Data processing method and system based on animal feed mixing control system

Abstract

The embodiments of the present invention relate to the field of data processing technology, and specifically to a data processing method and system based on an animal feed mixing control system. The method realizes comprehensive and accurate monitoring of the operating status of a target control system through innovative data processing and analysis technologies, and can timely identify potential abnormal problems, thereby providing strong technical support for the stable operation and efficient production of the animal feed mixing control system.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and more specifically, to a data processing method and system based on an animal feed mixing control system. Background Art

[0002] In modern industrial production, the monitoring and management of target control systems are crucial to ensuring production efficiency and product quality. However, traditional monitoring methods often rely on single sensor data, which makes it difficult to fully and accurately reflect the actual operating status of the control system. Especially in complex industrial environments, such as animal feed mixing control systems, their operating status is affected by many factors, including the humidity, temperature, and mixing ratio of the raw materials, and there are complex interactions between these factors. Therefore, how to effectively integrate and analyze multi-source sensor monitoring information to achieve comprehensive and accurate monitoring of the control system's operating status has become a technical problem that needs to be solved urgently.

[0003] In recent years, the rapid development of artificial intelligence technology has provided new ideas and methods for the monitoring and management of industrial control systems. In particular, the application of artificial intelligence has shown great potential in animal feed mixing control systems. By introducing advanced technologies such as machine learning and deep learning, a large amount of sensor monitoring data can be efficiently processed and analyzed, and valuable operating status information can be extracted from it, thereby realizing real-time monitoring of the operating status of the control system and abnormal identification. However, how to effectively extract, integrate and analyze this information in the context of multi-mode IOT sensor monitoring information is still a challenging technical problem. Summary of the invention

[0004] In view of this, the present invention provides a data processing method and system based on an animal feed mixing control system.

[0005] An embodiment of the present invention provides a data processing method based on an animal feed mixing control system, which is applied to a data processing system. The method includes: acquiring multi-mode IOT sensor monitoring information of a target control system; extracting control system operation status information of a plurality of sensor monitoring category labels from the multi-mode IOT sensor monitoring information, and mapping the extracted plurality of control system operation status information into a plurality of operation status embedded vector chains; using the number of first embedded vectors as a linear scanning frame size, and the number of second embedded vectors as a scanning period, performing linear scanning frame-based segmentation on each of the operation status embedded vector chains to obtain a first operation status embedded vector subset corresponding to each of the operation status embedded vector chains; performing state element integration on each of the first operation status embedded vector subsets, and interacting the integrated plurality of operation status integrated vectors to obtain a global sensor monitoring vector of the target control system; and determining the abnormal recognition viewpoint of the target control system based on the global sensor monitoring vector.

[0006] The present invention also provides a data processing system, comprising: a memory for storing program instructions and data; a processor for coupling with the memory and executing instructions in the memory to implement the above method.

[0007] The present invention also provides a computer storage medium, comprising instructions, and when the instructions are executed on a processor, the above method is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0009] Figure 1 A schematic flow chart of the steps of a data processing method based on an animal feed mixing control system provided by an embodiment of the present invention.

[0010] Figure 2 A structural block diagram of a data processing system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0011] The technical solution of the present invention will be described below in conjunction with the accompanying drawings.

[0012] When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention. It should be noted that the terms "first", "second", etc. in the specification of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0013] The method embodiments provided in the embodiments of the present invention can be executed in a data processing system, a computer device or a similar computing device. Taking running on a data processing system as an example, the data processing system may include one or more processors (the processor may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory for storing data. Optionally, the above-mentioned data processing system may also include a transmission device for communication functions. It can be understood by those skilled in the art that the above-mentioned structure is only illustrative and does not limit the structure of the above-mentioned data processing system. For example, the data processing system may also include more or fewer components than those shown above, or have a configuration different from that shown above.

[0014] The memory can be used to store computer programs, for example, software programs and modules of application software, such as a computer program corresponding to a data processing method based on an animal feed mixing control system in an embodiment of the present invention. The processor executes various functional applications and data processing by running the computer program stored in the memory, that is, implementing the above method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the data processing system via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0015] The transmission device is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the data processing system. In one example, the transmission device includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device can be a radio frequency (Radio Frequency, referred to as RF) module, which is used to communicate with the Internet wirelessly.

[0016] In conjunction with the above, please refer to Figure 1 , Figure 1 It is a flow chart of a data processing method based on an animal feed mixing control system provided by an embodiment of the present invention. The method is applied to a data processing system and may further include steps 110 to 130.

[0017] Step 110, obtaining multi-mode IOT sensor monitoring information of the target control system; extracting control system operation status information of a plurality of sensor monitoring category labels from the multi-mode IOT sensor monitoring information, and mapping the extracted plurality of control system operation status information into a plurality of operation status embedded vector chains.

[0018] Step 120, using the number of first embedded vectors as the linear scanning frame size and the number of second embedded vectors as the scanning period, segmenting each of the operating state embedded vector chains based on the linear scanning frame to obtain a first operating state embedded vector subset corresponding to each of the operating state embedded vector chains; integrating state elements of each of the first operating state embedded vector subsets, and interacting the integrated operating state integrated vectors to obtain a global sensing monitoring vector of the target control system.

[0019] Step 130: Determine an abnormality recognition perspective of the target control system based on the global sensor monitoring vector.

[0020] In the animal feed production process, the mixing control system plays a vital role. In order to ensure the accuracy and efficiency of feed mixing, an efficient data processing system is essential. The following is a specific application scenario example, which details how the data processing system is implemented in the animal feed mixing control system and realizes abnormality identification through a series of steps to improve the overall production efficiency: In a modern animal feed production plant, the feed mixing control system (target control system) is responsible for mixing a variety of different raw materials according to a precise formula to produce high-quality animal feed. The feed mixing control system is equipped with a variety of IOT sensors for real-time monitoring and control of various parameters in the mixing process, such as temperature, humidity, mixing time, feed amount of raw materials, etc. The data processing system is responsible for collecting, processing and analyzing these sensor monitoring information to ensure the stability and efficiency of the mixing process.

[0021] First, the data processing system first obtains real-time multi-mode IOT sensor monitoring information by connecting with the target control system. Multi-mode IOT sensor monitoring information includes various data from different sensors, such as the internal temperature data of the mixer provided by the temperature sensor, the raw material humidity data provided by the humidity sensor, and the raw material feed amount data provided by the flow sensor. After obtaining these sensor monitoring information, the data processing system pre-processes these information to extract the control system operation status information of several sensor monitoring category labels. For example, the "temperature" category operation status information is extracted from the temperature sensor data, the "humidity" category operation status information is extracted from the humidity sensor data, and the "feed amount" category operation status information is extracted from the flow sensor data. Next, the data processing system maps these extracted control system operation status information into several operation status embedded vector chains. Each operation status embedded vector chain represents the operation status of the target control system in a specific time period, and each embedded vector in the operation status embedded vector chain corresponds to a specific sensor monitoring category label.

[0022] Secondly, after obtaining several operation state embedding vector chains, the data processing system uses the number of first embedding vectors as the linear scanning frame size and the number of second embedding vectors as the scanning period to perform linear scanning frame segmentation on each operation state embedding vector chain. The purpose of this step is to segment the long operation state embedding vector chain into several shorter first operation state embedding vector subsets for subsequent state element integration. For each first operation state embedding vector subset, the data processing system performs state element integration and integrates all embedding vectors in the subset to obtain an operation state integration vector. This integration vector represents the overall operation state of the target control system under a specific sensor monitoring category label during the scanning period. After completing the state element integration of all first operation state embedding vector subsets, the data processing system obtains several operation state integration vectors. Next, these operation state integration vectors interact and obtain a global sensor monitoring vector through a specific interaction mechanism (such as an attention mechanism or a gating mechanism). This global sensor monitoring vector integrates the operation state information under all sensor monitoring category labels and represents the global operation state of the control system in the current time period.

[0023] Finally, after obtaining the global sensor monitoring vector, the data processing system determines the abnormality recognition viewpoint of the target control system based on this vector. Specifically, the data processing system pre-trains an abnormality recognition model (such as the subsequent material handling state analysis algorithm), which can judge whether the control system is in an abnormal state based on the characteristics of the global sensor monitoring vector. The abnormality recognition model can be trained using machine learning or deep learning methods. During the training process, the data processing system provides a large number of global sensor monitoring vectors of normal operating status and abnormal operating status as training data, allowing the model to learn how to distinguish between these two states. After the training is completed, the model can judge whether the operating state of the control system is normal based on the new global sensor monitoring vector. If the abnormality recognition model determines that the control system is in an abnormal state, the data processing system immediately issues an alarm and provides relevant abnormal information to the operator. The operator can conduct timely inspection and maintenance of the control system based on this information to avoid possible production failures and product quality problems.

[0024] In combination with the above contents, step 110, step 120 and step 130 are respectively described in detail below.

[0025] In step 110, multi-mode IOT sensor monitoring information refers to various types of monitoring data collected in real time by various Internet of Things (IOT) sensors installed in the animal feed mixing control system. These data come from different sensors, such as temperature sensors, humidity sensors, flow sensors, etc., which together constitute a comprehensive and multi-dimensional monitoring of the mixing control system. For example, the temperature sensor will provide real-time temperature data inside the mixer, the humidity sensor will provide humidity data of the raw materials, and the flow sensor will provide feed volume data of the raw materials. These data together form multi-mode IOT sensor monitoring information, which provides a rich data source for the data processing system to comprehensively analyze and judge the operating status of the mixing control system.

[0026] Sensor monitoring category labels are a way to classify and identify different types of data in multi-mode IoT sensor monitoring information. In the animal feed mixing control system, the data collected by each sensor belongs to a specific sensor monitoring category, such as "temperature", "humidity", "feed amount", etc. These category labels help the data processing system to distinguish and process different types of sensor monitoring information. For example, when the data processing system receives data from a temperature sensor, it will mark this data as sensor monitoring information of the "temperature" category and perform corresponding processing and analysis.

[0027] The control system operating status information refers to the operating status of the animal feed mixing control system in a specific time period reflected by the multi-mode IOT sensor monitoring information. This information includes the real-time data of the control system under various sensor monitoring categories, such as temperature, humidity, feed volume, etc., and the changes of these data over time. By analyzing and processing this information, the data processing system can understand the real-time operating status of the control system, determine whether it is working properly, and whether there are potential faults or abnormalities.

[0028] The operating state embedding vector chain is a vectorized representation method of the operating state information of the control system by the data processing system. In the animal feed mixing control system, the data processing system maps the operating state information of the control system under each sensor monitoring category into one or more embedding vectors, and connects these embedding vectors in chronological order to form an operating state embedding vector chain. This representation method helps the data processing system to perform more refined analysis and processing of the operating state of the control system, because the embedding vector can capture the complex features and patterns in the operating state information.

[0029] In detail, in step 110, the data processing system first establishes a connection with the target control system and obtains multi-mode sensor monitoring information in real time through the IOT sensor network. This information includes various data from different sensors, such as temperature, humidity, feed amount, etc., which together constitute a comprehensive, multi-dimensional real-time monitoring of the hybrid control system. After obtaining the multi-mode IOT sensor monitoring information, the data processing system preprocesses and parses the information. It will classify the information according to the sensor monitoring category label and extract the control system operation status information under each category. For example, it will extract the operation status information of the "temperature" category from the temperature sensor data, including real-time temperature data and temperature change trends. Finally, the data processing system maps the extracted control system operation status information into several operation status embedded vector chains. This mapping process is achieved by converting the data into vector form, each vector represents the operation status of the control system in a specific time period and sensor monitoring category. By connecting these vectors in series in chronological order, the data processing system can obtain a series of operation status embedded vector chains, which together describe the overall operation status and change trend of the control system. This step provides a basis for subsequent data analysis and abnormality identification.

[0030] In step 120, the linear scan frame size refers to the fixed-length window size used when segmenting the operating state embedded vector chain in the data processing system. This size determines the number of operating state embedded vectors included in each segmentation. In the application scenario of the animal feed mixing control system, the linear scan frame size can be set according to actual needs, for example, 10 or 20 consecutive operating state embedded vectors can be selected as a scan frame. Such a design helps to regularly segment the embedded vector chain, which is convenient for the subsequent state element integration and the generation of the global sensor monitoring vector.

[0031] The scanning cycle refers to the frequency or period at which the data processing system segments the operating state embedding vector chain and integrates the state elements. It determines how long or how many embedding vectors the system performs a new segmentation and integration operation. In the animal feed mixing control system, the scanning cycle can be set according to the actual monitoring needs and production process. For example, it can be selected to scan and integrate once every minute or after a certain amount of raw materials are processed. Such a design helps ensure that the data processing system can capture the operating state changes of the control system in a timely manner.

[0032] Linear scan frame-based segmentation refers to the data processing system performing continuous, non-overlapping segmentation operations on the running state embedding vector chain according to the preset linear scan frame size. Each scan frame contains a fixed number of continuous embedding vectors for subsequent state element integration. This segmentation method helps to decompose long embedding vector chains into multiple shorter subsets, which is convenient for data processing and analysis.

[0033] The first operating state embedding vector subset refers to a subset containing a fixed number of operating state embedding vectors obtained by a segmentation operation based on a linear scanning frame. Each subset represents the operating state information of the control system in a specific time period. In the animal feed mixing control system, these subsets provide basic data for the subsequent state element integration.

[0034] State element integration refers to the process in which the data processing system performs an integration operation on all embedded vectors in the first operating state embedded vector subset to obtain an integrated vector. This integrated vector combines the information of all embedded vectors in the subset and represents the overall operating state of the control system under a certain sensor monitoring category during the scanning cycle. The integration operation can use various mathematical methods, such as vector summation, averaging or weighted averaging.

[0035] The global sensor monitoring vector refers to the final vector obtained by the data processing system after integrating the state elements of all first operating state embedded vector subsets and performing interactive operations on several obtained operating state integrated vectors. This vector integrates the operating state information of the control system under all sensor monitoring categories and represents the global operating state of the control system in the current time period. The global sensor monitoring vector is an important basis for subsequent abnormality identification.

[0036] In detail, in step 120, the data processing system first uses the number of first embedded vectors as the linear scan frame size to perform linear scan frame segmentation on each operating state embedded vector chain. This operation will decompose the long embedded vector chain into multiple shorter first operating state embedded vector subsets containing a fixed number of embedded vectors. Such a segmentation method is helpful for subsequent independent state element integration and analysis of the subsets. Next, the data processing system performs state element integration on each first operating state embedded vector subset. This operation integrates all embedded vectors in the subset to obtain an operating state integrated vector. The integration operation can use various mathematical methods to ensure that the integrated vector can accurately reflect the comprehensive information of all embedded vectors in the subset. Through state element integration, the data processing system can convert the scattered embedded vectors into a more compact and meaningful integrated vector, which is convenient for subsequent analysis and processing. Finally, the data processing system performs interactive operations on the integrated several operating state integrated vectors to obtain the global sensor monitoring vector of the target control system. This interactive operation can use various mathematical or machine learning methods to ensure that the global sensor monitoring vector can accurately reflect the overall operating state of the control system under all sensor monitoring categories. The global sensor monitoring vector is an important basis for subsequent abnormality identification. It integrates the operating status information of the control system in all aspects and provides comprehensive and multi-dimensional monitoring data for the data processing system.

[0037] In step 130, the abnormal identification viewpoint refers to the judgment or evaluation given by the data processing system on whether the operating state of the target control system is abnormal after analyzing and processing the global sensor monitoring vector. In the application scenario of the animal feed mixing control system, the abnormal identification viewpoint is the conclusion drawn by the data processing system after reasoning and judging through a specific algorithm or model based on the comprehensive information in the global sensor monitoring vector. This viewpoint represents the data processing system's view on whether the current operating state of the control system is normal, and is an important basis for subsequent abnormal alarm or processing. Specifically, the abnormal identification viewpoint can include different levels or classifications such as "normal operation", "minor abnormality", and "serious abnormality" to reflect the severity of the operating state of the control system. The data processing system determines the most appropriate abnormal identification viewpoint based on the characteristics, trends in the global sensor monitoring vector, and the comparison results with other normal or abnormal states.

[0038] In the application scenario of the animal feed mixing control system, step 130 is the key step for the data processing system to perform the final analysis and judgment on the global sensor monitoring vector. Specifically, this step includes the following sub-steps: the data processing system first receives and inputs the global sensor monitoring vector, which integrates the operating status information of the target control system under all sensor monitoring categories and represents the global operating status of the control system in the current time period; next, the data processing system applies a pre-trained or set abnormal recognition algorithm or model to analyze and judge the global sensor monitoring vector. This algorithm or model can be constructed based on machine learning, deep learning or statistical analysis methods (which will be introduced later), and can determine whether the operating status of the control system is abnormal based on the characteristics, trends in the global sensor monitoring vector and the comparison results with other normal or abnormal states; finally, the data processing system determines the abnormal recognition view of the target control system based on the output results of the abnormal recognition algorithm or model. This view can be different levels or classifications such as "normal operation", "minor abnormality" and "serious abnormality" to reflect the severity of the operating status of the control system. The data processing system uses this view as the final analysis result for subsequent abnormal alarm, processing or further analysis and decision-making.

[0039] It is worth mentioning that in the animal feed mixing control system, the detailed refinement of the abnormal identification viewpoint is the key to ensure that the system can accurately identify and respond to various potential problems. When determining the abnormal identification viewpoint of the target control system based on the global sensor monitoring vector, the data processing system considers multiple aspects and factors. The following is a detailed description of various possible situations:

[0040] (1) Normal operation: When all indicators in the global sensor monitoring vector are within the normal range and there is no significant deviation from the historical normal data, the data processing system determines that the target control system is in normal operation. This means that the key parameters of the mixer, such as temperature, humidity, and feed volume, are within the expected control range and there is no abnormality or deviation from the normal trend.

[0041] (2) Minor abnormality: If some indicators in the global sensor monitoring vector deviate slightly from the normal range, but have not yet reached the level of serious abnormality, the data processing system will identify it as a minor abnormality. For example, the temperature of the mixer may be slightly higher than the normal set value, but still within the acceptable range. In this case, the system may issue a warning to prompt the operator to pay attention and take appropriate adjustment measures.

[0042] (3) Severe abnormality: When multiple indicators in the global sensor monitoring vector deviate significantly from the normal range and are significantly abnormal compared with historical normal data, the data processing system determines that the target control system is in a serious abnormal state. For example, the temperature of the mixer rises sharply, far exceeding the normal set value, and the feed volume also fluctuates abnormally. In this case, the system immediately sends an alarm signal, requiring the operator to take immediate measures to prevent equipment damage or production accidents.

[0043] (4) Trending abnormalities: In addition to immediate abnormality identification, the data processing system also analyzes trend information in the global sensor monitoring vector. If certain parameters continue to deviate from the normal range for a period of time and show a clear abnormal trend, the system will also identify it as an abnormality. For example, although the temperature of the mixer has not yet exceeded the normal range, it has been rising continuously and the trend is obvious. In this case, the system may issue an early warning so that the operator can intervene in time and prevent potential problems from occurring.

[0044] (5) Combined abnormalities: In actual applications, multiple abnormalities may exist at the same time. The data processing system comprehensively considers various indicators and trends in the global sensor monitoring vector to identify combined abnormalities. For example, the temperature, humidity and feed volume of the mixer fluctuate abnormally at the same time. In this case, the system comprehensively analyzes various abnormal information and gives corresponding abnormal identification opinions and processing suggestions.

[0045] It can be seen that when determining the abnormality identification viewpoint of the target control system based on the global sensor monitoring vector, the data processing system comprehensively considers multiple aspects and factors, including real-time indicators, historical data, trend information, and the combination of multiple abnormalities. By refining the abnormality identification viewpoint in detail, the data processing system can more accurately identify and respond to potential problems, thereby ensuring the stable operation and production efficiency of the animal feed mixing control system.

[0046] In summary, the embodiment of the present invention achieves a comprehensive perception of the operating state of the control system by acquiring the multi-mode IOT sensor monitoring information of the target control system. By extracting the control system operating state information of several sensor monitoring category labels from this information and mapping it into several operating state embedded vector chains, the embodiment of the present invention effectively converts complex sensor monitoring data into a structured and easy-to-process form. Further, the operating state embedded vector chain is segmented based on the linear scanning frame with a specific number of embedded vectors as the linear scanning frame size and scanning period, which not only improves the efficiency of data processing, but also makes the monitoring of the operating state of the control system more detailed and accurate. By integrating the state elements of the first operating state embedded vector subset after segmentation, and interacting the several integrated operating state integrated vectors obtained by integration, a global sensor monitoring vector is finally obtained. This step greatly improves the accuracy and comprehensiveness of abnormal identification. Finally, the abnormal identification viewpoint of the target control system is determined based on the global sensor monitoring vector, which provides strong support for timely discovery and processing of potential problems in the control system, thereby significantly improving the stability and operation efficiency of the control system.

[0047] In some optional embodiments, the state element integration is performed on each of the first operating state embedded vector subsets, and several integrated operating state integrated vectors are interacted to obtain a global sensor monitoring vector of the target control system, including: determining the first material processing state analysis algorithm corresponding to each of the first operating state embedded vector subsets; inputting the first operating state embedded vector subset corresponding to each of the operating state embedded vector chains into the corresponding first material processing state analysis algorithm to perform state element integration to obtain an operating state integrated vector corresponding to each of the operating state embedded vector chains; and performing knowledge vector collision on several of the operating state integrated vectors to obtain a global sensor monitoring vector of the target control system.

[0048] In this embodiment, the data processing system is responsible for the in-depth analysis and processing of multi-mode IOT sensor monitoring information to achieve accurate monitoring and abnormality identification of the target control system's operating status.

[0049] First, the data processing system extracts the control system operation status information of several sensor monitoring category tags from the multi-mode IOT sensor monitoring information of the animal feed mixing control system. This information covers multiple key parameters such as temperature, humidity, mixing ratio, and feeding speed, and fully reflects the real-time operation status of the control system. Subsequently, the data processing system maps the extracted control system operation status information into several operation status embedded vector chains, each of which corresponds to the change in the operation status of the control system within a certain period of time.

[0050] Next, the data processing system performs linear scan frame segmentation on each operating state embedded vector chain with a specific number of embedded vectors as the linear scan frame size and another specific number of embedded vectors as the scan period. The purpose of this step is to segment the long operating state embedded vector chain into several shorter first operating state embedded vector subsets to facilitate more detailed and efficient state element integration. The selection of the linear scan frame size and scan period is based on a comprehensive consideration of the speed of change of the control system operating state and the data processing efficiency, to ensure that the segmented vector subsets can accurately reflect the changes in the operating state and are not too long, resulting in reduced processing efficiency.

[0051] After obtaining the first operating state embedded vector subset, the data processing system needs to determine the first material processing state analysis algorithm corresponding to each vector subset. The first material processing state analysis algorithm here is a series of algorithm models that are pre-designed and trained according to the characteristics and requirements of different material processing stages in the animal feed mixing process. These algorithm models can extract the key elements reflecting the material processing state according to the operating state embedded vectors in the vector subset, and integrate them to obtain the operating state integrated vector corresponding to each operating state embedded vector chain.

[0052] Specifically, when the data processing system inputs the first operating state embedded vector subset corresponding to a certain operating state embedded vector chain into the corresponding first material processing state analysis algorithm, the algorithm will first decode and parse each embedded vector in the vector subset, and extract the key parameter values ​​such as temperature, humidity, and mixing ratio. Then, the algorithm will calculate a series of state elements reflecting the material processing state based on these parameter values ​​and their changing trends and relationships, such as the mixing uniformity of the material, the humidity control accuracy, the stability of the feeding speed, etc. Finally, the algorithm integrates these state elements to form an operating state integrated vector containing the overall material processing state information of the operating state embedded vector chain.

[0053] After obtaining several operating state integrated vectors, the data processing system needs to perform knowledge vector collision on them to obtain the global sensor monitoring vector of the target control system. Knowledge vector collision is an advanced data processing technology that simulates the association and reasoning process of the human brain when processing complex information. Specifically, the data processing system takes several operating state integrated vectors as input and crosses, compares and merges them with each other through preset collision rules and algorithms. In this process, the commonalities and differences between different operating state integrated vectors will be identified, and a new, more comprehensive and in-depth global sensor monitoring vector will be formed. This global sensor monitoring vector not only contains the key parameters and state element information of the control system under various operating states, but also reflects the internal connection and change law between these information and elements, providing strong data support for subsequent abnormality identification.

[0054] Finally, the data processing system determines the abnormality recognition view of the target control system based on the global sensor monitoring vector. This step is the core goal of the technical solution. Through the previous series of data processing and analysis processes, the data processing system has obtained a global sensor monitoring vector that comprehensively and accurately reflects the operating status of the control system. At this time, the data processing system uses a pre-trained abnormality recognition model or algorithm to conduct in-depth mining and analysis of the global sensor monitoring vector. The abnormality recognition model will determine whether the control system is in an abnormal state based on the key parameters and state element information in the vector, as well as the inherent connection and change rules between them, and give corresponding abnormality recognition views. These views may include different levels or classifications such as "normal operation", "minor abnormality", and "serious abnormality", providing operators with timely and accurate abnormality alarms and processing suggestions.

[0055] In this way, the multi-mode IOT sensor monitoring information of the animal feed mixing control system is comprehensively and deeply analyzed and processed by the data processing system, so as to realize the accurate monitoring and abnormal identification of the operating state of the control system. First, by extracting the operating state information of the control system of several sensor monitoring category labels from the multi-mode IOT sensor monitoring information and mapping it into several operating state embedded vector chains, the embodiment of the present invention effectively realizes the structured and vectorized processing of complex sensor monitoring data, which provides convenience for subsequent data analysis and processing. By segmenting the long operating state embedded vector chain into several shorter first operating state embedded vector subsets through the segmentation method based on the linear scanning frame, the embodiment of the present invention improves the efficiency and accuracy of data processing. This step makes the monitoring of the operating state of the control system more detailed and comprehensive, and can timely discover and deal with potential abnormal problems. By determining the first material processing state analysis algorithm corresponding to each first operating state embedded vector subset, and performing state element integration and knowledge vector collision, the embodiment of the present invention realizes the in-depth analysis and mining of the operating state of the control system. This step not only extracts the key element information reflecting the material processing state, but also reveals the inherent connection and change law between these information, which provides strong data support for abnormal identification. By determining the abnormal identification viewpoint of the target control system based on the global sensor monitoring vector, the embodiment of the present invention realizes timely and accurate identification of the abnormal state of the control system. This step provides operators with timely and accurate abnormal alarms and processing suggestions, effectively improving the stability and operating efficiency of the control system.

[0056] In other optional embodiments, performing a knowledge vector collision on a plurality of the operating state integrated vectors to obtain a global sensor monitoring vector of the target control system includes: determining a first confidence coefficient of each of the operating state integrated vectors; optimizing each of the operating state integrated vectors according to the first confidence coefficient to obtain a plurality of operating state optimized vectors; performing a knowledge vector collision on the plurality of operating state optimized vectors to obtain a global sensor monitoring vector of the target control system.

[0057] Furthermore, as the core executive body, the data processing system is responsible for in-depth analysis and processing of information from multi-mode IOT sensor monitoring, aiming to achieve comprehensive and accurate monitoring of the operating status of the target control system and abnormality identification.

[0058] First, the data processing system extracts several control system operating status information with sensor monitoring category tags from the multi-mode IOT sensor monitoring information of the animal feed mixing control system. This information comprehensively covers multiple key parameters such as temperature, humidity, mixing ratio, and feeding speed, and can reflect the operating status of the control system in real time. Subsequently, the data processing system maps the extracted information into several operating status embedded vector chains, each of which corresponds to the operating status change trajectory of the control system within a certain period of time.

[0059] Next, the data processing system uses a specific number of embedded vectors as the size of the linear scan frame and another specific number of embedded vectors as the scan period to perform linear scan frame-based segmentation on each operating state embedded vector chain. The purpose of this step is to segment the long operating state embedded vector chain into several shorter first operating state embedded vector subsets in order to perform more detailed and efficient state element integration. Through this segmentation method, the data processing system can more accurately capture the changes in the operating state of the control system and facilitate subsequent data processing.

[0060] After obtaining the first operating state embedding vector subset, the data processing system determines the first material processing state analysis algorithm corresponding to each vector subset. These algorithms are pre-designed and trained according to the characteristics and requirements of different material processing stages in the animal feed mixing process. By applying these algorithms, the data processing system can extract the key elements reflecting the material processing state from the operating state embedding vectors in the vector subset and integrate them to obtain the operating state integration vector corresponding to each operating state embedding vector chain.

[0061] However, after obtaining the running state integrated vector, the data processing system does not directly perform knowledge vector collision, but first further optimizes each running state integrated vector. Specifically, the data processing system determines the first confidence coefficient of each running state integrated vector. This confidence coefficient reflects the accuracy and reliability of the running state information contained in the vector. By calculating the confidence coefficient, the data processing system can assign weights to different running state integrated vectors, thereby more accurately reflecting the global sensor monitoring information of the control system in the subsequent knowledge vector collision.

[0062] After determining the first confidence coefficient, the data processing system optimizes each operating state integrated vector according to this coefficient to obtain several operating state optimized vectors. The optimization process may include vector weighting, smoothing, denoising and other processing steps to improve the quality and accuracy of the vector. Through such optimization processing, the data processing system can further improve the accuracy and reliability of the global sensor monitoring vector.

[0063] Finally, the data processing system performs knowledge vector collision on several operating state optimization vectors to obtain the global sensor monitoring vector of the target control system. Knowledge vector collision is an advanced data processing technology that simulates the association and reasoning process of the human brain when processing complex information. Through this step, the data processing system can cross, compare and fuse different operating state optimization vectors to form a global sensor monitoring vector that reflects the operating state of the control system more comprehensively and deeply. This vector not only contains the key parameters and state element information of the control system under various operating states, but also reflects the inherent connection and change law between these information and elements.

[0064] Based on the global sensor monitoring vector, the data processing system can further determine the abnormality recognition viewpoint of the target control system. By applying the pre-trained abnormality recognition model or algorithm, the data processing system conducts in-depth mining and analysis of the global sensor monitoring vector. The abnormality recognition model will determine whether the control system is in an abnormal state based on the key parameters and state element information in the vector and the inherent connection and change rules between them, and give corresponding abnormality recognition views. These views provide operators with timely and accurate abnormality alarms and processing suggestions, which help improve the stability and operation efficiency of the control system.

[0065] In this way, the embodiment of the present invention performs comprehensive and in-depth analysis and processing of the multi-mode IOT sensor monitoring information of the animal feed mixing control system through the data processing system, thereby realizing accurate monitoring and abnormal identification of the operating state of the control system. By determining the first confidence coefficient of each operating state integrated vector and optimizing the vector, the embodiment of the present invention improves the accuracy and reliability of the operating state integrated vector. This step enables the subsequent knowledge vector collision to more accurately reflect the global sensor monitoring information of the control system, providing more powerful data support for abnormal identification. By performing knowledge vector collision on several operating state optimization vectors, the embodiment of the present invention realizes comprehensive and in-depth analysis and mining of the operating state of the control system. This step not only extracts the key parameters and state element information reflecting the operating state of the control system, but also reveals the inherent connection and change law between these information, providing a more comprehensive and in-depth data basis for abnormal identification. By determining the abnormal identification viewpoint of the target control system based on the global sensor monitoring vector, the embodiment of the present invention realizes timely and accurate identification of the abnormal state of the control system. This step provides operators with timely and accurate abnormal alarms and processing suggestions, which helps to timely discover and deal with potential problems in the control system and improve the stability and operation efficiency of the control system.

[0066] In some further optional embodiments, the first confidence coefficient is determined by the following steps: obtaining a feed production task report, the feed production task report containing mode data corresponding to the sensor equipment of the multi-mode IOT sensor monitoring information; and determining the first confidence coefficient of each operating status integrated vector based on the correlation characteristics between each sensor monitoring category label and the feed production task report.

[0067] Furthermore, the data processing system is also responsible for in-depth analysis and processing of information from multi-mode IOT sensor monitoring to achieve comprehensive and accurate monitoring of the target control system's operating status and abnormality identification. In order to further improve the accuracy of monitoring and identification, the data processing system adopts a more refined method when determining the confidence coefficient of each operating status integrated vector.

[0068] First, the data processing system obtains the feed production task report. This report is an important document in the feed production process. It records various information about the production task in detail, including the goals, requirements, plans, and actual execution of the production task. The feed production task report also contains the mode data corresponding to the sensor equipment of the multi-mode IOT sensor monitoring information. These data reflect the working mode, working status, and working parameters of the sensor equipment in the production task, and are an important basis for the data processing system to determine the confidence coefficient.

[0069] Next, the data processing system determines the first confidence coefficient of each operating state integrated vector based on the correlation characteristics between each sensor monitoring category label and the feed production task report. This step is the key to the data processing system to optimize the operating state integrated vector. Specifically, the data processing system analyzes the correlation between the operating state information represented by each sensor monitoring category label and the relevant information in the feed production task report. This correlation may be reflected in many aspects, such as the consistency between the operating state information and the production task objectives, the matching degree between the operating state information and the actual execution of the production task, and the correlation between the operating state information and the working mode of the sensor equipment.

[0070] By deeply analyzing this correlation feature, the data processing system can evaluate the accuracy and reliability of each operating state integrated vector in reflecting the actual operating state of the control system. Based on this evaluation, the data processing system assigns a first confidence coefficient to each operating state integrated vector. This coefficient is a value between 0 and 1, which reflects the confidence level of the operating state information contained in the vector. When the first confidence coefficient is close to 1, it means that the operating state information contained in the vector is very accurate and reliable; when the first confidence coefficient is close to 0, it means that the operating state information contained in the vector may have a large error or uncertainty.

[0071] After determining the first confidence coefficient, the data processing system optimizes each operating state integrated vector according to this coefficient. The optimization process may include steps such as vector weighting, smoothing, and denoising to improve the quality and accuracy of the vector. Through such optimization, the data processing system can further improve the accuracy and reliability of the global sensor monitoring vector.

[0072] Finally, the data processing system performs knowledge vector collision on the optimized operation state integrated vector to obtain the global sensor monitoring vector of the target control system. This step is the key to the data processing system to achieve abnormal identification. Through knowledge vector collision, the data processing system can cross, compare and fuse different operation state optimization vectors to form a global sensor monitoring vector that more comprehensively and deeply reflects the operation state of the control system. This vector not only contains the key parameters and state element information of the control system under various operation states, but also reflects the inherent connection and change law between these information and elements.

[0073] Based on the global sensor monitoring vector, the data processing system can further determine the abnormality recognition viewpoint of the target control system. By applying the pre-trained abnormality recognition model or algorithm, the data processing system conducts in-depth mining and analysis of the global sensor monitoring vector. The abnormality recognition model will determine whether the control system is in an abnormal state based on the key parameters and state element information in the vector and the inherent connection and change rules between them, and give corresponding abnormality recognition views. These views provide operators with timely and accurate abnormality alarms and processing suggestions, which help improve the stability and operation efficiency of the control system.

[0074] In this way, the embodiment of the present invention performs comprehensive and in-depth analysis and processing of the multi-mode IOT sensor monitoring information of the animal feed mixing control system through the data processing system, thereby realizing accurate monitoring and abnormal identification of the operating state of the control system. By obtaining the feed production task report and analyzing its correlation characteristics with the sensor monitoring category label, the embodiment of the present invention provides a more accurate and reliable basis for determining the first confidence coefficient of the operating state integrated vector. This step makes the determination of the confidence coefficient more in line with the needs and characteristics of the actual production task, and improves the accuracy and practicality of the confidence coefficient. By optimizing the operating state integrated vector according to the first confidence coefficient, the embodiment of the present invention further improves the accuracy and reliability of the global sensor monitoring vector. This step makes the global sensor monitoring vector more accurately reflect the actual operating state of the control system, and provides more powerful data support for abnormal identification. By determining the abnormal identification viewpoint of the target control system based on the global sensor monitoring vector, the embodiment of the present invention realizes timely and accurate identification of the abnormal state of the control system. This step provides operators with timely and accurate abnormal alarms and processing suggestions, which helps to timely discover and deal with potential problems in the control system and improve the stability and operation efficiency of the control system. By combining feed production task reports and multi-mode IOT sensor monitoring information, the embodiment of the present invention achieves comprehensive and accurate monitoring and abnormality identification of the operating status of the animal feed mixing control system.

[0075] In a preferred design idea, after extracting the control system operating status information of a plurality of sensor monitoring category labels from the multi-mode IOT sensor monitoring information and mapping the extracted plurality of control system operating status information into a plurality of operating status embedded vector chains, it also includes: using the number of third embedded vectors as the linear scanning frame size and the number of fourth embedded vectors as the scanning period, performing linear scanning frame-based segmentation on each of the operating status embedded vector chains to obtain a second operating status embedded vector subset corresponding to each of the operating status embedded vector chains; then performing state element integration on each of the first operating status embedded vector subsets, and The global sensor monitoring vector of the target control system is obtained by interacting the integrated several operating state integrated vectors, including: state element integration for each of the first operating state embedded vector subsets, and interacting the integrated several first state element integrated vectors to obtain a first-stage sensor monitoring vector, and state element integration for each of the second operating state embedded vector subsets, and interacting the integrated several second state element integrated vectors to obtain a second-stage sensor monitoring vector; the first-stage sensor monitoring vector is interacted with the second-stage sensor monitoring vector to obtain a global sensor monitoring vector of the target control system.

[0076] In detail, after the data processing system extracts the control system operating status information of several sensor monitoring category labels from the multi-mode IOT sensor monitoring information and maps this information into several operating status embedded vector chains, it will perform further refined processing to improve the monitoring accuracy and comprehensiveness of the control system operating status.

[0077] Specifically, the data processing system sets a third number of embedded vectors as the size of the linear scan frame, and a fourth number of embedded vectors as the scanning period. The selection of these two parameters is based on a deep understanding of the characteristics of the control system's operating status information and monitoring requirements. The size of the linear scan frame determines the length of the vector subset segmented each time, while the scanning period determines the frequency of the segmentation operation. Through such settings, the data processing system can segment each operating state embedded vector chain based on the linear scan frame, thereby obtaining the second operating state embedded vector subset corresponding to the vector chain. The purpose of this step is to cut the originally continuous operating state embedded vector chain into several shorter vector subsets for subsequent more refined processing and analysis.

[0078] Next, the data processing system will perform state element integration on these first operating state embedded vector subsets and second operating state embedded vector subsets. State element integration is a process of extracting key information from vector subsets and integrating them into a single vector. Through this process, the data processing system can extract key elements that reflect the operating state of the control system from each vector subset, and integrate these elements into several first state element integrated vectors and second state element integrated vectors. These integrated vectors not only contain the key information in the vector subsets, but also enhance the representation ability and robustness of the information through specific integration methods (such as weighted averaging, maximum pooling, etc.).

[0079] Subsequently, the data processing system will interact these first state element integrated vectors and second state element integrated vectors to obtain two-stage sensor monitoring vectors. Interaction is a process of fusing, comparing and analyzing multiple vectors. Through this process, the data processing system can further mine the control system operation state information contained in the vector and integrate this information into two-stage sensor monitoring vectors. Specifically, the data processing system will first interact several first state element integrated vectors to obtain the first stage sensor monitoring vector. This vector reflects the operation state of the control system at a certain stage or in a certain aspect, and is obtained by fine processing and analysis of the first operation state embedded vector subset. Then, the data processing system will interact several second state element integrated vectors to obtain the second stage sensor monitoring vector. This vector reflects the operation state of the control system at another stage or in another aspect, and is obtained by performing the same fine processing and analysis on the second operation state embedded vector subset.

[0080] Finally, the data processing system will further interact the sensor monitoring vectors of these two stages to obtain the global sensor monitoring vector of the target control system. This global sensor monitoring vector is a high-dimensional vector that comprehensively and deeply reflects the operating status of the control system. It not only contains the key operating status information of the control system at all stages and aspects, but also further enhances the integrity and accuracy of the information through the interaction process. By analyzing and mining this global sensor monitoring vector, the data processing system can achieve comprehensive and accurate monitoring of the operating status of the control system, and promptly discover and handle potential anomalies or problems.

[0081] The embodiment of the present invention uses a data processing system to perform refined processing and analysis on the multi-mode IOT sensor monitoring information of the animal feed mixing control system, thereby realizing comprehensive and accurate monitoring of the operating state of the control system. By setting the size of the linear scanning frame and the scanning period to segment the operating state embedded vector chain, the embodiment of the present invention can cut the originally continuous operating state information into several shorter vector subsets. Such a processing method makes the subsequent state element integration and interactive analysis more refined and accurate, which helps to improve the precision and accuracy of monitoring. At the same time, by adjusting the size and scanning period of the linear scanning frame, it can flexibly adapt to the characteristics and monitoring requirements of the operating state information of different control systems, thereby improving the versatility and flexibility of the technical solution. By integrating the state elements of the first operating state embedded vector subset and the second operating state embedded vector subset, and interacting the integrated vectors to obtain two-stage sensor monitoring vectors, the embodiment of the present invention realizes multi-stage and multi-faceted monitoring of the operating state of the control system. Such a monitoring method can more comprehensively reflect the operating state of the control system and improve the comprehensiveness and depth of monitoring. At the same time, through the state element integration and interactive process, the operating state information of the control system contained in the vector can be further mined, thereby improving the sensitivity and accuracy of monitoring. By further interacting the two-stage sensor monitoring vectors to obtain a global sensor monitoring vector, the embodiment of the present invention realizes comprehensive and integrated monitoring of the operating status of the control system. This global sensor monitoring vector is a high-dimensional vector that contains key operating status information of the control system at all stages and in all aspects. By analyzing and mining this global sensor monitoring vector, the data processing system can realize comprehensive and accurate monitoring of the operating status of the control system, and promptly discover and handle potential anomalies or problems. This helps to improve the stability and operating efficiency of the control system and ensure the high quality and efficiency of animal feed mixed production.

[0082] In an alternative embodiment, determining the abnormality identification viewpoint of the target control system based on the global sensor monitoring vector includes: inputting the global sensor monitoring vector into a second material processing state analysis algorithm to make an abnormal state decision to obtain a first abnormal state decision score; and determining the abnormal location keywords of the target control system based on the first abnormal state decision score.

[0083] In this embodiment, after obtaining the global sensor monitoring vector, the data processing system will determine the abnormality recognition point of view of the target control system based on this vector.

[0084] Specifically, the data processing system will input the global sensor monitoring vector into the second material processing state analysis algorithm for abnormal state decision making. The second material processing state analysis algorithm is an algorithm specifically used to analyze the operating state of the control system during material processing. It can identify abnormal states in the control system by deeply mining and comparing the high-dimensional data in the global sensor monitoring vector and give a corresponding abnormal state decision score. This score reflects the degree of deviation between the current operating state of the control system and the normal state, and is an important basis for judging whether there is an abnormality in the control system.

[0085] After obtaining the first abnormal state decision score, the data processing system will determine the abnormal location keywords of the target control system based on this score. Abnormal location keywords refer to keywords that can accurately describe the abnormal state or abnormal location of the control system. By analyzing and mining the key information in the first abnormal state decision score, the data processing system can identify the specific factors or locations that cause the control system abnormality and present these factors or locations in the form of keywords. These keywords not only help to quickly locate and control the abnormal state of the system, but also provide an important reference for subsequent abnormal processing and optimization.

[0086] In order to more specifically illustrate how the technical solution of this embodiment is implemented, a practical example can be given. In the animal feed mixing control process, the data processing system collects a large amount of operating status information through multi-mode IOT sensor monitoring, and obtains a global sensor monitoring vector after a series of processing and analysis. Then, the data processing system inputs this vector into the second material processing state analysis algorithm for abnormal state decision-making. After deep mining and comparison of the algorithm, the data processing system obtains a higher first abnormal state decision score, which indicates that there is a large deviation between the current operating state of the control system and the normal state.

[0087] Based on this score, the data processing system began to analyze the specific factors or locations that caused the control system abnormality. Through further mining and comparison, the data processing system found that the main factor causing the abnormality was the excessive stirring speed of the feed mixer, which led to uneven feed mixing and machine overheating. Therefore, the data processing system output the three keywords "excessive stirring speed", "uneven feed mixing" and "machine overheating" as abnormality positioning keywords. These keywords accurately describe the abnormal state and location of the control system, providing an important reference for subsequent abnormality handling and optimization.

[0088] In this way, the global sensor monitoring vector is input into the second material processing state analysis algorithm through the data processing system to make an abnormal state decision, and the abnormal location keywords of the target control system are determined based on the decision score, so as to realize the accurate identification and location of the abnormal state of the animal feed mixing control system. By deeply mining and comparing the global sensor monitoring vector through the second material processing state analysis algorithm, this embodiment can accurately identify the abnormal state in the control system and give the corresponding abnormal state decision score. This score reflects the degree of deviation between the current operating state of the control system and the normal state, and provides an important basis for judging whether the control system is abnormal. At the same time, by adjusting the parameters and models of the algorithm, the accuracy and sensitivity of abnormal state identification can be further improved. By analyzing and mining the key information in the first abnormal state decision score, this embodiment can determine the specific factors or positions that cause the control system to be abnormal, and present these factors or positions in the form of keywords. These abnormal location keywords not only help to quickly locate and control the abnormal state of the system, but also provide an important reference basis for subsequent abnormal processing and optimization. At the same time, by continuously optimizing the keyword extraction method and algorithm, the accuracy and comprehensiveness of abnormal location can be further improved.

[0089] In some possible examples, several of the first material processing state analysis algorithms and the second material processing state analysis algorithms are obtained by collaborative debugging based on algorithm learning examples, and the debugging steps of several of the first material processing state analysis algorithms and the second material processing state analysis algorithms include the following steps: obtaining algorithm learning example information, the algorithm learning example information includes several multi-mode IOT sensor monitoring information examples of control system examples and prior abnormal keywords of each control system example; extracting several control system operation state information examples with sensor monitoring category labels from the multi-mode IOT sensor monitoring information examples, and mapping the extracted several control system operation state information examples into several operation state embedded vector chain examples; using the number of the first embedded vectors as the linear scanning frame size, and the number of the second embedded vectors as the scanning period, for each of the operation state embedded vector chains The example is segmented based on a linear scanning frame to obtain a first operating state embedded vector subset example corresponding to each of the operating state embedded vector chain examples; the first operating state embedded vector subset example corresponding to each of the operating state embedded vector chain examples is loaded into the corresponding first material processing state analysis algorithm for state element integration, and the integrated operating state integration vector example corresponding to each of the operating state embedded vector chain examples obtained by integration is interacted to obtain a global sensor monitoring vector example; the global sensor monitoring vector example is input into the second material processing state analysis algorithm for abnormal state decision to obtain a second abnormal state decision score, and the abnormal judgment error is determined based on the second abnormal state decision score and the corresponding prior abnormal keyword; the algorithm model weights of several of the first material processing state analysis algorithms and the second material processing state analysis algorithm are improved based on the abnormal judgment error.

[0090] In this application scenario, in order to achieve accurate monitoring and abnormal identification of the feed mixing control system, the data processing system will use multiple algorithms, including several first material processing status analysis algorithms and one second material processing status analysis algorithm. These algorithms do not exist in isolation, but complete the task together through collaborative debugging.

[0091] The coordinated debugging of several first material processing state analysis algorithms and second material processing state analysis algorithms is a key step, which ensures the accuracy and effectiveness of the entire data processing process. This debugging step is based on algorithm learning examples and specifically includes the following key links:

[0092] First, the data processing system needs to obtain example information for algorithm learning. This information is the basis for algorithm debugging, including examples of multi-mode IOT sensor monitoring information for several control system examples and a priori abnormal keywords for each control system example. The multi-mode IOT sensor monitoring information examples are collected from actual feed mixing control systems and contain data monitored by various sensors, such as temperature, humidity, pressure, etc. A priori abnormal keywords are known keywords related to the abnormal state of the control system, such as "mixer failure", "temperature sensor abnormality", etc.

[0093] Next, the data processing system extracts several control system operating status information examples with sensor monitoring category labels from the multi-mode IOT sensor monitoring information examples. The purpose of this step is to convert the original sensor monitoring data into a form that is easier to process and analyze. Then, the data processing system maps these extracted control system operating status information examples into several operating status embedding vector chain examples. The operating status embedding vector chain is a high-dimensional vector that can more comprehensively describe the operating status of the control system.

[0094] Subsequently, the data processing system will perform linear scan frame segmentation on each running state embedded vector chain example with the number of first embedded vectors as the linear scan frame size and the number of second embedded vectors as the scan period. The purpose of this step is to segment the running state embedded vector chain into smaller subsets for subsequent processing and analysis. Through the segmentation of the linear scan frame, the data processing system can obtain the first running state embedded vector subset example corresponding to each running state embedded vector chain example.

[0095] Next, the data processing system will load the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example into the corresponding first material processing state analysis algorithm for state element integration. State element integration is a process of fusing key information from multiple embedded vectors into an integrated vector. Through this step, the data processing system can obtain an operating state integrated vector example corresponding to each operating state embedded vector chain example. Then, the data processing system interacts these operating state integrated vector examples to obtain a global sensor monitoring vector example. The global sensor monitoring vector is a high-dimensional vector that integrates multiple operating state integrated vectors and can more comprehensively reflect the overall operating state of the control system.

[0096] Finally, the data processing system will input the global sensor monitoring vector example into the second material processing state analysis algorithm for abnormal state decision-making. The second material processing state analysis algorithm is an algorithm specifically used to identify abnormal states of control systems. Through this step, the data processing system can obtain the second abnormal state decision score, which reflects the degree of deviation between the current operating state of the control system and the normal state. Then, the data processing system will determine the abnormality discrimination error based on the second abnormal state decision score and the corresponding prior abnormal keywords. The abnormality discrimination error is an important indicator for measuring the accuracy of the algorithm in identifying abnormal states. By continuously reducing the abnormality discrimination error, the data processing system can gradually optimize the algorithm model weights of several first material processing state analysis algorithms and the second material processing state analysis algorithm, thereby improving the accuracy and effectiveness of the entire data processing process.

[0097] In summary, the coordinated debugging of several first material processing state analysis algorithms and second material processing state analysis algorithms is a complex and delicate process. Through this process, the data processing system can gradually optimize the model weights of the algorithm and improve the accuracy of monitoring the operating state of the entire feed mixing control system and identifying abnormalities.

[0098] In the next step, after extracting several control system operation status information examples with sensor monitoring category labels from the multi-mode IOT sensor monitoring information examples and mapping the extracted several control system operation status information examples into several operation status embedded vector chain examples, it also includes: using the third number of embedded vectors as the linear scanning frame size and the fourth number of embedded vectors as the scanning period, performing linear scanning frame-based segmentation on each of the operation status embedded vector chain examples to obtain a second operation status embedded vector subset example corresponding to each of the operation status embedded vector chain examples; then loading the first operation status embedded vector subset example corresponding to each of the operation status embedded vector chain examples into the corresponding first material processing state analysis algorithm for state element integration, and interacting the operation status integrated vector examples corresponding to each of the operation status embedded vector chain examples obtained by the integration Obtaining a global sensor monitoring vector example, including: loading a first operating state embedded vector subset example corresponding to each operating state embedded vector chain example into the corresponding first material processing state analysis algorithm for state element integration, and interacting the first state element integration vector example corresponding to each operating state embedded vector chain example obtained by integration to obtain a first-stage sensor monitoring vector example; loading a second operating state embedded vector subset example corresponding to each operating state embedded vector chain example into the corresponding first material processing state analysis algorithm for state element integration, and interacting the second state element integration vector example corresponding to each operating state embedded vector chain example obtained by integration to obtain a second-stage sensor monitoring vector example; interacting the first-stage sensor monitoring vector example with the second-stage sensor monitoring vector example to obtain a global sensor monitoring vector example of the control system example.

[0099] Based on this embodiment, in order to achieve accurate monitoring and abnormal identification of the feed mixing control system, the data processing system will use multiple algorithms and debug these algorithms in a coordinated manner. During the debugging process, the data processing system will extract several control system operation status information examples of sensor monitoring category labels from the multi-mode IOT sensor monitoring information examples, and map these extracted information examples into several operation status embedded vector chain examples. This step is the basis of data processing, which ensures the accuracy and effectiveness of subsequent analysis.

[0100] Next, the data processing system will further segment each operating state embedded vector chain example based on a linear scan frame. Specifically, the data processing system will segment each operating state embedded vector chain example using the number of the third embedded vectors as the linear scan frame size and the number of the fourth embedded vectors as the scan period. Through this step, the data processing system can obtain a second operating state embedded vector subset example corresponding to each operating state embedded vector chain example. These subset examples are the basis for subsequent state element integration and global sensor monitoring vector generation.

[0101] After obtaining the first operating state embedded vector subset example and the second operating state embedded vector subset example, the data processing system will load these subset examples into the corresponding first material processing state analysis algorithm for state element integration. State element integration is a process of fusing key information in multiple embedded vectors into an integrated vector. Through this step, the data processing system can obtain the first state element integrated vector example and the second state element integrated vector example corresponding to each operating state embedded vector chain example.

[0102] Subsequently, the data processing system will interact with the first state element integrated vector example corresponding to each operation state embedding vector chain example obtained by integration, and obtain the first stage sensor monitoring vector example. Similarly, the data processing system will also interact with the second state element integrated vector example corresponding to each operation state embedding vector chain example obtained by integration, and obtain the second stage sensor monitoring vector example. The two stage sensor monitoring vector examples respectively represent the operation state information of the control system in different time periods or different states.

[0103] Finally, the data processing system will interact the first-stage sensor monitoring vector example with the second-stage sensor monitoring vector example to obtain the global sensor monitoring vector example of the control system example. The global sensor monitoring vector is a high-dimensional vector that integrates multiple operating state integrated vectors, which can more comprehensively reflect the overall operating state of the control system. Through this step, the data processing system completes the comprehensive monitoring and representation of the operating state of the entire feed mixing control system.

[0104] In this way, by extracting several control system operation status information examples of sensor monitoring category labels from the multi-mode IOT sensor monitoring information examples, and mapping these extracted information examples into several operation status embedded vector chain examples, the data processing system can more comprehensively capture and analyze the operation status information of the control system. This helps to improve the accuracy of monitoring and control of the entire feed mixing control process, thereby ensuring the quality and efficiency of feed mixing. By using the number of third embedded vectors as the linear scanning frame size and the number of fourth embedded vectors as the scanning period, each operation status embedded vector chain example is segmented based on the linear scanning frame, and the data processing system can more finely divide the operation status of the control system, thereby more accurately capturing and analyzing the dynamic changes of the control system. This helps to improve the accuracy of real-time monitoring and abnormal identification of the entire feed mixing control process. By loading the first operation status embedded vector subset example and the second operation status embedded vector subset example corresponding to each operation status embedded vector chain example into the corresponding first material processing state analysis algorithm for state element integration, and interacting the integrated vector examples to obtain the global sensor monitoring vector example, the data processing system can more comprehensively integrate and analyze the operation status information of the control system. This helps improve the overall monitoring of the entire feed mixing control system and the accuracy of abnormal identification, so that potential problems can be discovered and dealt with in a timely manner to avoid adverse effects on the feed mixing process.

[0105] In some other optional embodiments, each of the first material processing state analysis algorithms includes a residual connection branch and a state element integration branch, and the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example is loaded into the corresponding first material processing state analysis algorithm for state element integration, and the operating state integrated vector example corresponding to each integrated operating state embedded vector chain example is interacted to obtain a global sensor monitoring vector example, including: performing residual connection processing on the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example according to the residual connection branch to obtain a first residual connection vector example; performing state element integration on the first residual connection vector example corresponding to each operating state embedded vector chain example according to the state element integration branch to obtain an operating state integrated vector example corresponding to each operating state embedded vector chain example; interacting the operating state integrated vector example corresponding to each operating state embedded vector chain example to obtain a global sensor monitoring vector example.

[0106] In detail, each of the first material processing status analysis algorithms also includes a distribution feature mapping branch. After performing residual connection processing on the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example according to the residual connection branch to obtain the first residual connection vector example, it also includes: performing distribution feature mapping on the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example according to the distribution feature mapping branch to obtain a first operating state distribution heat map example; then performing state element integration on the first residual connection vector example corresponding to each operating state embedded vector chain example according to the state element integration branch to obtain an operating state integrated vector example corresponding to each operating state embedded vector chain example, including: performing state element integration on the first operating state distribution heat map example corresponding to each operating state embedded vector chain example according to the state element integration branch to obtain an operating state integrated vector example corresponding to each operating state embedded vector chain example.

[0107] In this application scenario, in order to achieve accurate monitoring and abnormal identification of the feed mixing control system, the data processing system will use multiple algorithms and coordinately debug these algorithms. In other optional embodiments, each first material processing state analysis algorithm is designed to include a residual connection branch and a state element integration branch, such a design enables the algorithm to better capture and analyze the operating state information of the control system.

[0108] Specifically, when the data processing system needs to load the first operating state embedding vector subset example corresponding to each operating state embedding vector chain example into the corresponding first material processing state analysis algorithm for state element integration, it will first perform residual connection processing on the first operating state embedding vector subset example corresponding to each operating state embedding vector chain example based on the residual connection branch. Residual connection is a commonly used technique in deep learning. It enables the network to learn the difference between input and output by directly adding the input to the output, thereby better capturing the characteristics of the data. Here, through residual connection processing, the data processing system can obtain the first residual connection vector example, which retains the key information of the original operating state embedding vector and highlights the residual features therein.

[0109] Next, the data processing system will perform state element integration on the first residual connection vector example corresponding to each running state embedding vector chain example according to the state element integration branch. State element integration is a process of fusing key information from multiple embedding vectors into an integrated vector. Here, through state element integration, the data processing system can obtain an example of the running state integrated vector corresponding to each running state embedding vector chain example. This integrated vector example combines the key information of the original running state embedding vector and the residual connection vector, and more comprehensively reflects the running state of the control system.

[0110] Furthermore, each first material processing state analysis algorithm also includes a distribution feature mapping branch. After obtaining the first residual connection vector example, the data processing system will perform distribution feature mapping on the first operating state embedding vector subset example corresponding to each operating state embedding vector chain example according to the distribution feature mapping branch. Distribution feature mapping is a technology that maps data to a high-dimensional space and analyzes its distribution characteristics. Here, through distribution feature mapping, the data processing system can obtain an example of a first operating state distribution heat map. This heat map example intuitively shows the distribution characteristics of the control system operating state, providing richer information for subsequent state element integration.

[0111] Finally, the data processing system will perform state element integration on the first operation state distribution heat map example corresponding to each operation state embedding vector chain example according to the state element integration branch, and obtain the operation state integration vector example corresponding to each operation state embedding vector chain example. Subsequently, the data processing system will interact the operation state integration vector example corresponding to each operation state embedding vector chain example to obtain the global sensor monitoring vector example. The global sensor monitoring vector is a high-dimensional vector that integrates multiple operation state integration vectors, which can more comprehensively reflect the overall operation state of the control system.

[0112] In this way, by introducing the residual connection branch and the state element integration branch, each first material processing state analysis algorithm can better capture and analyze the operating state information of the control system. The residual connection enables the algorithm to learn the difference between the input and the output, so as to more accurately capture the characteristics of the data; while the state element integration can fuse the key information in multiple embedded vectors into an integrated vector, which more comprehensively reflects the operating state of the control system. This helps to improve the accuracy of monitoring and control of the entire feed mixing control process. By introducing the distribution feature mapping branch, each first material processing state analysis algorithm can further enrich the representation of the operating state information. The distribution feature mapping enables the algorithm to map the data to a high-dimensional space and analyze its distribution characteristics, so as to obtain more intuitive and comprehensive operating state information. This helps to improve the accuracy of real-time monitoring and abnormal identification of the entire feed mixing control process. By performing residual connection processing, distribution feature mapping and state element integration on the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example, the data processing system can more comprehensively integrate and analyze the operating state information of the control system. This helps improve the overall monitoring of the entire feed mixing control system and the accuracy of abnormal identification, so that potential problems can be discovered and dealt with in a timely manner to avoid adverse effects on the feed mixing process.

[0113] In some independent embodiments, the distribution feature mapping branch is used to perform distribution feature mapping on the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example to obtain a first operating state distribution heat map example, including: determining the monitoring category influence factor corresponding to each operating state embedded vector subset example; and performing distribution feature mapping on the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example according to the monitoring category influence factor to obtain a first operating state distribution heat map example.

[0114] It is understandable that in order to achieve accurate monitoring of the feed mixing control system, the data processing system will use multiple algorithms and debug these algorithms in a coordinated manner. Among them, the distribution feature mapping branch is an important component in the data processing system. It is responsible for the distribution feature mapping of the operating state embedding vector, thereby obtaining a heat map reflecting the operating state of the control system.

[0115] In some independent embodiments, the data processing system performs distribution feature mapping on the first operating state embedding vector subset example corresponding to each operating state embedding vector chain example according to the distribution feature mapping branch to obtain a first operating state distribution heat map example.

[0116] First, the data processing system needs to determine the monitoring category influencing factors corresponding to each operating state embedding vector subset example. The monitoring category influencing factor is an important concept, which refers to the factors that have a significant impact on specific monitoring categories (such as mixing uniformity, raw material input, etc.) during the feed mixing control process. These factors may include the parameter settings of the control system, the physical properties of the raw materials, the operating status of the mixing equipment, etc. By determining the monitoring category influencing factors, the data processing system can more accurately grasp the operating status of the control system and provide a targeted basis for subsequent distribution feature mapping.

[0117] After determining the monitoring category influencing factors, the data processing system will perform distribution feature mapping on the first operating state embedded vector subset examples corresponding to each operating state embedded vector chain example based on these influencing factors. This process actually combines the key information in the operating state embedded vector with the monitoring category influencing factors, and transforms it into a heat map that can intuitively reflect the operating state of the control system through certain mathematical transformations and mapping relationships. In this heat map, different colors or brightness represent different operating states or the degree of quality of the operating states, thereby achieving an intuitive display of the operating state of the control system.

[0118] Specifically, the data processing system will perform weighted processing on the operating state embedding vector or select a specific mapping function according to the different influencing factors of the monitoring category. For example, for the monitoring category of mixing uniformity, the data processing system may pay more attention to the influencing factors related to the mixing equipment speed, mixing time, etc., and give these factors greater weights when mapping the distribution characteristics. For the monitoring category of raw material input, the data processing system may pay more attention to the influencing factors related to the raw material input speed and input amount.

[0119] Through such processing, the data processing system can obtain the first operating state distribution heat map example corresponding to each operating state embedded vector chain example. This heat map example not only intuitively shows the operating state of the control system, but also reflects the differences and advantages and disadvantages between different operating states through the depth of color and shape changes. This is of great significance for real-time monitoring and abnormal identification of the feed mixing control process.

[0120] In practical applications, the data processing system will further analyze and process the obtained operation status distribution heat map examples. For example, it can identify whether the operation status of the control system has changed by comparing the heat map examples at different time points; or locate possible faults or problems by analyzing the abnormal areas in the heat map. This information is of great guiding significance for the optimization and adjustment of the feed mixing control process.

[0121] In this way, by introducing the concept of monitoring category influencing factors and combining the distribution feature mapping branch to carry out targeted processing and analysis of the operating state embedding vector, a heat map example that can intuitively reflect the operating state of the control system is obtained. This technical solution not only improves the monitoring accuracy of the feed mixing control process, but also provides strong support for subsequent abnormality identification and fault location.

[0122] In some other independent embodiments, the state element integration branch includes a feature focusing module, a first knowledge quantification collision module, a feature activation module, and a second knowledge quantification collision module, and the state element integration is performed on the first residual connection vector example corresponding to each of the running state embedded vector chain examples according to the state element integration branch to obtain the running state integrated vector example corresponding to each of the running state embedded vector chain examples, including: performing a feature focusing operation on the first residual connection vector example corresponding to each of the running state embedded vector chain examples according to the feature focusing module to obtain a first equipment loss focusing vector; performing equipment risk quantification mapping on the first equipment loss focusing vector and the first residual connection vector example corresponding to each of the running state embedded vector chain examples according to the first knowledge quantification collision module to obtain a first equipment risk quantification mapping variable; activating the first equipment risk quantification mapping variable corresponding to each of the running state embedded vector chain examples according to the feature activation module to obtain a first activation result; performing equipment risk quantification mapping on the first equipment risk quantification mapping variable and the first activation result corresponding to each of the running state embedded vector chain examples according to the second knowledge quantification collision module to obtain an running state integrated vector example corresponding to each of the running state embedded vector chain examples.

[0123] In this embodiment, in order to more accurately monitor and evaluate the operating status of the feed mixing control system, the data processing system uses a series of advanced algorithms and models, among which the state factor integration branch is an important component.

[0124] In detail, the state element integration branch is designed to include a feature focusing module, a first knowledge quantification collision module, a feature activation module and a second knowledge quantification collision module. Such a design enables the data processing system to more comprehensively integrate and analyze the operating state information of the control system, thereby achieving more accurate monitoring and evaluation of the feed mixing process.

[0125] Specifically, when the data processing system needs to perform state element integration on the first residual connection vector example corresponding to each running state embedding vector chain example, it will first perform a feature focusing operation on the first residual connection vector example corresponding to each running state embedding vector chain example according to the feature focusing module. Feature focusing is a data processing technology that performs specific transformations and screening on the input vector so that the output vector can more prominently reflect the key features in the input vector. Here, through the feature focusing operation, the data processing system can obtain the first device loss focus vector, which is more focused on reflecting the information of the key state element of device loss.

[0126] Next, the data processing system will perform equipment risk quantification mapping on the first equipment loss focus vector and the first residual connection vector example corresponding to each operating state embedded vector chain example based on the first knowledge quantification collision module. Knowledge quantification collision is a technology that combines knowledge with data and converts it into quantitative variables through a specific mapping relationship. Here, through the equipment risk quantification mapping, the data processing system can obtain the first equipment risk quantification mapping variable, which integrates the key information in the equipment loss focus vector and the residual connection vector, and more specifically reflects the quantitative characteristics of the key state element of equipment risk.

[0127] Subsequently, the data processing system activates the first equipment risk quantification mapping variable corresponding to each operating state embedding vector chain example according to the feature activation module. Feature activation is a technique that performs nonlinear transformation on the input vector through a specific activation function, which can enhance the key features in the input vector and suppress unimportant features. Here, through the feature activation operation, the data processing system can obtain the first activation result, which more prominently reflects the key features in the equipment risk quantification mapping variable.

[0128] Finally, the data processing system will perform equipment risk quantification mapping on the first equipment risk quantification mapping variable and the first activation result corresponding to each operating state embedded vector chain example according to the second knowledge quantification collision module. Through this step, the data processing system can obtain an operating state integrated vector example corresponding to each operating state embedded vector chain example. This integrated vector example integrates all the previous key information and more comprehensively reflects the overall operating state of the control system.

[0129] In this way, by introducing the feature focusing module, the first knowledge quantification collision module, the feature activation module and the second knowledge quantification collision module, the data processing system can more comprehensively integrate and analyze the operating status information of the control system. This enables the data processing system to more accurately capture and reflect the key state elements of the control system, such as equipment loss and equipment risk, thereby improving the accuracy of monitoring and evaluation of the entire feed mixing control process. Through the feature focusing operation and equipment risk quantification mapping, the data processing system can focus more on the information reflecting the key state element of equipment loss and convert it into specific quantitative variables. This enables the data processing system to more specifically evaluate and monitor the loss of equipment, timely discover and deal with potential equipment problems, and avoid adverse effects on the feed mixing process. Through the feature activation operation, the data processing system can enhance the key features in the equipment risk quantification mapping variables and suppress unimportant features. This enables the data processing system to more accurately identify and evaluate equipment risks, providing strong support for subsequent fault prediction and maintenance plans. Through the equipment risk quantification mapping of the second knowledge quantification collision module, the data processing system can obtain the operating state integrated vector example corresponding to each operating state embedded vector chain example. This integrated vector example integrates all the previous key information and more comprehensively reflects the overall operating state of the control system. This enables the data processing system to more comprehensively monitor and evaluate the operating status of the feed mixing control system, providing strong support for optimizing the feed mixing process and improving production efficiency.

[0130] Further, Figure 2 The structural block diagram of the data processing system 300 is shown, which includes: a memory 310 for storing program instructions and data; a processor 320 for coupling with the memory 310 and executing the instructions in the memory 310 to implement the above method.

[0131] Furthermore, a computer storage medium is provided, comprising instructions, which implement the above method when executed on a processor.

[0132] In several embodiments provided in the embodiments of the present invention, it should be understood that the disclosed apparatus and method can also be implemented in other ways. The apparatus and method embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the apparatus, method and computer program product according to multiple embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, program segment or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0133] In addition, the functional modules in the various embodiments of the present invention may be integrated together to form an independent part, or each module may exist independently, or two or more modules may be integrated to form an independent part.

[0134] If the function is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code. It should be noted that in this article, the term "include", "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements that are not explicitly listed, or also includes elements inherent to such process, method, article or device. Without more constraints, an element defined by the phrase "comprising a..." does not exclude the existence of other identical elements in the process, method, article or apparatus comprising the element.

[0135] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A data processing method based on an animal feed mixing control system, characterized in that: Applied to a data processing system, the method comprises: Acquire multi-mode IOT sensor monitoring information of a target control system; wherein the target control system is an animal feed mixing control system, and the multi-mode IOT sensor monitoring information refers to various types of monitoring data collected in real time by installed IOT sensors in the animal feed mixing control system; the IOT sensors include temperature sensors, humidity sensors, and flow sensors; the temperature sensor provides real-time temperature data inside the mixer, the humidity sensor provides humidity data of the raw materials, and the flow sensor provides feed volume data of the raw materials; Extracting a plurality of control system operation status information of sensor monitoring category labels from the multi-mode IOT sensor monitoring information, and mapping the extracted plurality of control system operation status information into a plurality of operation status embedding vector chains; Taking the number of first embedded vectors as the linear scan frame size and the number of second embedded vectors as the scan period, segmenting each of the operating state embedded vector chains based on the linear scan frame to obtain a first operating state embedded vector subset corresponding to each of the operating state embedded vector chains; Performing state element integration on each of the first operating state embedded vector subsets, and interacting a plurality of integrated operating state integrated vectors to obtain a global sensor monitoring vector of the target control system; An abnormality recognition perspective of the target control system is determined based on the global sensor monitoring vector.

2. The method according to claim 1, characterized in that The step of integrating the state elements of each of the first operation state embedded vector subsets and interacting the integrated operation state integrated vectors to obtain the global sensor monitoring vector of the target control system includes: Determine a first material processing state analysis algorithm corresponding to each of the first operating state embedding vector subsets; Inputting the first operating state embedding vector subset corresponding to each operating state embedding vector chain into the corresponding first material processing state analysis algorithm for state element integration to obtain an operating state integration vector corresponding to each operating state embedding vector chain; A knowledge vector collision is performed on a plurality of the operating state integrated vectors to obtain a global sensor monitoring vector of the target control system.

3. The method according to claim 2, characterized in that The performing knowledge vector collision on a plurality of the operation state integrated vectors to obtain a global sensor monitoring vector of the target control system includes: Determining a first confidence coefficient for each of the operating state integrated vectors; Optimizing each of the operating state integrated vectors according to the first confidence coefficient to obtain a plurality of operating state optimization vectors; A knowledge vector collision is performed on the plurality of operation state optimization vectors to obtain a global sensor monitoring vector of the target control system.

4. The method according to claim 3, characterized in that: The first confidence coefficient is determined by the following steps: Obtaining a feed production task report, wherein the feed production task report includes mode data corresponding to the sensor device of the multi-mode IOT sensor monitoring information; According to the correlation characteristics between each of the sensor monitoring category labels and the feed production task report, a first confidence coefficient of each of the operating status integrated vectors is determined.

5. The method according to claim 1, characterized in that After extracting a plurality of control system operation status information of sensor monitoring category labels from the multi-mode IOT sensor monitoring information and mapping the extracted plurality of control system operation status information into a plurality of operation status embedded vector chains, the method further includes: using the number of third embedded vectors as the linear scanning frame size and the number of fourth embedded vectors as the scanning period, performing linear scanning frame-based segmentation on each of the operation status embedded vector chains to obtain a second operation status embedded vector subset corresponding to each of the operation status embedded vector chains; The state element integration is performed on each of the first operating state embedded vector subsets, and several integrated operating state integrated vectors are interacted to obtain the global sensor monitoring vector of the target control system, including: state element integration is performed on each of the first operating state embedded vector subsets, and several first state element integrated vectors are interacted to obtain the first stage sensor monitoring vector, and state element integration is performed on each of the second operating state embedded vector subsets, and several second state element integrated vectors are interacted to obtain the second stage sensor monitoring vector; the first stage sensor monitoring vector is interacted with the second stage sensor monitoring vector to obtain the global sensor monitoring vector of the target control system.

6. The method according to claim 2, characterized in that The determining of the abnormality recognition viewpoint of the target control system based on the global sensor monitoring vector includes: Inputting the global sensor monitoring vector into a second material processing state analysis algorithm to perform abnormal state decision making, and obtaining a first abnormal state decision score; Determine an abnormality location keyword of the target control system based on the first abnormal state decision score; Among them, several of the first material processing state analysis algorithms and the second material processing state analysis algorithms are obtained by collaborative debugging based on algorithm learning examples, and the debugging steps of several of the first material processing state analysis algorithms and the second material processing state analysis algorithms include the following steps: Acquire algorithm learning example information, wherein the algorithm learning example information includes multi-mode IOT sensor monitoring information examples of several control system examples and a priori abnormal keywords of each control system example; Extracting a plurality of control system operation status information examples of sensor monitoring category labels from the multi-mode IOT sensor monitoring information examples, and mapping the extracted plurality of control system operation status information examples into a plurality of operation status embedding vector chain examples; Taking the first number of embedded vectors as a linear scan frame size and the second number of embedded vectors as a scan period, segmenting each of the running state embedded vector chain examples based on a linear scan frame to obtain a first running state embedded vector subset example corresponding to each of the running state embedded vector chain examples; Loading the first operating state embedded vector subset example corresponding to each of the operating state embedded vector chain examples into the corresponding first material processing state analysis algorithm for state element integration, and interacting the operating state integrated vector examples corresponding to each of the operating state embedded vector chain examples obtained by integration to obtain a global sensor monitoring vector example; Inputting the global sensor monitoring vector example into a second material processing state analysis algorithm to perform abnormal state decision, obtain a second abnormal state decision score, and determine an abnormal state discrimination error according to the second abnormal state decision score and a corresponding prior abnormal keyword; The algorithm model weights of a plurality of the first material processing state analysis algorithms and the second material processing state analysis algorithms are improved according to the abnormal discrimination error.

7. The method according to claim 6, characterized in that After extracting a plurality of control system operation status information examples of sensor monitoring category labels from the multi-mode IOT sensor monitoring information examples, and mapping the extracted plurality of control system operation status information examples into a plurality of operation status embedded vector chain examples, the method further includes: using the number of third embedded vectors as the linear scanning frame size, and using the number of fourth embedded vectors as the scanning period, performing linear scanning frame-based segmentation on each of the operation status embedded vector chain examples, and obtaining a second operation status embedded vector subset example corresponding to each of the operation status embedded vector chain examples; The method of loading the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example into the corresponding first material processing state analysis algorithm for state element integration, and interacting the operating state integrated vector example corresponding to each operating state embedded vector chain example obtained by integration to obtain a global sensor monitoring vector example includes: loading the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example into the corresponding first material processing state analysis algorithm for state element integration, and interacting the first state element integrated vector example corresponding to each operating state embedded vector chain example obtained by integration to obtain a first-stage sensor monitoring vector example; loading the second operating state embedded vector subset example corresponding to each operating state embedded vector chain example into the corresponding first material processing state analysis algorithm for state element integration, and interacting the second state element integrated vector example corresponding to each operating state embedded vector chain example obtained by integration to obtain a second-stage sensor monitoring vector example; interacting the first-stage sensor monitoring vector example with the second-stage sensor monitoring vector example to obtain a global sensor monitoring vector example of the control system example.

8. The method according to claim 6, characterized in that Each of the first material processing state analysis algorithms includes a residual connection branch and a state element integration branch, and the first operating state embedding vector subset example corresponding to each of the operating state embedding vector chain examples is loaded into the corresponding first material processing state analysis algorithm for state element integration, and the operating state integration vector examples corresponding to each of the integrated operating state embedding vector chain examples are interacted to obtain a global sensor monitoring vector example, including: Performing residual connection processing on a first running state embedding vector subset example corresponding to each running state embedding vector chain example according to the residual connection branch to obtain a first residual connection vector example; Performing state element integration on the first residual connection vector example corresponding to each of the running state embedding vector chain examples according to the state element integration branch to obtain a running state integration vector example corresponding to each of the running state embedding vector chain examples; Interacting the operation state integration vector example corresponding to each of the operation state embedding vector chain examples to obtain a global sensor monitoring vector example; Among them, each of the first material processing status analysis algorithms also includes a distribution feature mapping branch. After the residual connection branch performs residual connection processing on the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example to obtain the first residual connection vector example, it also includes: performing distribution feature mapping on the first operating state embedded vector subset example corresponding to each operating state embedded vector chain example according to the distribution feature mapping branch to obtain a first operating state distribution heat map example; performing state element integration on the first residual connection vector example corresponding to each operating state embedded vector chain example according to the state element integration branch to obtain an operating state integrated vector example corresponding to each operating state embedded vector chain example, including: performing state element integration on the first operating state distribution heat map example corresponding to each operating state embedded vector chain example according to the state element integration branch to obtain an operating state integrated vector example corresponding to each operating state embedded vector chain example.

9. A data processing system, characterized in that: include: Memory, used to store program instructions and data; A processor, coupled to a memory, and configured to execute instructions in the memory to implement the method according to any one of claims 1 to 8.

10. A computer storage medium, characterized in that: Contains instructions, which, when executed on a processor, implement the method according to any one of claims 1 to 8.

Images