time dry ash caking state information, execute the plurality of sieving-breaking control adjustment strategies, and obtain strategy execution configuration parameters;

[0094] An intelligent control module 17, configured to connect to the remote terminal, send an execution instruction through the remote terminal, and perform intelligent control of online sieving-breaking of the target sieving-breaking device based on the execution instruction according to the strategy execution configuration parameters.

[0095] Furthermore, the real-time monitoring module 11 is further configured to perform the following steps:

[0096] Retrieve the historical conveying data record file of the target dry ash, and obtain a plurality of position information and a plurality of flow state information of the target dry ash at multiple positions in the dry ash conveying path, wherein the plurality of position information and the plurality of flow state information are in a corresponding relationship;

[0097] Determine a plurality of key positions of the dry ash conveying path based on the plurality of position information and the plurality of flow state information;

[0098] Arrange the plurality of sensors according to the plurality of key positions;

[0099] Monitor the real-time flow state of the plurality of key positions in the dry ash conveying path according to the plurality of sensors, and extract effective flow characteristics according to the monitoring results;

[0100] Form comprehensive flow state information of the dry ash conveying path based on the effective flow characteristics, and output the real-time flow state information.

[0101] Furthermore, the recognition model construction module 12 is further configured to perform the following steps:

[0102] Construct a caking information database based on historical caking data, perform feature retrieval on the caking information database, and obtain a caking feature set;

[0103] Traverse the real-time flow state information to match the caking feature set, and generate a caking feature vector set;

[0104] Determine a data training feature set, a data supervision feature set, and a data verification feature set based on the caking feature vector set, use a deep learning algorithm to perform data training on the data training feature set and the data supervision feature set, and perform iterative testing on the training result based on the data verification feature set until convergence to obtain the caking recognition model.

[0105] Furthermore, the recognition result acquisition module 13 is further configured to perform the following steps:

[0106] Activate the caking recognition model to determine whether the target dry ash in the dry ash conveying path is less than a preset flow threshold. If so, it is regarded as a caking phenomenon. Traverse the caking feature vector set to perform data transformation and determine the deep caking features;

[0107] Analyze the change trend of the target dry ash flow state based on the deep caking features;

[0108] Perform caking state recognition according to the change trend, and output the caking recognition result based on the caking state.

[0109] Furthermore, the broken screen control adjustment module 15 is further configured to perform the following steps:

[0110] Formulate a feedback target based on the optimization direction, and establish the feedback adjustment unit according to the feedback target;

[0111] Evaluate the execution effects of the multiple broken screen control strategies through the feedback adjustment unit to obtain multiple strategy execution effects;

[0112] Judge whether the multiple strategy execution effects meet the preset target. If not, generate a feedback result;

[0113] Determine the optimization adjustment parameters, the optimization adjustment direction, and the optimization adjustment intensity according to the feedback result and the multiple strategy execution effects;

[0114] Add the optimized adjustment parameters, the optimized adjustment direction, and the optimized adjustment intensity to the optimized adjustment information set.

[0115] Furthermore, the screening breakage control adjustment module 15 is further configured to perform the following steps:

[0116] For each screening breakage control strategy among the multiple screening breakage control strategies, sequentially perform optimized adjustment according to the optimized adjustment parameters, the optimized adjustment direction, and the optimized adjustment intensity in the optimized adjustment information set, and obtain multiple optimized adjustment results;

[0117] Conduct a screening breakage simulation run based on the multiple optimized adjustment results, evaluate the adjustment effect according to the evaluation index based on the operation result, and generate multiple adjustment evaluation results;

[0118] Continuously optimize and adjust the multiple screening breakage control strategies according to the multiple adjustment evaluation results, and output the multiple screening breakage control adjustment strategies.

[0119] Furthermore, the intelligent control module 17 is further configured to perform the following steps:

[0120] Start the target screening breakage device and perform real-time operation monitoring on the target screening breakage device to generate a screening breakage operation control file;

[0121] Based on the screening breakage operation control file, perform operation status identification according to the operation nodes to generate multiple operation status identification results;

[0122] Extract the abnormal operation status identification data set according to the multiple operation status identification results;

[0123] Trace the abnormality of the corresponding operation nodes according to the abnormal operation status identification data set to generate an abnormal data source;

[0124] Perform automatic abnormal correction based on the abnormal data source to obtain an abnormal correction result, and perform operation deviation adjustment on the started target screening breakage device according to the abnormal correction result.

[0125] Through the foregoing detailed description of an online adaptive screening breakage control method for dry ash caking in this specification, those skilled in the art can clearly know an online adaptive screening breakage control method and system in this embodiment. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description in the method section.

[0126] The foregoing description of the disclosed embodiments enables those skilled in the art to practice or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An online adaptive broken sieve control method for dry ash caking, characterized in that, The method is applied to an online adaptive sieving-breaking control system for dry ash caking. The online adaptive sieving-breaking control system for dry ash caking is communicatively connected to a feedback adjustment unit and a remote terminal. The method includes: Deploy a plurality of sensors based on the dry ash conveying path, and monitor the target dry ash in real time through the plurality of sensors to obtain the real-time flow state information of the target dry ash; Construct a caking recognition model based on the real-time flow state information; Activate the caking recognition model to perform caking recognition on the target dry ash, and obtain a caking recognition result; Perform caking classification identification based on the caking recognition result, and formulate a plurality of sieving-breaking control strategies according to the caking classification identification result; Obtain an optimization adjustment information set through the feedback adjustment unit, and continuously optimize and adjust the plurality of sieving-breaking control strategies according to the optimization adjustment information set to obtain a plurality of sieving-breaking control adjustment strategies; Extract the real-time dry ash conveying state information and the real-time dry ash caking state information, execute the plurality of sieving-breaking control adjustment strategies, and obtain strategy execution configuration parameters; Connect to the remote terminal, send an execution instruction through the remote terminal, and based on the execution instruction, start the target sieving-breaking device for online sieving-breaking intelligent control according to the strategy execution configuration parameters.

2. The method according to claim 1, characterized in that, Deploy a plurality of sensors based on the dry ash conveying path, and monitor the target dry ash in real time through the plurality of sensors to obtain the real-time flow state information of the target dry ash. The method includes: Retrieve the historical conveying data record file of the target dry ash, and obtain a plurality of position information and a plurality of flow state information of the target dry ash in the dry ash conveying path. Among them, the plurality of position information and the plurality of flow state information are in a corresponding relationship; Determine a plurality of key positions of the dry ash conveying path based on the plurality of position information and the plurality of flow state information; Deploy the plurality of sensors according to the plurality of key positions; Monitor the real-time flow state of the plurality of key positions in the dry ash conveying path according to the plurality of sensors, and extract effective flow characteristics according to the monitoring results; Form the comprehensive flow state information of the dry ash conveying path based on the effective flow characteristics, and output the real-time flow state information.

3. The method according to claim 1, characterized in that, Construct a caking recognition model based on the real-time flow state information. The method includes: Construct a caking information library based on historical caking data, perform feature retrieval on the caking information library, and obtain a caking feature set; Traverse the real-time flow state information to match the caking feature set, and generate a caking feature vector set; Determine a data training feature set, a data supervision feature set, and a data verification feature set based on the caking feature vector set, use a deep learning algorithm to perform data training on the data training feature set and the data supervision feature set, and perform iterative testing on the training result based on the data verification feature set until convergence to obtain the caking recognition model.

4. The method according to claim 3, wherein Activate the caking recognition model to perform caking recognition on the target dry ash, and obtain a caking recognition result. The method includes: Activate the agglomeration recognition model to determine whether the target dry ash is less than a preset flow threshold within the dry ash conveying path. If so, it is regarded as an agglomeration phenomenon, traverse the agglomeration feature vector set for data transformation, and determine the deep agglomeration features; Analyze the change trend of the flow state of the target dry ash based on the deep agglomeration features; Perform agglomeration state recognition according to the change trend, and output the agglomeration recognition result based on the agglomeration state.

5. The method according to claim 1, wherein, Obtain an optimization adjustment information set through the feedback adjustment unit. The method includes: Formulate a feedback target based on the optimization direction, and establish the feedback adjustment unit according to the feedback target; Evaluate the execution effects of the multiple screen-breaking control strategies through the feedback adjustment unit to obtain multiple strategy execution effects; Judge whether the multiple strategy execution effects meet the preset target. If not, generate a feedback result; Determine the optimization adjustment parameters, optimization adjustment direction, and optimization adjustment intensity according to the feedback result and the multiple strategy execution effects; Add the optimization adjustment parameters, the optimization adjustment direction, and the optimization adjustment intensity to the optimization adjustment information set.

6. The method according to claim 5, wherein Continuously optimize and adjust the multiple screen-breaking control strategies according to the optimization adjustment information set to obtain multiple screen-breaking control adjustment strategies. The method includes: For each screen-breaking control strategy in the multiple screen-breaking control strategies, perform optimization and adjustment in sequence according to the optimization adjustment parameters, the optimization adjustment direction, and the optimization adjustment intensity in the optimization adjustment information set to obtain multiple optimization adjustment results; Perform screen-breaking simulation operation according to the multiple optimization adjustment results, and evaluate the adjustment effect according to the evaluation index based on the operation result to generate multiple adjustment evaluation results; Continuously optimize and adjust the multiple screen-breaking control strategies according to the multiple adjustment evaluation results, and output the multiple screen-breaking control adjustment strategies.

7. The method according to claim 1, characterized in that, Start the intelligent control of online screen-breaking for the target screen-breaking device. The method further includes: Start the target screen-breaking device and perform real-time operation monitoring on the target screen-breaking device to generate a screen-breaking operation control file; Perform operation state identification according to the screen-breaking operation control file according to the operation nodes to generate multiple operation state identification results; Extract the abnormal operation state identification data set according to the multiple operation state identification results; Trace the abnormality of the corresponding operation node according to the abnormal operation state identification data set to generate an abnormal data source; Perform automatic abnormal correction based on the abnormal data source to obtain an abnormal correction result, and perform operation deviation adjustment on the started target screen-breaking device according to the abnormal correction result.

8. An on-line adaptive sieving-breaking control system for dry ash caking, characterized in that The system is communicatively connected to the feedback adjustment unit and the remote terminal, and is used to implement an online adaptive screen-breaking control method for dry ash agglomeration according to any one of claims 1-7. The system includes: A real-time monitoring module, which is used to arrange multiple sensors based on the dry ash conveying path, and perform real-time monitoring on the target dry ash through the multiple sensors to obtain the real-time flow state information of the target dry ash; An identification model construction module, which is used to construct an agglomeration recognition model based on the real-time flow state information; An identification result acquisition module, which is used to activate the agglomeration recognition model to perform agglomeration recognition on the target dry ash and obtain an agglomeration recognition result; The broken sieve control strategy formulation module is used to conduct agglomerate classification identification based on the agglomerate identification result, and formulate multiple broken sieve control strategies according to the agglomerate classification identification result; The broken sieve control adjustment module is used to obtain the optimization adjustment information set through the feedback adjustment unit, and continuously optimize and adjust the multiple broken sieve control strategies according to the optimization adjustment information set to obtain multiple broken sieve control adjustment strategies; The configuration parameter acquisition module is used to extract the real-time dry ash conveying status information and the real-time dry ash agglomerate status information, execute the multiple broken sieve control adjustment strategies, and obtain the strategy execution configuration parameters; The intelligent control module is used to connect to the remote terminal, send an execution instruction through the remote terminal, and based on the execution instruction, start the target broken sieve device for online broken sieve intelligent control according to the strategy execution configuration parameters.

Images