A system construction method combining Internet of Things and virtual reality technology
The system construction method addresses data accuracy and model prediction issues by identifying errors, establishing a twin model, and optimizing it based on physical status, enhancing predictive accuracy and safety in IoT and virtual reality systems.
Patent Information
- Application Number
- CN202411327787.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-24
- Publication Date
- 2024-12-24
- Estimated Expiration
- 2044-09-24
AI Technical Summary
Existing IoT and virtual reality combined systems face challenges in ensuring data accuracy and model prediction accuracy due to errors and incomplete representation of device status behaviors, leading to misjudgments and inaccuracies in device status analysis.
A system construction method combining IoT and virtual reality technologies, involving error identification, feature extraction, twin model establishment, periodic monitoring, and optimization of the twin model based on physical status to enhance data accuracy and predictive capabilities.
Ensures accurate data collection and model building by identifying and correcting errors, optimizing the twin model based on actual equipment status, and improving predictive accuracy and safety of equipment operation.
Smart Images
Figure CN118861944B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of Internet of Things (IoT) twin technology, specifically a system construction method that combines IoT and virtual reality technologies. Background Technology
[0002] Digital twin technology, a field that has garnered significant attention in recent years, is gradually playing a crucial role within the edge computing framework. By establishing a virtual model corresponding to the actual equipment, it reflects and predicts the equipment's operational performance and failure risks in real time, providing strong support for equipment optimization and maintenance.
[0003] When using some existing IoT and virtual reality combined systems, the accuracy of the data obtained for device status analysis is difficult to guarantee, which may lead to misjudgments in the overall device status analysis. At the same time, when using models to predict device status, different device status behaviors cannot be fully reflected and predicted by the models, resulting in errors in model predictions. Summary of the Invention
[0004] To address the shortcomings of existing technologies, this invention provides a system construction method that combines Internet of Things (IoT) and virtual reality technologies. This method solves the problem of not being able to guarantee the accuracy of data and the existence of errors in model building when using data.
[0005] To achieve the above objectives, the present invention provides the following technical solution: a system construction method combining Internet of Things (IoT) and virtual reality technologies, specifically comprising the following steps:
[0006] Step 1: Perform error identification on the collected construction data to obtain the identification results, which include correct data and erroneous data. Then, determine the impact of the erroneous data in the identification results and generate the judgment results, which include whether there is an impact or not.
[0007] Step 2: Build a model based on the acquired construction data, perform feature analysis and extraction on the construction data, establish a corresponding twin model based on the extracted features, and then analyze the equipment based on the twin model to obtain analysis results, including whether there is a danger or not.
[0008] Step 3: Periodically monitor the obtained twin model, and combine it with the physical status of the equipment to determine whether the twin model needs to be optimized and adjusted. Optimize the twin model based on the corresponding abnormal data and generate optimization information.
[0009] Step 4: Generate different functional modules based on the corresponding construction features of the twin model. At the same time, perform comprehensive analysis and calculation of the management value of the functional module by combining the historical data of different functional modules, and adjust and analyze the functional module according to the changes in the management value.
[0010] As a further aspect of the present invention, the specific method for generating the judgment result in step one is as follows:
[0011] The system retrieves the build data and evaluates it. If errors are found in the build data, the corresponding error data is retrieved. If no errors are found in the build data, no action is taken.
[0012] The acquired erroneous data is analyzed to determine whether the erroneous data has an impact on the overall data. The specific determination method is as follows: the data type corresponding to the erroneous data is obtained and the analysis type is remembered. At the same time, the quantity ratio corresponding to the analysis type is calculated and compared with the preset value. When the quantity ratio is greater than the preset value, it indicates that the proportion of the same type of erroneous data is large, and an impact result is generated. Conversely, when the quantity ratio is less than the preset value, a secondary analysis signal is generated and processed.
[0013] As a further aspect of the present invention: the specific method for processing the secondary analysis signal in step one is as follows:
[0014] Retrieve historical data, and simultaneously retrieve usage records of similar data within the historical data, and calculate the corresponding usage percentage. Then compare the usage percentage with a threshold. If the usage percentage is greater than the threshold, an impact result is generated; otherwise, if the usage percentage is less than the threshold, an no-impact result is generated.
[0015] As a further aspect of the present invention, the specific method for obtaining the analysis results in step two is as follows:
[0016] All construction data is acquired, and the construction data is classified according to data type to obtain data classification information. Any type of construction data is acquired, analyzed, and the target object is remembered. Then, the features of the target object are analyzed and extracted, and the extracted features are selected to obtain construction features.
[0017] Next, a twin model is constructed based on the obtained construction features. At the same time, the potential hazards of the equipment are identified and judged based on the obtained twin model, the real-time features of the equipment are obtained, and the obtained real-time features are substituted into the twin model. The twin model analyzes and judges whether there are potential hazards in the current operation of the equipment, and generates analysis results, including whether there are hazards or not.
[0018] As a further aspect of the present invention: the specific method for determining whether the twin model needs optimization and adjustment in step three is as follows:
[0019] The system acquires the physical state of the device and the corresponding actual data. It then inputs the actual data into the twin model to obtain the prediction result. The prediction result is then compared with the physical state. If the prediction result is the same as the physical state, it indicates that there is no anomaly in the twin model, and continuous monitoring information is generated. Conversely, if the prediction result is different from the physical state, it indicates that there is an anomaly in the twin model, and adjustment and optimization signals are generated. The adjustment and optimization signals are then analyzed.
[0020] As a further aspect of the present invention: the specific method for analyzing the adjustment and optimization signal in step three is as follows:
[0021] The twin model is used to construct features that differ from the physical device state. These features are then adjusted based on the physical device state features. Historical data from the twin model is also obtained, and the number of optimizations in the historical data is denoted as i, where i = 1, 2, ..., j, and j represents the number of optimizations. The optimization time corresponding to optimization number i is denoted as Ti, and the optimization time interval is calculated. The mean of the optimization time interval is then calculated, and optimization information is generated based on the mean.
[0022] As a further aspect of the present invention: the specific method for comprehensively analyzing and calculating the management value of the functional modules by combining historical data from different functional modules in step four is as follows:
[0023] The construction features of the twin model are obtained, and corresponding functional modules are generated based on these features. These functional modules are labeled n, where n = 1, 2, ..., m, and m represents the number of functional modules. Next, the usage of functional module n is obtained, including the number of uses Cn, the amount of data generated Kn, and the amount of abnormal data Pn. These parameters are then substituted into the formula. The management value Qn corresponding to functional module n is calculated.
[0024] As a further aspect of the present invention: the specific method for analyzing and adjusting the functional modules based on changes in the management value in step four is as follows:
[0025] The functional modules n are sorted in ascending order of their management values Qn, and the management values Qn of the functional modules n are monitored and analyzed in a time period t. If the management value Qn of a functional module increases within the time period t, the corresponding functional module is identified and a monitoring signal is generated. Conversely, if the management value Qn of a functional module decreases or remains unchanged within the time period t, a normal signal is generated.
[0026] This invention provides a system construction method combining Internet of Things (IoT) and virtual reality technologies. Compared with existing technologies, it has the following advantages:
[0027] This invention first assesses the correctness of the acquired data, then performs an overall impact analysis on any erroneous data collected. If an impact exists, the data is re-collected; otherwise, it is deleted. This ensures the accuracy of subsequent model building and reduces overall error. Secondly, during model building, the model is analyzed based on the characteristics of the data. Simultaneously, the established twin model is used to predict the state of the equipment, and the model is optimized and adjusted based on the actual state to improve the predictive accuracy of the twin model. By collecting equipment data and using the twin model, the equipment is monitored in real time, improving the safety of equipment operation. Attached Figure Description
[0028] Figure 1 This is a flowchart of the method of the present invention;
[0029] Figure 2 This is a flowchart of the present invention. Detailed Implementation
[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0031] Please see Figure 1 This application provides a system construction method combining Internet of Things and virtual reality technologies, which specifically includes the following steps:
[0032] Step 1: Perform error identification on the collected construction data to obtain the identification results, which include correct data and erroneous data. Then, assess the impact of the erroneous data in the identification results and generate a judgment result, which includes whether there is an impact or not. The specific identification and judgment methods are as follows:
[0033] The system obtains the build data and then checks it. If there are errors in the build data, the corresponding error data is obtained; otherwise, no errors are processed.
[0034] In an automated manufacturing system, a robotic arm performs tasks based on pre-set construction data (such as movement coordinates, speed, and angles). The system first checks this construction data. Receiving construction data for the robotic arm's next movement from the control software, the system's built-in logic judgment module analyzes this data for issues such as values exceeding preset ranges, format errors, or logical conflicts. If data errors are found, such as a movement coordinate exceeding the robot's working range, the system marks these erroneous data points. If the data is completely correct, the robotic arm will execute the task according to the received construction data.
[0035] The acquired erroneous data is analyzed to determine whether it has any impact on the overall data structure. The specific determination method is as follows: the data type corresponding to the erroneous data is obtained and the analysis type is remembered. At the same time, the quantity ratio corresponding to the analysis type is calculated. Here, the quantity ratio represents the ratio of the data volume of the erroneous data belonging to the same type. The quantity ratio is compared with a preset value, and the specific value of the preset value is set by the operator. When the quantity ratio is greater than the preset value, it indicates that the proportion of the same type of erroneous data is large, and an impact result is generated. Conversely, when the quantity ratio is less than the preset value, a secondary analysis signal is generated.
[0036] Next, the generated secondary analysis signal is processed to obtain historical data, and the usage records of the same type of data in the historical data are obtained. The corresponding usage percentage is calculated and then compared with a threshold. The threshold is used to judge and compare the calculated usage percentage. The specific value of the threshold is set by the operator. When the usage percentage is greater than the threshold, an impact result is generated. Conversely, when the usage percentage is less than the threshold, an no-impact result is generated.
[0037] Unlike existing technologies, this application performs data identification and analysis during data collection. Based on the results of the identification and analysis, the data is classified into correct data and incorrect data. Furthermore, for incorrect data, it determines whether to remove it based on the impact of the incorrect data on the overall data. The specific determination method is to identify the data based on the proportion of the data volume corresponding to the incorrect data.
[0038] Step Two: Based on the acquired construction data, a model is built, and feature analysis and extraction are performed on the construction data. Simultaneously, a corresponding twin model is established based on the extracted features. Then, the equipment is analyzed using the twin model to obtain analysis results, including whether or not a hazard exists. The specific method for obtaining the analysis results is as follows:
[0039] All construction data is acquired, and the acquired construction data is confirmed to be correct. The construction data is then categorized according to data type to obtain data classification information. Any type of construction data is then analyzed and the target object is remembered. Next, the features of the target object are analyzed and extracted. These extracted features are then selected to obtain the construction features. The feature analysis and extraction are performed using deep learning algorithms. Specifically, `sklearn.feature_extraction.DictVectorizer` can be used to convert categorical features into one-hot encoded forms. The construction data is then analyzed to find the corresponding features. Further feature selection is performed using methods including Filter, Wrapper, and Embedded, based on statistical testing, model performance, and the algorithm itself, respectively.
[0040] Specifically, different feature extraction methods are used for different types of constructed data. For example, for text data, feature analysis is performed to obtain text features, the text data is converted into numerical features, and the bag-of-words model and TF-IDF (Term Frequency-Inverse Document Frequency) are used for corresponding feature extraction. For image data, deep learning methods such as convolutional neural networks (CNN) are the current mainstream image feature extraction technology.
[0041] Next, a digital twin model is constructed based on the obtained construction features. Here, the construction features are the features corresponding to different types of construction data. First, constructing a digital twin model requires understanding and performing data modeling. This process involves transforming physical systems or processes in the real world into a computable digital form. This includes collecting, organizing, and analyzing relevant data. Then, mathematical, statistical, and computer science methods are used to construct the model. At the same time, the potential hazards of the equipment are identified and judged based on the obtained twin model. The real-time features of the equipment are obtained, which represent the phenomena generated by the equipment during operation. The obtained real-time features are substituted into the twin model. The twin model analyzes and judges whether there are potential hazards in the current operation of the equipment, and generates analysis results, including whether there are hazards or not.
[0042] Specifically, the first step is to collect real-time operating data from the equipment, including key indicators such as temperature, vibration, and wear. This data reflects the equipment's status and behavior during actual operation. Using the collected data, a digital twin model is constructed. This model, through high-precision algorithms and machine learning techniques, can simulate the equipment's real-time operating state and predict its future performance. By analyzing the equipment's real-time characteristics through the twin model, it can determine whether there are any potential hazards. For example, if the temperature of a component suddenly rises, the model will immediately detect this change and, based on past data and the current operating environment, assess whether this change will lead to equipment failure or safety hazards.
[0043] For example, in a manufacturing company, key equipment on the production line (such as CNC machine tools) is equipped with various sensors to monitor its operating status in real time. By building digital twin models and inputting data from the sensors in real time, technicians can observe the equipment's operation in the control center. When the model detects abnormal vibration data of a component, it immediately analyzes possible causes, such as bearing wear or imbalance. The system quickly assesses the potential for mechanical failure or production accidents if this situation continues, and alerts operators to conduct inspections and maintenance through an early warning system.
[0044] Step 3: Periodically monitor the obtained twin model, and combine this with the physical status of the equipment to determine whether the twin model needs optimization and adjustment. Optimize the twin model based on the corresponding abnormal data, and generate optimization information. The specific method for generating optimization information is as follows:
[0045] The process involves acquiring the physical state of the device, which refers to the actual state information of the device, along with the corresponding actual data. This data is then input into the digital twin model to obtain a prediction result. The prediction result is compared with the physical state. If the prediction result matches the physical state, the digital twin model is considered normal and requires no optimization, and continuous monitoring information is generated. Conversely, if the prediction result differs from the physical state, the digital twin model is considered abnormal and requires optimization, and adjustment and optimization signals are generated. The key aspect of this stage is collecting real-time operational data from various sensors installed on the device. This data accurately reflects the actual working state of the device. The collected data is input into the established digital twin model. The model analyzes and calculates the data to output a prediction of the future device state. This prediction result is compared with the current physical state of the device. If the prediction result is highly consistent with the physical state, the digital twin model accurately reflects the actual operation of the physical device, is normal, and does not require immediate optimization. Conversely, if the prediction result deviates significantly from the physical state, the digital twin model may have problems and requires further optimization.
[0046] Next, the generated adjustment and optimization signals are analyzed to obtain the construction features that differ from the physical state of the equipment in the twin model. The construction features are then adjusted based on the physical state of the equipment. At the same time, historical data of the twin model is obtained, and the number of optimizations in the historical data is denoted as i, where i = 1, 2, ..., j, and j represents the number of optimizations. The optimization time corresponding to the number of optimizations i is also obtained and denoted as Ti. The optimization time interval is calculated, and then the time mean of the optimization time interval is calculated. The optimization information is generated based on the mean.
[0047] Unlike existing technologies, this method continuously monitors and processes the constructed model. During model monitoring, it analyzes the actual state parameters of the equipment and optimizes the model based on the differences between the processed and actual results. This improves the overall accuracy of the model's predictions. Furthermore, for results requiring optimization, it analyzes any anomalies and processes them using historical data. It also calculates periodic optimization intervals, ensuring regular model verification and optimization during subsequent use. This allows the model to be updated and further improves its predictive accuracy.
[0048] Step 4: Based on the twin model, corresponding features are used to generate different functional modules. Simultaneously, historical data from different functional modules is used for comprehensive analysis and calculation of their management values. The functional modules are then analyzed and adjusted based on changes in these management values. The specific analysis and adjustment methods are as follows:
[0049] The construction features of the twin model are obtained, and corresponding functional modules are generated based on these features. Taking a wind turbine as an example, the following functional modules can be designed based on its construction features: a performance analysis module, which uses aerodynamic models and real-time wind speed data to predict power generation and analyze performance under current wind speeds; and a fault prediction module, which analyzes historical maintenance records and real-time monitoring data (such as vibration and temperature) to predict wear and potential faults in key components; and a similar module generation process for all construction features, labeling each module as n (n = 1, 2, ..., m), where m represents the number of functional modules. The usage of functional module n is then obtained, including usage count Cn, data generation volume Kn, and abnormal data volume Pn. These parameters are then substituted into the formula. The management value Qn corresponding to functional module n is calculated;
[0050] Functional modules n are sorted in ascending order of their management value Qn. The management value Qn of each functional module n is monitored and analyzed at time intervals t. If the management value Qn of a functional module increases within time interval t, the corresponding functional module is identified, and a monitoring signal is generated. Conversely, if the management value Qn decreases or remains unchanged within time interval t, a normal signal is generated. For functional modules that generate monitoring signals, the monitored data needs to be analyzed to identify the components that generated the corresponding data and further inspect and analyze them. Functional modules that generate normal signals are not processed, but their corresponding data is monitored.
[0051] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.
[0052] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.
Claims
1. A system construction method combining the Internet of Things and virtual reality technology, characterized in that: The specific steps include: Step 1: Perform error recognition on the collected construction data to obtain recognition results, wherein the recognition results include correct data and erroneous data, and perform impact judgment on the erroneous data in the recognition results, and generate judgment results at the same time, wherein the judgment results include the existence of impact and the absence of impact. For the erroneous data, determine whether it needs to be eliminated based on the impact of the erroneous data on the whole. The specific judgment method is: determine and recognize based on the data capacity ratio corresponding to the erroneous data; Step 2: Build a model based on the acquired construction data, analyze and extract features of the construction data, and establish a corresponding twin model based on the extracted features. Then, analyze the equipment based on the twin model to obtain analysis results, and the analysis results include whether there is danger or not. Step 3: Periodically monitor the obtained twin model, and determine whether the twin model needs to be optimized and adjusted based on the physical status of the equipment. Optimize the twin model based on the corresponding abnormal data and generate optimization information. Step 4: Generate different functional modules according to the construction features corresponding to the twin model, and perform comprehensive analysis and calculation of the management value of the functional modules based on the historical data of different functional modules. Analyze and adjust the functional modules according to the changes in the management value. The specific methods are as follows: Obtain the construction features of the twin model, and generate corresponding functional modules according to the construction features. The functional modules are labeled as n, and n=1, 2, ..., m, where m represents the number of functional modules. Then obtain the usage of functional module n, where the usage includes the number of times used Cn, the amount of data generated Kn and the amount of abnormal data Pn, and substitute the obtained parameters into the formula Calculate and obtain the management value Qn corresponding to the functional module n; Sort the functional modules n according to the management value Qn from small to large, and monitor and analyze the management value Qn of the functional module n with time t as a period. If the management value Qn of the functional module increases within the time period t, the corresponding functional module is obtained and a monitoring signal is generated. Otherwise, if the management value Qn of the functional module decreases or remains unchanged within the time period t, a normal signal is generated.
2. The method for constructing a system combining the Internet of Things and virtual reality technology according to claim 1, characterized in that: The specific method of generating the judgment result in step 1 is: Acquire the build data and judge the acquired build data. If there is an error in the build data, obtain the corresponding error data. If there is no error in the build data, do not process it. Analyze the acquired erroneous data to determine whether the erroneous data has any impact on the overall construction of the data, and the specific judgment method is: obtain the data type corresponding to the erroneous data and record it as the analysis type, and calculate the quantity ratio corresponding to the analysis type, and compare the quantity ratio with the preset value. When the quantity ratio is greater than the preset value, it means that the erroneous data of the same type accounts for a large proportion, and an impact result is generated. Conversely, when the quantity ratio is less than the preset value, a secondary analysis signal is generated and processed at the same time.
3. The method for constructing a system combining the Internet of Things and virtual reality technology according to claim 2, characterized in that: The specific method of processing the secondary analysis signal in step 1 is: Obtain historical data, and at the same time obtain the usage records of the same type of data in the historical data, and calculate the corresponding usage ratio value, and then compare the usage ratio value with the threshold. When the usage ratio value is greater than the threshold, an impact result is generated. Conversely, when the usage ratio value is less than the threshold, a no impact result is generated.
4. The method for constructing a system combining the Internet of Things and virtual reality technology according to claim 1, characterized in that: The specific method of obtaining the analysis result in step 2 is: All construction data are obtained, and the construction data are classified according to data types to obtain data classification information, and any type of construction data is obtained for analysis and recorded as a target object, and then the features of the target object are analyzed and extracted, and the extracted features are selected to obtain construction features; Then, a model is constructed based on the obtained construction features to obtain a twin model. At the same time, the potential dangers of the equipment are identified and judged based on the obtained twin model, and the real-time characteristics of the equipment are obtained. The acquired real-time characteristics are substituted into the twin model. The twin model is used to analyze and judge whether there are potential dangers in the current operation of the equipment, and the analysis results are generated at the same time. The analysis results include the existence of danger and the non-existence of danger.
5. The method for constructing a system combining the Internet of Things and virtual reality technology according to claim 1, characterized in that: The specific method for determining whether the twin model needs to be optimized and adjusted in step 3 is: Obtain the physical state of the equipment and the actual data corresponding to the physical state, and substitute the actual data into the twin model to obtain the prediction result. Then compare the prediction result with the physical state. If the prediction result is the same as the physical state, it means that there is no abnormality in the twin model and generates continuous monitoring information. Otherwise, if the prediction result is different from the physical state, it means that there is an abnormality in the twin model and generates an adjustment optimization signal, which is analyzed at the same time.
6. The method for constructing a system combining the Internet of Things and virtual reality technology according to claim 5, characterized in that: The specific method of analyzing the adjustment optimization signal in step 3 is: Obtain the construction features in the twin model that are different from the physical state of the device, and adjust the construction features based on the features of the physical state of the device. At the same time, obtain the historical data of the twin model, and obtain the number of optimizations in the historical data, denoted as i, and i=1, 2,…, j, where j represents the number of optimizations. At the same time, obtain the optimization time corresponding to the optimization number i, denoted as Ti, and calculate the optimization time interval, then calculate the time mean of the optimization time interval, and generate optimization information based on the mean.