Grain storage industry monitoring and early warning method and device based on virtual-real fusion

Abstract

The invention relates to the technical field of monitoring and early warning, and provides a grain storage industry monitoring and early warning method and device based on virtual-real fusion. The method comprises the following steps: constructing a three-dimensional virtual space according to an actual granary structure to obtain a virtual model; according to the collected real-time monitoring data, performing visual display at a corresponding position in the virtual model; performing mildew risk prediction and equipment fault risk prediction according to the real-time monitoring data to obtain a prediction result; and triggering a corresponding early warning mechanism according to the prediction result, and presenting the risk position distribution corresponding to the prediction result based on the virtual model, thereby solving the problems that in the prior art, data isolation cannot realize effective integration, the visualization degree is low, data resources cannot be effectively borne, accurate early warning is difficult to realize through data association analysis, and the risk is poor. And a global and visual management and control mode cannot be formed.

Description

Technical Field

[0001] This invention relates to the field of monitoring and early warning technology, and in particular to a monitoring and early warning method and device for the grain storage industry based on virtual-real fusion. Background Technology

[0002] Currently, the grain storage industry typically includes facilities such as grain silos, conveying equipment, ventilation equipment, temperature control equipment, and grain condition monitoring equipment. Different grain silos, depending on the type of grain stored (such as soybeans and corn) and the storage period, utilize various equipment to set up different monitoring points, such as grain pile temperature, silo humidity, gas concentration, and equipment operating status. During grain storage, grain is susceptible to factors such as temperature and humidity, pests, and mold, and equipment malfunctions can lead to grain loss or safety accidents.

[0003] To address the aforementioned issues, although some companies in the industry have deployed grain condition monitoring systems and equipment monitoring systems, there are still problems such as data isolation between systems and lack of effective integration, low visualization, and inability to effectively support data resources. This makes it difficult to achieve accurate early warning through data correlation analysis and to form a global and visualized management and control model.

[0004] Therefore, overcoming the shortcomings of the existing technology is an urgent problem to be solved in this technical field. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a monitoring and early warning method and device for the grain storage industry based on virtual-real fusion.

[0006] The present invention adopts the following technical solution: Firstly, this invention provides a monitoring and early warning method for the grain storage industry based on virtual-real fusion. A three-dimensional virtual space is constructed according to the actual grain warehouse structure to obtain a virtual model; based on the collected real-time monitoring data, the corresponding positions in the virtual model are visualized. Based on the real-time monitoring data, mold risk prediction and equipment failure risk prediction are performed to obtain prediction results; The prediction results trigger the corresponding early warning mechanism, and the risk location distribution corresponding to the prediction results is presented based on the virtual model.

[0007] Furthermore, the step of visualizing the corresponding location in the virtual model based on the collected real-time monitoring data includes: Extract key influencing parameters from the real-time monitoring data; Based on the aforementioned key impact parameters, a multi-dimensional threshold judgment logic is set to obtain real-time risk signals; The real-time risk signal is then bound to the corresponding grain pile partition location in the virtual model; The real-time risk signals and corresponding key impact parameters are displayed at the bound grain pile zoning locations for visualization purposes.

[0008] Furthermore, the step of setting a multi-dimensional threshold judgment logic based on the key impact parameters to obtain a real-time risk signal includes: When predicting the risk of mold growth, the system calls the preset variety-specific thresholds in the database according to the storage characteristics of each grain variety. When the key influencing parameters of the preset number of dimensions simultaneously exceed the corresponding variety-specific thresholds, a real-time risk signal is generated.

[0009] Furthermore, the step of setting a multi-dimensional threshold judgment logic based on the key impact parameters to obtain real-time risk signals includes: When predicting equipment failure risks, anomaly modes are set based on multiple key influencing parameters, and a risk mapping relationship between the anomaly modes and equipment failure types is established to set multi-dimensional threshold judgment logic. According to the risk mapping relationship, a real-time risk signal is generated when the key impact parameters meet the criteria of the abnormal pattern.

[0010] Furthermore, the prediction results obtained by performing mold risk prediction and equipment failure risk prediction based on the real-time monitoring data include: Historical data is used as training samples to train the original model to enable risk prediction capabilities, thus obtaining a prediction model. The prediction model was used to predict the risk of mold growth and the risk of equipment failure for the key influencing parameters, and the prediction results were obtained.

[0011] Furthermore, the original model includes a first original model and a second original model, the prediction model includes a mold prediction model and a fault prediction model, and the historical data includes historical mold data and historical equipment fault data; The process of using historical data as training samples to train the original model to have risk prediction capabilities and obtain a prediction model includes: When performing mold risk prediction, the first original model uses the historical mold data to learn the mapping relationship between the parameter change trends of each zone of the grain pile and the risk probability, and obtains the mold prediction model. When performing equipment failure risk prediction, the second original model extracts local abnormal features from the historical equipment failure data, learns the long-term change trends of parameters for various equipment types, and obtains a failure prediction model.

[0012] Furthermore, the step of using historical data as training samples to train the original model to have risk prediction capabilities and obtain a prediction model also includes: Using the misjudgment cases from the historical data, the prediction model is iteratively trained at preset intervals to optimize the prediction model.

[0013] Furthermore, the early warning mechanism includes a major risk early warning mechanism, a relatively large risk early warning mechanism, and a general risk early warning mechanism; The method for triggering the corresponding early warning mechanism based on the prediction result includes: If a mold risk prediction is performed, a major risk warning mechanism is triggered when the predicted risk probability is greater than the risk threshold; or, a general risk warning mechanism is triggered when the predicted trend of any parameter will exceed the upper limit of the safety threshold after a preset period. If equipment failure risk prediction is performed, a major risk warning mechanism is triggered when the prediction result is that the failure probability is greater than or equal to the probability threshold, and / or when the prediction result is that the long-term change trend of any equipment parameter is abnormal but does not affect operation.

[0014] Secondly, the present invention also provides a monitoring and early warning device for the grain storage industry based on virtual-real fusion, used to implement the monitoring and early warning method for the grain storage industry based on virtual-real fusion described in the first aspect, the device comprising: At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the processor for performing the grain storage industry monitoring and early warning method based on virtual-real fusion as described in the first aspect.

[0015] Thirdly, the present invention also provides a non-volatile computer storage medium storing computer-executable instructions, which are executed by one or more processors to perform the monitoring and early warning method for the grain storage industry based on virtual-real fusion described in the first aspect.

[0016] This invention utilizes a virtual model constructed based on the actual structure of a grain silo, combined with real-time monitoring data for visualization. This enables a mapping between grain storage status, equipment operation, and the physical space of the grain silo. It centrally analyzes and displays data from various individual points collected in real time, effectively integrating and timely analyzing fragmented data. The high degree of visualization effectively utilizes data resources. Furthermore, it predicts mold and equipment failure risks through correlation analysis, triggering corresponding early warning mechanisms based on the prediction results. The virtual model presents the distribution of these risks, achieving precise early warning and spatial visualization of warning information, thus facilitating comprehensive management of the grain silo. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0018] Figure 1 This is a flowchart illustrating a monitoring and early warning method for the grain storage industry based on virtual-real fusion, provided by an embodiment of the present invention. Figure 2 This is a schematic diagram of the overall process of a monitoring and early warning method for the grain storage industry based on virtual-real fusion provided by an embodiment of the present invention; Figure 3 This is a schematic diagram of the system structure of a monitoring and early warning cockpit system for the grain storage industry based on virtual-real fusion, provided in an embodiment of the present invention. Figure 4 This is a flowchart illustrating step 10 provided in an embodiment of the present invention; Figure 5 This is a flowchart illustrating step 102 provided in an embodiment of the present invention; Figure 6 This is a flowchart illustrating another step 102 provided in an embodiment of the present invention; Figure 7 This is a flowchart illustrating step 20 provided in an embodiment of the present invention; Figure 8 This is a flowchart illustrating step 201 provided in an embodiment of the present invention; Figure 9 This is a flowchart illustrating step 30 provided in an embodiment of the present invention; Figure 10 This is a schematic diagram of the architecture of a monitoring and early warning device for the grain storage industry based on virtual-real fusion, provided by an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] In the description of this invention, the terms "inner", "outer", "longitudinal", "lateral", "upper", "lower", "top", "bottom", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and do not require that this invention must be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0021] In this invention, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0022] In this application, unless otherwise expressly specified and limited, the term "connection" should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral part; it can be a direct connection or an indirect connection through an intermediate medium. Furthermore, the term "coupled" can refer to an electrical connection that enables signal transmission.

[0023] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0024] Example 1: To solve the above problems, such as Figure 1 As shown, this embodiment of the invention provides a monitoring and early warning method for the grain storage industry based on virtual-real fusion, including: Step 10: Construct a three-dimensional virtual space according to the actual grain warehouse structure to obtain a virtual model; based on the collected real-time monitoring data, visualize the corresponding position in the virtual model.

[0025] The specific types of real-time monitoring data are determined by those skilled in the art based on the specific application scenario. In one optional embodiment, the real-time monitoring data may include grain storage status and equipment operation status, specifically: grain pile temperature, humidity inside the grain silo, gas concentration, grain moisture content, equipment operating parameters, environmental data, and grain entry and exit information. Among these, equipment operating parameters may include: start / stop signals, ash hopper temperature, airlock ash unloading, inlet / outlet air pressure difference, conveyor status, current, and voltage, etc.; environmental data may include outdoor temperature and humidity, etc.; and grain entry and exit information may include quantity, time, and variety, etc.

[0026] like Figure 2 As shown, on the one hand, various monitoring devices are installed at multiple locations in the actual grain warehouse according to the actual monitoring needs of existing technologies, and data is acquired from these devices. After preprocessing, corresponding real-time monitoring data is obtained. On the other hand, in order to integrate these real-time monitoring data and achieve effective analysis and visualization, a virtual model is built. Among them, a three-dimensional virtual space is constructed based on the actual grain warehouse structure using digital twin technology. The specific method is selected by those skilled in the art according to the specific application scenario and is not limited here.

[0027] like Figure 3As shown in the figure, this embodiment of the invention also provides a monitoring and early warning cockpit system for the grain storage industry based on virtual-real fusion. Its system structure consists of the following modules: grain storage data acquisition module 1; data acquisition and preprocessing module 2; virtual modeling module 3; risk monitoring and analysis module 4; early warning triggering and level classification module 5; cockpit visualization module 6; and emergency response support module 7. The following description, in conjunction with the system structure, illustrates the monitoring and early warning system for the grain storage industry based on virtual-real fusion of this invention.

[0028] In one embodiment, the grain warehouse data acquisition module 1 is an existing or newly built system in the industry, and the data within the system can be real-time monitoring data. The data acquisition and preprocessing module 2 is a data integration and processing functional module that communicates with the grain warehouse data acquisition module 1 via the Message Queuing Telemetry Transport (MQTT) protocol or the OPC Unified Architecture (OPCUCA) protocol. Figure 2 As shown, after data collection, edge computing nodes are used to clean and standardize the collected raw data (i.e., data preprocessing). The processed data is then transmitted to the system database via a Kafka message queue. In one embodiment, the raw data can be cleaned by removing noise and filling in missing values.

[0029] The virtual modeling module 3 is used to construct a three-dimensional virtual model of the grain silo and achieve virtual-real mapping. Following step 10 above, a virtual space consistent with the actual grain silo structure is constructed using digital twin technology. Simultaneously, the real-time monitoring data output by the data acquisition and preprocessing module 2 is bound to the corresponding location in the virtual model, achieving real-time synchronous updates between the physical entity's state and the virtual model. For example, in one embodiment, when the real-time monitoring data is the grain pile temperature, the real-time monitoring data of the grain pile temperature is displayed and updated in real-time at the corresponding location in the virtual model.

[0030] This invention achieves spatial visualization of grain storage status through digital twin technology. Managers can intuitively grasp the grain condition, equipment and environmental status of the entire grain warehouse in the visualized scene of the virtual model. The location of risks and the scope of impact are clear at a glance, solving the problems of data fragmentation and positioning difficulties in traditional systems.

[0031] Step 20: Based on the real-time monitoring data, predict the risk of mold growth and the risk of equipment failure, and obtain the prediction results.

[0032] In one embodiment, the risk monitoring and analysis module 4 constructs a two-tiered risk analysis system of "real-time identification + trend prediction" based on standardized real-time monitoring data (i.e., grain pile temperature, grain silo humidity, gas concentration, grain moisture content, and equipment operating parameters, etc.) output by the data acquisition and preprocessing module 2 and historical data stored in the system database (i.e., historical risk event records, parameter change curves, handling result feedback, and risk threshold standards corresponding to grain varieties, etc.). This system focuses on predicting two core risks: mold risk and equipment failure risk. A specific example of predicting mold risk and equipment failure risk will be provided below.

[0033] Step 30: Trigger the corresponding early warning mechanism based on the prediction results, and present the risk location distribution corresponding to the prediction results based on the virtual model.

[0034] In one embodiment, the early warning triggering and classification module 5 triggers the corresponding early warning mechanism based on the prediction results output by the risk monitoring and analysis module 4 and in conjunction with preset risk assessment rules. The early warning levels are divided into three levels: Level 1 (major risk, such as localized high temperature in a grain pile that continues to rise), Level 2 (relatively high risk, such as abnormal equipment parameters that do not affect operation), and Level 3 (general risk, such as potential risks caused by environmental factors). A specific example of triggering the corresponding early warning mechanism based on the prediction results will be provided below. For example, when the prediction result indicates that the temperature of a conveyor bearing at a certain location in a physical grain silo is abnormal but does not affect operation, the display at that conveyor bearing will show "Temperature abnormal but does not affect operation."

[0035] This invention relates to a grain storage monitoring and early warning mode based on virtual-real fusion technology. By constructing a three-dimensional virtual model of the grain warehouse and binding real-time monitoring data with the virtual space, it achieves spatial visualization and precise positioning of risks.

[0036] In existing technologies, monitoring and early warning in the grain storage industry largely rely on traditional single-point data monitoring and manual inspections. This invention, however, uses a virtual model constructed based on the actual grain storage structure, combined with real-time monitoring data for visualization. This achieves a correlation mapping between grain storage status, equipment operation, and the physical space of the grain storage facility. It centrally analyzes and displays various real-time collected single-point data, effectively integrating and timely analyzing fragmented data with a high degree of visualization, thus effectively utilizing data resources. Furthermore, it predicts mold and equipment failure risks through correlation analysis and triggers corresponding early warning mechanisms based on the prediction results. The virtual model presents the distribution of corresponding risk locations, achieving precise early warning and spatial visualization of early warning information, facilitating comprehensive management of the grain storage facility.

[0037] In one embodiment, such as Figure 4 As shown, in step 10, the visualization display at the corresponding location in the virtual model based on the collected real-time monitoring data includes: Step 101: Extract key influencing parameters from the real-time monitoring data.

[0038] To address the risk of grain mold, the first step is to identify the risk in real time. In one embodiment, key influencing parameters are extracted from real-time monitoring data: grain pile temperature, grain warehouse relative humidity, grain moisture content, and gas concentration inside the grain warehouse. These key influencing parameters are used as core monitoring dimensions.

[0039] Regarding equipment failure risks, the first step is to identify risks in real time. In one embodiment, key influencing parameters are extracted from real-time monitoring data, such as stop signal, ash hopper temperature, airlock ash discharge, inlet and outlet air pressure difference, conveyor status, current and voltage, etc., to use these key influencing parameters as core monitoring dimensions.

[0040] Step 102: Set up multi-dimensional threshold judgment logic based on the key impact parameters to obtain real-time risk signals.

[0041] Thresholds are set for key influencing parameters of core monitoring dimensions, and the relationship between multiple key influencing parameters and their thresholds is combined to determine the risk of grain mold or equipment failure. When the risk of grain mold and / or equipment failure is identified, a real-time risk signal is generated.

[0042] Step 103: Bind the real-time risk signal to the corresponding grain pile partition location in the virtual model.

[0043] The virtual model corresponds to an actual grain pile structure with multiple partitions, each with its own partition location.

[0044] Step 104: Display the real-time risk signal and corresponding key impact parameters at the bound grain pile partition location for visualization.

[0045] In one embodiment, the real-time risk signal is bound to the grain pile partition location in the virtual model modeling module 3 to clarify the specific physical area where the risk occurs, and the specific physical area where the risk occurs, the risk type and / or risk level are displayed in real time in the cockpit visualization display module 6.

[0046] Regarding the risk of grain mold, in one embodiment, such as Figure 5 As shown, step 102 includes: Step 1021a: When performing mold risk prediction, the preset variety-specific thresholds in the system database are called according to the storage characteristics of each grain variety.

[0047] Based on the storage characteristics of different grain varieties (such as soybeans and corn), the system invokes preset variety-specific thresholds from the system database. These thresholds are selected by those skilled in the art according to specific usage scenarios and stored in the system database for use whenever key influencing parameters are updated. For example, the specific thresholds for the safe grain pile temperature for corn storage include the following aspects: safe grain pile temperature, safe moisture content, safe humidity in the grain silo, and carbon dioxide concentration; that is, a safe grain pile temperature of less than or equal to 25 degrees Celsius, a safe moisture content of less than or equal to 14%, a safe humidity in the grain silo of less than or equal to 70%, and a carbon dioxide concentration of less than or equal to 0.5%.

[0048] Step 1022a: When the key influencing parameters of the preset dimension number simultaneously exceed the corresponding variety-specific threshold, a real-time risk signal is generated.

[0049] The preset number of dimensions is selected by those skilled in the art based on the specific application scenario. In one embodiment, the preset number of dimensions can be 2. For example, when any two or more dimension parameters in the real-time data simultaneously exceed the corresponding threshold (e.g., grain pile temperature reaches 28 degrees Celsius, moisture content is 15.2%, carbon dioxide concentration is 0.8%), or when a single dimension parameter exceeds the threshold by more than 15% (e.g., the local temperature of the grain pile suddenly rises to 32 degrees Celsius), a real-time risk signal is generated to preliminarily determine it as a "real-time mold risk signal".

[0050] Regarding the risk of equipment failure, in one embodiment, such as Figure 6 As shown, step 102 includes: Step 1021b: When performing equipment failure risk prediction, anomaly modes are set based on multiple key influencing parameters, and a risk mapping relationship between the anomaly modes and equipment failure types is established to set multi-dimensional threshold judgment logic.

[0051] Based on the equipment's operating mechanism, a mapping relationship between abnormal modes and fault types is established to avoid misjudgment due to fluctuations in a single parameter: for example, if the fan current is greater than 120% of the rated value for 30 seconds and the air pressure difference is greater than 20% (i.e., indicating that the duct is blocked and the load is increased), it is judged as "overload fault risk". Other equipment risk judgment criteria can be derived from historical experience, expert discussions, etc.

[0052] Step 1022b: According to the risk mapping relationship, when the key impact parameters meet the criteria of the abnormal mode, generate a real-time risk signal.

[0053] Existing monitoring and early warning technologies are mostly limited to alarms triggered by a single parameter exceeding a standard. This invention, however, comprehensively assesses multiple key influencing parameters to generate timely and preliminary real-time risk signals. By visualizing these signals, a virtual model is used to display the preliminary risk analysis, enabling staff to make timely judgments and responses. Furthermore, combining these key influencing parameters with subsequent risk trend prediction allows for more comprehensive risk monitoring.

[0054] When predicting the risk trend of grain mold, in one embodiment, such as Figure 7 As shown, step 20 includes: Step 201: Use historical data as training samples to train the original model to enable it to predict risks, thus obtaining a prediction model.

[0055] The original model includes a first original model and a second original model; the prediction model includes a mold prediction model and a fault prediction model; and the historical data includes historical mold data and historical equipment fault data. The first original model is used for mold risk prediction, and the second original model is used for equipment fault risk prediction. The specific network structures of the first and second original models are determined by those skilled in the art based on the specific application scenario. In one embodiment, the first original model can use a Long Short-Term Memory (LSTM) network model, and the second original model can use a Long Short-Term Memory Convolutional Neural Network (CNN) model.

[0056] Step 202: Use the prediction model to predict the risk of mold growth and the risk of equipment failure of the key influencing parameters, and obtain the prediction results.

[0057] The process of using key influencing parameters for prediction is similar to the process of training historical data, and will not be repeated here.

[0058] Specifically, in one embodiment, such as Figure 8 As shown, step 201 includes: Step 2011: When performing mold risk prediction, the first original model uses the historical mold data to learn the mapping relationship between the parameter change trends of each zone of the grain pile and the risk probability, and obtains the mold prediction model.

[0059] Historical mold data can utilize key influencing parameters within a specific timeframe. For instance, using historical data from the system database as training samples, including time-series variation curves of "temperature, humidity, moisture, and gas concentration" for different grain varieties and storage periods over the past five years, the occurrence time points of historical mold events, abnormal parameter characteristics in the early stages of mold, and data on treatment measures and their effects, the network parameters of the first original model are optimized using gradient descent. This enables the first original model to map "parameter change trends to mold risk probabilities," resulting in a trained mold prediction model. The specific training methods for the first and second original models are determined by those skilled in the art based on the specific application scenario and are not limited here.

[0060] Step 2012: When performing equipment failure risk prediction, the second original model extracts local abnormal features from the historical equipment failure data, learns the long-term change trend of parameters for various equipment types, and obtains a failure prediction model.

[0061] For example, historical equipment failure data includes: time-series curves of operating parameters (e.g., current fluctuation curves), failure occurrence time, failure type (e.g., motor overload, mechanical blockage), duration of abnormal parameters before failure, and maintenance records for different equipment types. A second original model extracts local anomaly features (e.g., sudden current peaks) from the time-series data of historical equipment failures, and a gated recurrent unit (GRU) captures long-term parameter trends (e.g., a slow, monthly increase in fan current). These features are then combined to train the second original model, outputting the equipment failure probability and duration prediction as the prediction result.

[0062] In one embodiment, step 201 further includes: using misjudged cases from the historical data to iteratively train the prediction model at preset intervals to optimize the prediction model. The preset interval is determined by those skilled in the art based on the specific application scenario; for example, by combining "misjudged cases" from historical data (e.g., parameters exceeding limits due to short-term environmental fluctuations but without mold growth), the mold prediction model can be iteratively trained periodically (once a month), incorporating the latest treatment results feedback data to gradually reduce prediction errors and improve the accuracy of trend prediction.

[0063] This invention introduces a risk prediction mechanism based on deep learning algorithms. It performs multi-dimensional risk identification and trend prediction based on real-time monitoring data and historical data, overcoming the lag of traditional threshold alarms and achieving early risk warning.

[0064] In one embodiment, the early warning mechanism includes a major risk early warning mechanism, a significant risk early warning mechanism, and a general risk early warning mechanism; such as Figure 9As shown, in step 30, triggering the corresponding early warning mechanism based on the prediction result includes: Step 301: If mold risk prediction is performed, a major risk warning mechanism is triggered when the predicted risk probability is greater than the risk threshold; or, a general risk warning mechanism is triggered when the predicted trend of any parameter will exceed the upper limit of the safety threshold after a preset period.

[0065] The risk threshold and the upper limit of the safety threshold shall be determined by those skilled in the art based on the specific use case.

[0066] For example, when performing mold risk prediction according to step 202, the preprocessed continuous 6-hour time series data of parameters (sampling frequency of 5 minutes / time) is acquired in real time and input into the trained mold prediction model. The model outputs the mold risk probability (0 to 100%) of each zone of the grain pile and the trend curve of key parameter changes in the next 24 hours. When the predicted risk probability is greater than or equal to 60%, or when a certain parameter (e.g., grain pile temperature) is predicted to continue to rise and exceed the upper limit of the safety threshold within 12 hours, it is determined as a "potential mold risk warning signal".

[0067] The early warning system is divided into three levels. Level 1, or major risk, is triggered when the predicted probability of risk exceeds a risk threshold (e.g., localized and sustained high temperatures in a grain pile). Level 3, or general risk, is triggered when the predicted trend of any parameter is expected to exceed the upper limit of the safety threshold after a preset time period. Specific response plans for the major and general risk early warning mechanisms should be determined by those skilled in the art based on the specific application scenario and are not limited here.

[0068] Step 302: If equipment failure risk prediction is performed, a major risk warning mechanism is triggered when the prediction result is that the failure probability is greater than or equal to the probability threshold, and / or when the prediction result is that the long-term change trend of any equipment parameter is abnormal but does not affect operation.

[0069] For example, when performing equipment failure risk prediction according to step 202, high-frequency operating data of the equipment is collected continuously for 2 hours in real time (sampling frequency is once per minute). After being input into the failure prediction model, the model outputs the probability of equipment failure in the next 48 hours (0 to 100%). When the predicted failure probability is greater than or equal to 50%, or when a certain parameter (such as the temperature of the conveyor bearing) is predicted to exceed the safety threshold within 8 hours, it is determined as a "potential equipment failure risk warning signal" and triggers a secondary warning, i.e. a major risk warning mechanism.

[0070] In existing technologies, managers struggle to quickly locate risks and assess their impact, and they cannot perform intelligent predictions by combining historical data and environmental factors, resulting in low risk response efficiency and poor control effectiveness. This invention combines multi-dimensional data analysis and deep learning prediction to achieve early risk identification and accurate warnings. Compared to traditional threshold alarms, it can detect potential risks earlier, shifting the focus of safety risk control upstream.

[0071] like Figure 3 As shown, while triggering the early warning mechanism based on the prediction results, the cockpit visualization module 6 integrates the 3D virtual model with the early warning information and monitoring data, intuitively presenting the risk location and distribution through a 3D virtual scene. It also displays key parameter changes in the form of dashboards, trend charts, and heat maps, supporting detailed information queries for risk points (e.g., historical data, associated equipment, and impact range). The emergency response assistance module 7 automatically matches the preset response plan library based on the early warning level and risk type sent by the early warning triggering and level classification module 5, generating a response plan that includes the risk location, recommended measures, responsible personnel, and operational steps. This plan can be pushed to relevant personnel via system messages, mobile applications, and SMS, while simultaneously recording the response process and results for data feedback.

[0072] like Figure 2 As shown, this embodiment of the invention integrates interactive cockpit visualization and emergency response assistance functions. It displays risk distribution through a three-dimensional scene, automatically matches response plans, pushes instructions through multiple channels, and records feedback results, forming a closed-loop management of "monitoring-early warning-response-feedback".

[0073] In one embodiment, based on such Figure 3 The grain storage industry monitoring and early warning dashboard system shown in the diagram has a grain storage data acquisition module 1 that collects real-time data on grain pile temperature, grain storage humidity, and ventilation fan current. After being cleaned and processed by the data acquisition and preprocessing module 2, the data is transmitted to the virtual modeling module 3. This module binds temperature data to the partitions of the virtual grain pile model, humidity data to the grain storage space, and current data to the virtual ventilation fan model, achieving synchronization between the virtual model and the physical entity. The risk monitoring and analysis module 4, based on the data transmitted from the data acquisition and preprocessing module 2, detects that the temperature of a certain grain pile partition has been rising continuously for a long time and exceeding the threshold. Combined with historical data, it is judged to be at risk of mold growth, triggering the early warning triggering and level classification module 5 to issue a level one warning. The dashboard visualization module 6 marks the risk area in the 3D virtual model with a red heat map and pops up a details window containing temperature trends and the status of related ventilation equipment. The emergency response assistance module 7 pushes a response plan of "starting the ventilation equipment in the area and arranging manual spot checks" based on the triggered warning level and risk type. Relevant personnel receive and execute the plan, and the system records the data changes after the response for data feedback.

[0074] In another embodiment, based on such Figure 3 The grain storage industry monitoring and early warning cockpit system shown has a grain storage data acquisition module 1 that collects real-time data on grain pile temperature, grain storage humidity, and ventilation fan current. After being cleaned and processed by the data acquisition and preprocessing module 2, the data is transmitted to the virtual modeling module 3. This module binds the temperature data to the partitions of the virtual grain pile model, the humidity data to the grain storage space, and the current data to the virtual ventilation fan model, achieving synchronization between the virtual model and the physical entity. The risk monitoring and analysis module 4 detects abnormal fluctuations in the current of a certain conveyor based on the data transmitted from the data acquisition and preprocessing module 2. However, the data returns to normal after 30 seconds of abnormal fluctuation and does not reappear for a long time. Since the number of triggers within a certain period is lower than the system's preset trigger frequency, it is determined that no early warning will be triggered, and the process ends, entering continuous monitoring mode.

[0075] The monitoring and early warning method for the grain storage industry based on virtual-real fusion, as described in this invention, improves management efficiency through a dashboard visualization display and emergency response assistance functions. Managers can quickly formulate response strategies through virtual models, shortening emergency response time. At the same time, the historical data traceability and optimization mechanism helps to continuously improve the level of grain storage safety management.

[0076] Example 2: This embodiment is a preferred embodiment of the present invention.

[0077] Embodiment 1 of the present invention monitors and warns of grain storage based on grain pile zoning. However, regional dynamic risk assessment is limited to a single type of indicator and cannot effectively integrate multi-source data for safety management, resulting in an inability to truly reflect the regional risk level and a limited level of safety management.

[0078] Furthermore, existing dynamic risk assessments suffer from two main problems: First, the indicators used in dynamic risk assessments are too simplistic, focusing only on monitoring physical quantities such as grain pile conditions and equipment operation, while neglecting the impact of other management indicators (e.g., education and training, hazard identification, safety investment, and emergency management) on regional risks. Second, even when some existing technologies attempt to incorporate multi-dimensional indicators for risk assessment, the allocation of indicator weights still relies primarily on qualitative judgments based on expert experience, failing to establish an objective weighting mechanism based on data characteristics or actual industry risk patterns. This approach is susceptible to the influence of experts' subjective perceptions, industry experience preferences, and differences in professional backgrounds, leading to a mismatch between weight settings and the actual degree of risk impact, thus causing deviations in risk assessment results and reducing the accuracy and reliability of the assessment conclusions.

[0079] To address the aforementioned issues, this embodiment also provides a dynamic risk assessment method, including: Step 401: Collect monitoring data from multiple grain pile zones to obtain multi-source indicator data; wherein, the multi-source indicator data includes grain condition indicators, equipment indicators, and management indicators.

[0080] The specific types of indicators included in the monitoring data and multi-source indicator data are determined by those skilled in the art based on the specific application scenario. The monitoring data is the raw data collected, and the multi-source indicator data is extracted after processing the raw data. In one embodiment, the multi-source indicator data may include grain condition indicators, equipment indicators, and management indicators. Among them, grain condition indicators include the average temperature of the grain pile and the average humidity of the grain warehouse; equipment indicators include the equipment failure rate; and management indicators include the rate of hazard investigation and rectification.

[0081] Based on Example 1, such as Figure 3 As shown, the virtual-real fusion-based grain storage industry monitoring and early warning cockpit system of this embodiment also includes a dynamic risk assessment module 8; the following describes the virtual-real fusion-based grain storage industry monitoring and early warning system of this embodiment in conjunction with the system structure.

[0082] In a specific example, the grain warehouse data acquisition module 1 collects real-time monitoring data from three areas, A, B, and C, including grain condition indicators (e.g., average grain pile temperature, average grain warehouse humidity), equipment indicators (e.g., equipment failure rate), and management indicators (e.g., hazard identification and rectification rate). The data acquisition and preprocessing module 2 cleans and standardizes the raw data before transmitting it to the dynamic risk assessment module 8. The dynamic risk assessment module 8 extracts multi-source indicator data for the current period, as shown in Table 1 below. Table 1 Multi-source indicator data table

[0083] Step 402: Standardize the multi-source indicator data to obtain the standard score of each indicator.

[0084] In one embodiment, the indicators in the multi-source indicator data are divided into positive and negative indicators: grain pile temperature, grain silo humidity, and equipment failure rate are negative indicators; the hidden danger investigation and rectification rate is a positive indicator. The theoretical ranges for each indicator are set as follows: grain pile temperature [15, 35]℃, grain silo humidity [50, 90]%, equipment failure rate [0, 100]%, and hidden danger investigation and rectification rate [0, 100]%. Then, according to the theoretical range corresponding to each indicator, and considering whether the corresponding indicator is a positive or negative indicator, the standard score for each indicator is calculated.

[0085] The standard score for a positive indicator is expressed as follows: ; in, Indicates the standard score. This represents the actual value of the indicator in the multi-source indicator data. This represents the theoretical worst value of the indicator in multi-source indicator data. This represents the theoretical optimal value of the indicator in multi-source indicator data. This indicates the sequence number of the grain pile section. Indicates the index number.

[0086] The standard score for the negative indicator is expressed as follows: ; in, Indicates the first The grain pile section is number one. The standard score for each indicator This represents the theoretical optimal value of the indicator in multi-source indicator data. This represents the actual value of the indicator in the multi-source indicator data. This represents the theoretical worst value of the indicator in multi-source indicator data.

[0087] Table 2 below shows a specific example of obtaining the standard scores for each indicator after standardized calculation: Table 2 Standard Sub-Table of Multi-Source Indicator Data

[0088] Step 403: Determine the weight of each indicator in the multi-source indicator data.

[0089] Calculate the mean of each indicator. The expression is as follows: ; in, This indicates the total number of grain pile sections.

[0090] Calculate the standard deviation of each indicator The expression is as follows: ; A specific example of the mean and standard deviation of each indicator is shown in Table 3 below: Table 3. Specific examples of mean and standard deviation

[0091] Calculate the coefficient of variation for each indicator. : ; in, This represents the standard deviation of the indicator. This represents the mean of the indicator.

[0092] Table 4. Examples of Coefficients of Variation for Each Indicator

[0093] Determine weights : ; in, Represents the coefficient of variation. This indicates the total number of indicators in the multi-source indicator data.

[0094] Table 5 Weighting of each indicator

[0095] Step 404: Determine the final risk score of each grain pile partition according to the standard score and weight of each indicator in the multi-source indicator data, so as to determine the risk level of each grain pile partition.

[0096] The final risk score is determined using the following formula: ; in, Indicates the first The weight of each indicator.

[0097] In a specific instance, the final risk scores for the dynamic risk assessment of areas A, B, and C are as follows: A: 100×(0.650×0.409+0.550×0.444+0.915×0.058+0.920×0.089)=64.5; B: 100×(0.450×0.409+0.375×0.444+0.877×0.058+0.850×0.089)=47.8; C: 100×(0.800×0.409+0.700×0.444+0.948×0.058+0.960×0.089)=77.8.

[0098] In one embodiment, the risk level of a grain pile area with a final risk score greater than the low-risk threshold can be determined as low-risk, the risk level of a grain pile area with a final risk score less than the low-risk threshold but greater than the high-risk threshold can be determined as medium-risk, and the risk level of a grain pile area with a final risk score less than the high-risk threshold can be determined as high-risk. The high-risk threshold and the low-risk threshold are determined by those skilled in the art according to the specific application scenario. In an optional embodiment, the high-risk threshold can be 30.0 and the low-risk threshold can be 70.0. In this case, area A and area B are medium-risk and are highlighted in yellow in the cockpit virtual model; area C is low-risk and is highlighted in green in the cockpit virtual model.

[0099] This embodiment of the dynamic risk assessment method combines a standardized processing method based on theoretical optimal and worst values ​​with an objective weighting method based on the data itself. By objectively assigning weights to indicators, it overcomes the subjectivity and instability of traditional expert experience methods, making the risk assessment results more accurate and reliable. Through multi-source data standardization processing and weighted fusion calculation, it achieves quantitative assessment and dynamic updating of regional risk levels, solving the problems of strong subjectivity in weights and single indicator dimensions in traditional risk assessment, thus improving the scientificity and accuracy of risk assessment. Simultaneously, it integrates multi-dimensional data such as grain conditions, equipment, and management to construct a comprehensive risk assessment system that can truly reflect the overall risk level of the region. The dynamic update mechanism based on real-time data allows the risk level to automatically adjust according to changes in actual conditions, providing managers with continuous and accurate risk decision support.

[0100] Example 3: like Figure 10 The diagram shown is an architectural schematic of a grain storage industry monitoring and early warning device based on virtual-real fusion according to an embodiment of the present invention. This embodiment of the grain storage industry monitoring and early warning device based on virtual-real fusion includes one or more processors 21 and a memory 22. Figure 10 Take a processor 21 as an example.

[0101] Processor 21 and memory 22 can be connected via a bus or other means. Figure 10 Taking the example of a connection between China and Israel via a bus.

[0102] The memory 22, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs and non-volatile computer-executable programs, such as the grain storage industry monitoring and early warning method based on virtual-real fusion in Embodiment 1. The processor 21 executes the grain storage industry monitoring and early warning method based on virtual-real fusion by running the non-volatile software programs and instructions stored in the memory 22.

[0103] Memory 22 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 22 may optionally include memory remotely located relative to processor 21, which can be connected to processor 21 via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0104] The program instructions / modules are stored in the memory 22. When executed by one or more processors 21, they execute the grain storage industry monitoring and early warning method based on virtual-real fusion in Embodiment 1 above, for example, executing each step of the grain storage industry monitoring and early warning method based on virtual-real fusion described above.

[0105] It is worth noting that the information interaction and execution process between the modules and units in the above-mentioned device and system are based on the same concept as the processing method embodiment of the present invention. For details, please refer to the description in the method embodiment of the present invention, and will not be repeated here.

[0106] Those skilled in the art will understand that all or part of the steps in the various methods of the embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc.

[0107] The above description is only a preferred embodiment of the present invention and is 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 within the protection scope of the present invention.

Claims

1. A monitoring and early warning method for the grain storage industry based on virtual-real fusion, characterized in that, include: A three-dimensional virtual space was constructed based on the actual grain warehouse structure to obtain a virtual model; Based on the collected real-time monitoring data, the corresponding locations in the virtual model are visualized. Based on the real-time monitoring data, mold risk prediction and equipment failure risk prediction are performed to obtain prediction results; The prediction results trigger the corresponding early warning mechanism, and the risk location distribution corresponding to the prediction results is presented based on the virtual model.

2. The monitoring and early warning method for the grain storage industry based on virtual-real fusion according to claim 1, characterized in that, The method includes: Extract key influencing parameters from the real-time monitoring data; Based on the aforementioned key impact parameters, a multi-dimensional threshold judgment logic is set to obtain real-time risk signals; The real-time risk signal is then bound to the corresponding grain pile partition location in the virtual model; The real-time risk signals and corresponding key impact parameters are displayed at the bound grain pile zoning locations for visualization purposes.

3. The monitoring and early warning method for the grain storage industry based on virtual-real fusion according to claim 2, characterized in that, The method includes: When predicting the risk of mold growth, the system calls the preset variety-specific thresholds in the database according to the storage characteristics of each grain variety. When the key influencing parameters of the preset number of dimensions simultaneously exceed the corresponding variety-specific thresholds, a real-time risk signal is generated.

4. The monitoring and early warning method for the grain storage industry based on virtual-real fusion according to claim 2, characterized in that, The method includes: When predicting equipment failure risks, anomaly modes are set based on multiple key influencing parameters, and a risk mapping relationship between the anomaly modes and equipment failure types is established to set multi-dimensional threshold judgment logic. According to the risk mapping relationship, a real-time risk signal is generated when the key impact parameters meet the criteria of the abnormal pattern.

5. The monitoring and early warning method for the grain storage industry based on virtual-real fusion according to claim 2, characterized in that, The method includes: Historical data is used as training samples to train the original model to enable risk prediction capabilities, thus obtaining a prediction model. The prediction model was used to predict the risk of mold growth and the risk of equipment failure for the key influencing parameters, and the prediction results were obtained.

6. The monitoring and early warning method for the grain storage industry based on virtual-real fusion according to claim 5, characterized in that, The original model includes a first original model and a second original model; the prediction model includes a mold prediction model and a fault prediction model; and the historical data includes historical mold data and historical equipment fault data. The method includes: When performing mold risk prediction, the first original model uses the historical mold data to learn the mapping relationship between the parameter change trends of each zone of the grain pile and the risk probability, and obtains the mold prediction model. When performing equipment failure risk prediction, the second original model extracts local abnormal features from the historical equipment failure data, learns the long-term change trends of parameters for various equipment types, and obtains a failure prediction model.

7. The monitoring and early warning method for the grain storage industry based on virtual-real fusion according to claim 6, characterized in that, The method further includes: Using the misjudgment cases from the historical data, the prediction model is iteratively trained at preset intervals to optimize the prediction model.

8. The monitoring and early warning method for the grain storage industry based on virtual-real fusion according to claim 6, characterized in that, The early warning mechanism includes a major risk early warning mechanism, a relatively large risk early warning mechanism, and a general risk early warning mechanism; The method includes: If a mold risk prediction is performed, a major risk warning mechanism is triggered when the predicted risk probability is greater than the risk threshold; or, a general risk warning mechanism is triggered when the predicted trend of any parameter will exceed the upper limit of the safety threshold after a preset period. If equipment failure risk prediction is performed, a major risk warning mechanism is triggered when the prediction result is that the failure probability is greater than or equal to the probability threshold, and / or when the prediction result is that the long-term change trend of any equipment parameter is abnormal but does not affect operation.

9. A non-volatile computer storage medium, characterized in that, The computer storage medium stores computer-executable instructions, which are executed by one or more processors to perform the monitoring and early warning method for the grain storage industry based on virtual-real fusion as described in any one of claims 1-8.

10. A monitoring and early warning device for the grain storage industry based on virtual-real fusion, characterized in that, include: At least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the processor for performing the grain storage industry monitoring and early warning method based on virtual-real fusion as described in any one of claims 1-8.

Images