CN118468164B - A big data real-time analysis and processing device based on deep learning - Google Patents

Abstract

The present invention relates to the field of big data analysis technology, and specifically to a big data real-time analysis and processing device based on deep learning, comprising: a plurality of data acquisition units and a plurality of acquisition and identification units, a plurality of data transmission units and a plurality of transmission identification units, a plurality of data analysis units and a plurality of analysis and identification units, a plurality of data processing units and a plurality of processing and identification units, and a data visualization module, which is used to receive the analysis results of first analysis data and the processing results of first processing data. The present invention ensures that the urgent data that needs to be urgently analyzed and extracted can be efficiently transmitted for analysis and processing during the process of big data real-time analysis and processing, thereby ensuring that the urgent data in the process of big data real-time analysis and processing can be quickly extracted and fed back.

Description

A big data real-time analysis and processing device based on deep learning

Technical Field

The present invention relates to the technical field of big data analysis, and in particular to a big data real-time analysis and processing device based on deep learning.

Background Art

In the energy field, a wind power system, also known as a wind power generation system, is a system that uses wind power to generate electricity. Fault monitoring during the operation of a wind power system is crucial. If fault information during the operation of a wind power system is not detected in a timely manner, it may lead to reduced wind power conversion efficiency, affecting power generation; causing downtime or even damage, affecting the normal operation of the wind turbine; and even causing reduced stability and safety of the wind turbine, resulting in safety accidents.

A large amount of data is generated during the operation of the wind power system, mainly including environmental data (wind speed, wind direction, temperature, humidity, etc.), performance data (power generation, power curve, etc.), operation data (speed, current, voltage, frequency, turbulence intensity, etc.), and monitoring data (pressure, vibration, oil level, etc.). These data are closely related to the operating status of the wind power system, and these data can be used to diagnose whether there is a fault in the current operation of the wind power system. However, due to the large amount of data, traditional big data processing technology is difficult to meet the needs of wind power system data processing and analysis. It is usually necessary to use a big data analysis and processing system for fault diagnosis, aiming to extract useful information from the massive wind power system operation data and provide support for diagnostic decisions.

However, the existing big data analysis and processing system lacks the identification and analysis of data types in the process of wind power system operation data collection, and usually transmits the collected or pre-processed wind power system operation data according to the preset link. However, there may be some emergency data in the wind power system operation data, such as: related data for prompts, faults, and warnings. These data usually have different levels of urgency due to their individual differences. If traditional collection, analysis, processing and transmission methods are used for these emergency data, it is impossible to ensure that these emergency data can be quickly extracted and analyzed, resulting in the inability to obtain wind power system operation faults in a timely manner and to generate troubleshooting strategies in a timely manner.

In summary, there is an urgent need for a big data real-time analysis and processing device based on deep learning that can quickly classify and identify wind power system operation data and ensure that the emergency data existing in the wind power system operation data can be efficiently analyzed and processed.

Summary of the invention

To solve the above problems, the present invention provides a big data real-time analysis and processing device based on deep learning, which is used to classify and identify the collected wind power system operation data, and reasonably allocate transmission, analysis and processing paths for emergency data, to ensure that the emergency data can be efficiently analyzed and processed, and to diagnose the operation status of the wind power system.

In order to achieve the above object, the technical solution of the present invention is as follows: A big data real-time analysis and processing device based on deep learning, comprising:

A plurality of data acquisition units and a plurality of acquisition and identification units, wherein the data acquisition units are used to acquire wind power system operation data, the acquisition and identification units are used to classify and identify the wind power system operation data acquired by the data acquisition units, and the data acquisition units are further used to identify the acquired wind power system operation data according to the classification and identification results of the wind power system operation data to generate first acquisition data;

A plurality of data transmission units and a plurality of transmission identification units, the data transmission unit is used to receive the first collected data, the transmission identification unit is used to classify and identify the first collected data received by the data transmission unit, and the data transmission unit is further used to identify the first collected data according to the classification and identification result of the first collected data to generate first transmission data;

A plurality of data analysis units and a plurality of analysis and identification units, the data analysis unit is used to receive the first transmission data, the analysis and identification unit is used to classify and identify the first transmission data received by the data analysis unit, the data analysis unit is used to identify the first transmission data according to the classification and identification result of the first transmission data, generate first analysis data, and analyze the first analysis data;

A plurality of data processing units and a plurality of processing identification units, the data processing units are used to receive first analysis data, the processing identification units are used to classify and identify the first analysis data received by the data processing units, the data processing units are used to identify the first analysis data according to the classification and identification results of the first analysis data, generate first processing data, and process the first processing data;

a data visualization module, configured to receive and visualize an analysis result of the first analysis data and a processing result of the first processing data, and diagnose an operating state of the wind power system based on the analysis result of the first analysis data and the processing result of the first processing data;

The collection and identification unit, the transmission and identification unit, the analysis and identification unit and the processing and identification unit are used to receive preset classification and identification rules, and identify the urgency of the collected wind power system operation data, the first collection data, the first transmission data and the first analysis data according to the classification and identification rules respectively. When the collected wind power system operation data, the first collection data, the first transmission data or the first analysis data is identified as emergency data, the corresponding identification unit generates data transmission path information, and the data collection module, the data transmission module, the data analysis module and the data processing module are used to transmit emergency data according to the transmission path information.

Furthermore, the collection and identification unit is constructed based on a deep learning model. The collection and identification unit is used to classify the collected wind power system operation data according to preset classification and identification rules, and to judge the urgency of the collected wind power system operation data based on a preset urgency characteristic value. When the urgency characteristic value is greater than a threshold value, the collection and identification unit identifies the collected wind power system operation data as collection emergency data.

Furthermore, it also includes a monitoring module, which is connected to the collection and identification unit, the transmission identification unit, the analysis and identification unit, the processing identification unit and the data transmission unit signal. The monitoring module is used to obtain the collection emergency data identified by the collection and identification unit, and monitor whether the data transmission unit receives the corresponding collection emergency data. If the collection emergency data is not received by the data transmission unit, the corresponding collection emergency data is transmitted to the data transmission unit whose operating load is lower than the threshold.

Further, the monitoring module is also signal-connected to the data analysis unit and the data processing unit;

The monitoring module is used to monitor whether the first transmission data corresponding to the collected emergency data is received by the corresponding data analysis unit, and if the first transmission data corresponding to the collected emergency data is not received by the corresponding data analysis unit, the first transmission data corresponding to the collected emergency data is transmitted to the data analysis unit whose operating load is lower than the threshold;

The monitoring module is used to monitor whether the first analysis data corresponding to the collected emergency data is received by the data processing unit. If the first analysis data corresponding to the collected emergency data is not received by the corresponding data processing unit, the first analysis data corresponding to the collected emergency data is transmitted to the data processing unit whose operating load is lower than the threshold.

Furthermore, when the monitoring module detects multiple data identified as collected emergency data within a preset time interval, the monitoring module is used to obtain the urgency level of the collected emergency data in the collection identification unit corresponding to the collected emergency data, and generate multiple data transmission path information corresponding to the collected emergency data in sequence according to the urgency level.

Further, the monitoring module is also used to monitor the progress of the collected emergency data being transmitted by the transmission identification unit, the progress of the first transmission data corresponding to the collected emergency data being analyzed by the data analysis module, and the progress of the first analysis data corresponding to the collected emergency data being processed by the data processing module, and generate first progress information, second progress information and third progress information respectively;

The monitoring module is used to receive a first processing speed threshold, and obtain the speed at which the collected emergency data, the first transmission data corresponding to the collected emergency data, or the first analysis data corresponding to the collected emergency data are transmitted, analyzed or processed based on the first progress information, the second progress information and the third progress information, and send a prompt signal when the transmission, analysis or processing speed is lower than the first processing speed threshold.

Further, the monitoring module is further used to receive a second processing speed threshold, and generate a switching signal when the speed at which the emergency data, the first transmission data corresponding to the emergency data, or the first analysis data corresponding to the emergency data are transmitted, analyzed, or processed is lower than the second processing speed threshold;

The data transmission unit, the data analysis unit and the data processing unit are used to monitor the switching signal, and when the corresponding switching signal is monitored, stop data transmission, analysis or processing, and extract the unprocessed collected emergency data, the first transmission data corresponding to the collected emergency data, or the first analysis data corresponding to the collected emergency data, and transmit the unprocessed collected emergency data, the first transmission data corresponding to the collected emergency data, or the first analysis data corresponding to the collected emergency data to other data transmission units, data analysis units or data processing units whose operating load is lower than the threshold.

Furthermore, it also includes a state acquisition module, which is connected to the data transmission unit, the data analysis unit and the data processing unit by signal, and is used to collect the operation load of each data transmission unit, the data analysis unit and the data processing unit;

The data transmission unit, the data analysis unit and the data processing unit are all used to feed back the data processing requirements to the status acquisition module when the operating load is lower than the threshold.

Furthermore, the monitoring module is used to temporarily store the collected emergency data when the operating load of each data transmission unit is not less than a threshold, and switch between the data transmission units to transmit the collected emergency data until the operating load of the data transmission unit is less than the threshold or the collected emergency data is received by the data transmission unit.

Further, the data collection unit is used to add a timestamp to the first collected data identified as collected urgent data;

The data transmission unit is used to calculate the delay between the time when the first collected data marked as collected urgent data is actually transmitted and the timestamp, and the monitoring module is used to determine whether the data transmission speed is lower than the first processing speed threshold and the second processing speed threshold according to the delay.

The above scheme has the following beneficial effects:

1. The present invention classifies and identifies data in the data collection stage, data transmission stage, data analysis stage and data processing stage in the process of acquiring the operating status of the wind power system. Compared with the prior art, for the emergency data (relevant data for prompts, faults, warnings, etc.) that need to be urgently analyzed and processed, it is ensured that they can be efficiently transmitted for analysis and processing during the real-time analysis and processing of big data, thereby ensuring that the emergency data in the real-time analysis and processing of big data can be quickly extracted and the abnormal operating status of the wind power system can be acquired in time.

2. The present invention adopts multi-level classification and identification. For some data that need urgent transmission but do not need urgent analysis and processing, urgent transmission is realized in the transmission stage; for some data that need urgent analysis but do not need urgent processing, urgent transmission is realized in the transmission stage, and urgent analysis is realized in the analysis stage; for data that need urgent processing, urgent transmission is realized in the transmission stage, urgent analysis is realized in the analysis stage, and urgent processing is realized in the processing stage. Compared with the prior art, it ensures that in the process of acquiring wind power system operation data, wind power system operation data with different types of requirements can be adaptively and efficiently allocated to meet data requirements.

3. The present invention, through active processing at each unit end combined with coordinated allocation at the monitoring end, fully utilizes the device's ability to analyze and process wind power system operation data in real time. Compared with the prior art, it improves the data processing balance between the units in the wind power system operation data analysis and processing device, balances the data processing load of some units, and ensures that various types of wind power system operation data after classification and identification can be fed back normally and effectively for obtaining the operating status of the wind power system.

Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of the structure of an embodiment of a device for real-time analysis and processing of big data based on deep learning according to the present invention;

FIG2 is a schematic diagram of the monitoring module structure of an embodiment of a big data real-time analysis and processing device based on deep learning of the present invention;

FIG3 is a schematic diagram of a state acquisition module of an embodiment of a big data real-time analysis and processing device based on deep learning of the present invention;

FIG4 is a flowchart of the construction of each recognition unit based on the deep learning model in an embodiment of the big data real-time analysis and processing device based on deep learning of the present invention;

FIG5 is a data processing flow chart of an embodiment of a big data real-time analysis and processing device based on deep learning of the present invention.

DETAILED DESCRIPTION

The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

The following is further described in detail through specific implementation methods:

Embodiment 1:

As shown in Figures 1 to 3: A big data real-time analysis and processing device based on deep learning is used for data collection during the operation of a wind power system. It is mainly composed of a data collection module, a data collection module, a data analysis module, a data processing module and a data visualization module. The data collection module, the data collection module, the data analysis module, the data processing module and the data visualization module interact and communicate in sequence.

Among them, the data acquisition module includes several data acquisition units and several acquisition and identification units, and the acquisition and identification units are set one by one with the data acquisition units; the data acquisition unit is used to collect the operation data of the wind power system in real time, and the real-time collected operation data of the wind power system mainly includes: environmental data (wind speed, wind direction, temperature, humidity, etc.), performance data (power generation, power curve, etc.), operation data (speed, current, voltage, frequency, turbulence intensity, etc.), and monitoring data (pressure, vibration, oil level, etc.).

The collection and identification unit is used to classify and identify the wind power system operation data collected by the data collection unit. The data collection unit is also used to identify the collected wind power system operation data according to the classification and identification results of the wind power system operation data to generate the first collection data. As the front end of the wind power system operation data collection, the data collection module is crucial for determining the type of wind power system operation data, and for obtaining the urgent needs of the wind power system operation data, so as to ensure that the collected wind power system operation data can be efficiently transmitted, analyzed or processed later.

As shown in FIG. 4 , the acquisition and recognition unit of this embodiment is constructed based on a deep learning model, and the construction steps are as follows:

Data collection and acquisition, determining the data source that needs to be analyzed, namely, wind power system operation data, mainly environmental data, performance data, operation data and monitoring data during the operation of the wind power system. The specific data content is the same as the above-mentioned real-time collected wind power system operation data;

Data preprocessing: processing invalid wind power system operation data, missing values, outliers, etc. to ensure the quality of wind power system operation data; reducing the dimension and standardizing the wind power system operation data, and selecting and extracting useful features according to the analysis task;

Model selection and construction: Select an appropriate deep learning model according to the task type. Convolutional neural network (CNN), recurrent neural network (RNN), long short-term memory network (LSTM), etc. can be used. This embodiment is mainly used to identify wind power system operation data. There is a large amount of wind power system operation data as a training set. Therefore, the long short-term memory network (LSTM) is used, which performs well in processing sequence data with long-term dependencies such as wind power system operation data;

Model training: divide the preprocessed wind power system operation data into training set, validation set and test set, and use the training data to optimize the model's weights and parameters so that the model can better fit the data.

In addition, the wind power system operation data newly collected by the collection module can be identified and analyzed in real time. Specifically, the collection and identification unit is used to classify the collected wind power system operation data according to the preset classification and identification rules, and to judge the urgency of the collected wind power system operation data based on the preset urgency characteristic value. When the urgency characteristic value is greater than the threshold, it is marked as collecting emergency data. Among them, the urgency characteristic value is the emergency information contained in the wind power system operation data, for example, the voltage contained in the wind power system operation data is too high or too low, the current is too high or too low, the frequency is too high or too low, etc. However, since the collected wind power system operation data may often include some abnormal data (for example: the frequency is in a state of fluctuation for a long time, but the fluctuation amount is maintained within a certain range under normal operation), in order to avoid the collected wind power system operation data from being mistakenly identified as emergency data, therefore, by setting a threshold, the wind power system operation data that really needs to be urgently analyzed and processed can be effectively screened out.

Accordingly, the data transmission module includes a plurality of data transmission units and a plurality of transmission identification units, and the transmission identification units are arranged in a one-to-one correspondence with the data transmission units; the data transmission unit is used to receive the first collected data, and the transmission identification unit is used to classify and identify the first collected data received by the data transmission unit, and the data transmission unit is also used to identify the first collected data according to the classification and identification result of the first collected data to generate the first transmission data. Among them, the transmission identification unit is also constructed based on the deep learning model (the construction method is similar to that of the collection and identification unit, which will not be repeated here), and the transmission identification unit performs a second identification on the wind power system operation data type, the purpose of which is to confirm whether the wind power system operation data requires emergency analysis or emergency processing.

Correspondingly, the data analysis module and the data processing module also adopt the same architecture as the data acquisition module and the data transmission module. The data analysis module includes a plurality of data analysis units and a plurality of analysis and identification units, and the analysis and identification units are arranged one by one with the data analysis units; the data analysis unit is used to receive the first transmission data, and the analysis and identification unit is used to classify and identify the first transmission data received by the data analysis unit, and the data analysis unit is used to identify the first transmission data according to the classification and identification result of the first transmission data, generate the first analysis data, and analyze the first analysis data. The analysis of the first analysis data is mainly to judge whether the wind power system operation data will cause abnormal operation of the wind power system, and the type of abnormality, for example: abnormal data such as the internal temperature of the system, the internal vibration frequency of the system, etc., which may cause abnormal types including abnormal fan blades, bearings, generators, etc.; abnormal data such as current, voltage, frequency, etc., which may cause abnormal types including power grid failure, power grid frequency or voltage abnormality, etc.

The data processing module includes a plurality of data processing units and a plurality of processing identification units, and the data processing units and the processing identification units are arranged in a one-to-one correspondence; the data processing unit is used to receive the first analysis data, and the processing identification unit is used to classify and identify the first analysis data received by the data processing unit, and the data processing unit is used to identify the first analysis data according to the classification and identification result of the first analysis data, generate the first processing data, and process the first processing data. The processing of the first processing data is mainly to formulate a troubleshooting strategy based on the operation data of the wind power system.

The analysis and identification unit and the processing and identification unit are also built based on the deep learning model. The analysis and identification unit performs a third identification on the wind power system operation data type, the purpose of which is to confirm whether the wind power system operation data needs to be urgently processed; the processing and identification unit performs a fourth identification on the data type, the purpose of which is to confirm whether the wind power system operation data needs to be urgently visualized. Through the above-mentioned multiple stages of wind power system operation data type identification, it is ensured that the collected wind power system operation data efficiently meets the data collection, transmission, analysis, processing and visualization requirements.

The data visualization module is used to receive and visualize the analysis results of the first analysis data and the processing results of the first processing data. The data acquisition module is usually a mobile terminal. The mobile terminal provides a data display platform on the one hand and a user human-computer control platform on the other hand.

For multi-stage data classification and identification, specifically, the collection and identification unit, the transmission and identification unit, the analysis and identification unit and the processing and identification unit respectively receive preset classification and identification rules, and respectively identify the urgency of the collected wind power system operation data, the first collection data, the first transmission data and the first analysis data according to the classification and identification rules. When the collected wind power system operation data, the first collection data, the first transmission data or the first analysis data is identified as emergency data, the corresponding identification unit generates data transmission path information, and the data collection module, the data transmission module, the data analysis module and the data processing module are used to transmit the emergency data according to the transmission path.

Among them, the transmission path information mainly includes the transmission path of the next stage or the transmission paths of subsequent multiple stages of the wind power system operation data. Specifically, it is based on the judgment of the type of wind power system operation data by the acquisition identification unit, the transmission identification unit, the analysis identification unit and the processing identification unit. For example, if the wind power system operation data needs to be urgently analyzed and processed, after the acquisition identification unit identifies it as emergency data, the transmission path of the wind power system operation data through the data transmission module and the data analysis module is generated.

After the wind power system operation data type is identified, the monitoring module is used to realize the monitoring, transmission, analysis, processing and visualization of the wind power system operation data. The monitoring module is connected with the acquisition identification unit, the transmission identification unit, the analysis identification unit, the processing identification unit and the data transmission unit. The monitoring module is used to obtain the acquisition emergency data identified by the acquisition identification unit, and monitor whether the data transmission unit receives the corresponding acquisition emergency data. If the acquisition emergency data is not received by the data transmission unit, the corresponding acquisition emergency data is transmitted to the data transmission unit whose operating load is lower than the threshold. Similarly, the data in the data analysis module and the data processing module that has not been analyzed by the corresponding data analysis unit or processed by the data processing unit, the first transmission data or the first analysis data corresponding to the acquisition emergency data is transmitted to the data analysis unit or the data processing unit whose operating load is lower than the threshold, to ensure that the wind power system operation data that needs to be urgently transmitted, analyzed, processed or visualized can be efficiently processed.

The operating load of each data transmission unit, data analysis unit, and data processing unit is obtained through the state acquisition module, which is connected to the data transmission unit, data analysis unit, and data processing unit signals. The state acquisition module is used to collect the operating load of each data transmission unit, data analysis unit, and data processing unit (active acquisition). The data transmission unit, data analysis unit, and data processing unit are all used to feedback the data processing requirements to the state acquisition module when the operating load is lower than the threshold (passive acquisition).

In conjunction with FIG5 , the real-time processing flow of the wind power system operation data of this embodiment is briefly described. First, the wind power system operation data is collected. After the collection, the collected wind power system operation data is classified to identify whether it is emergency data. If it is emergency data, the transmission path is planned. If it is non-emergency (conventional) data, it is transmitted according to the original transmission path; then, the above-mentioned similar steps are repeated to sequentially perform transmission classification identification on the first collected data, analysis classification identification on the first transmission data, and processing classification identification on the first analysis data. During the process, the transmission path is re-planned for emergency data, and non-emergency (conventional) data is transmitted according to the original transmission path. The final data analysis and data processing results are visualized to facilitate technical personnel to timely and accurately obtain whether there are abnormalities in the operation status of the wind power system, and to formulate troubleshooting strategies in a timely manner.

Embodiment 2:

The difference from Example 1 is that, for wind power system operation data identified as emergency data, it is also necessary to meet its transmission, analysis, processing or visualization requirements according to its urgency. Specifically, when the monitoring module detects multiple data identified as collected emergency data within a preset time interval, the monitoring module is used to obtain the urgency level of the collected emergency data in the collection identification unit corresponding to the collected emergency data, and sequentially generate multiple data transmission path information corresponding to the collected emergency data according to the urgency level.

For example, the collected wind power system operation data includes fault information of wind power system equipment, and the fault information includes data that causes fluctuations in the operation of the wind power system equipment and data that causes operational errors of the wind power system equipment. Based on the urgency of the wind power system operation data, generally, data that may cause operational errors of the wind power system equipment has a higher urgency level and therefore needs to be transmitted, analyzed, processed or visualized first.

In addition, the monitoring module is used to temporarily store the collected emergency data when the operating load of each data transmission unit is not less than the threshold, and switch between the data transmission units to collect emergency data transmission until the operating load of the data transmission unit is less than the threshold or the collected emergency data is received by the data transmission unit, ensuring that the wind power system data classified as emergency can be quickly transmitted, analyzed, processed or visualized.

Embodiment 3:

The difference from Example 2 is that the monitoring module is used to monitor the progress of the collected emergency data being transmitted by the transmission identification unit, the progress of the first transmission data corresponding to the collected emergency data being analyzed by the data analysis module, and the progress of the first analysis data corresponding to the collected emergency data being processed by the data processing module, and generate first progress information, second progress information and third progress information respectively.

The monitoring module is used to receive a first processing speed threshold, and obtain the speed at which the collected emergency data, the first transmission data corresponding to the collected emergency data, or the first analysis data corresponding to the collected emergency data are transmitted, analyzed or processed based on the first progress information, the second progress information and the third progress information, and send a prompt signal when the transmission, analysis or processing speed is lower than the first processing speed threshold.

The monitoring module is also used to receive a second processing speed threshold and generate a switching signal when the speed at which the emergency data, the first transmission data corresponding to the emergency data, or the first analysis data corresponding to the emergency data is transmitted, analyzed or processed is lower than the second processing speed threshold.

The data transmission unit, the data analysis unit and the data processing unit are used to monitor the switching signal, and when the corresponding switching signal is monitored, stop data transmission, analysis or processing, and extract the unprocessed collected emergency data, the first transmission data corresponding to the collected emergency data, or the first analysis data corresponding to the collected emergency data, and transmit the unprocessed collected emergency data, the first transmission data corresponding to the collected emergency data, or the first analysis data corresponding to the collected emergency data to other data transmission units, data analysis units or data processing units whose operating load is lower than the threshold.

Based on the above, it can be clearly seen that this embodiment sets the first processing speed threshold and the second processing speed threshold, and does not further interfere with the wind power system data whose processing speed is greater than the first processing speed threshold. For the wind power system data whose processing speed is between the first processing speed threshold and the second processing speed threshold, it is necessary to give priority to the user prompt to let the user know the existence of this situation. For the wind power system data whose processing speed is less than the second processing speed threshold, it is necessary to switch the unit to ensure that it can be effectively transmitted, analyzed, processed or visualized.

The judgment of the data processing speed of the wind power system is carried out in the following manner: the data acquisition unit adds a timestamp to the first collected data identified as collecting emergency data, the data transmission unit calculates the time delay between the time when the first collected data identified as collecting emergency data is actually transmitted and the timestamp, and the monitoring module is used to judge whether the data transmission speed is lower than the first processing speed threshold and the second processing speed threshold according to the time delay. For the data between the data analysis unit and the data transmission unit, the data transmission unit adds a timestamp to the first transmission data corresponding to the identified as collecting emergency data, and the data analysis unit calculates the time delay between the time when the first transmission data identified as collecting emergency data is actually analyzed and the timestamp. The same is true for the wind power system data between the data processing unit and the data analysis unit.

Obviously, the above embodiments are merely examples for the purpose of clear explanation, and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the invention.

Claims (9)

1. A big data real-time analysis and processing device based on deep learning, characterized by comprising: A data acquisition module, comprising a plurality of data acquisition units and a plurality of acquisition and identification units, wherein the data acquisition units are used to acquire wind power system operation data, the wind power system operation data including environmental data, performance data, operation data and monitoring data, and the acquisition and identification units are used to classify and identify the wind power system operation data acquired by the data acquisition units, and the data acquisition units are further used to identify the acquired wind power system operation data according to the classification and identification results of the wind power system operation data, and generate first acquisition data; The data transmission module includes a plurality of data transmission units and a plurality of transmission identification units, wherein the data transmission unit is used to receive the first collected data, the transmission identification unit is used to classify and identify the first collected data received by the data transmission unit, and the data transmission unit is further used to identify the first collected data according to the classification and identification result of the first collected data to generate the first transmission data; The data analysis module includes a plurality of data analysis units and a plurality of analysis and identification units, wherein the data analysis unit is used to receive the first transmission data, the analysis and identification unit is used to classify and identify the first transmission data received by the data analysis unit, the data analysis unit is used to identify the first transmission data according to the classification and identification result of the first transmission data, generate first analysis data, and analyze the first analysis data; The data processing module includes a plurality of data processing units and a plurality of processing identification units, wherein the data processing unit is used to receive the first analysis data, the processing identification unit is used to classify and identify the first analysis data received by the data processing unit, the data processing unit is used to identify the first analysis data according to the classification and identification result of the first analysis data, generate first processing data, and process the first processing data; a data visualization module, configured to receive and visualize an analysis result of the first analysis data and a processing result of the first processing data, and diagnose an operating state of the wind power system based on the analysis result of the first analysis data and the processing result of the first processing data; The collection identification unit, the transmission identification unit, the analysis identification unit and the processing identification unit are used to receive preset classification identification rules, and identify the urgency of the collected wind power system operation data, the first collection data, the first transmission data and the first analysis data according to the classification identification rules. When the collected wind power system operation data, the first collection data, the first transmission data or the first analysis data is identified as emergency data, the corresponding identification unit generates data transmission path information, and the data collection module, the data transmission module, the data analysis module and the data processing module are used to transmit the emergency data according to the transmission path information; The acquisition and identification unit is constructed based on a deep learning model, and is used to identify the collected wind power system operation data as collected emergency data, and then generate transmission path information, wherein the transmission path information includes the next stage transmission path or subsequent multiple stage transmission paths of the wind power system operation data identified as collected emergency data; The collection and identification unit is also used to judge the urgency of the collected wind power system operation data based on a preset urgency characteristic value. When the urgency characteristic value is greater than a threshold value, the collection and identification unit will identify the collected wind power system operation data as collection emergency data, wherein the urgency characteristic value is data in a long-term fluctuating state in the wind power system operation data, and the threshold value is the basis for judging whether the data in a long-term fluctuating state is emergency information. 2. According to the deep learning-based big data real-time analysis and processing device according to claim 1, it is characterized in that it also includes a monitoring module, which is signal-connected to the collection and identification unit, the transmission identification unit, the analysis and identification unit, the processing identification unit and the data transmission unit. The monitoring module is used to obtain the collection emergency data identified by the collection and identification unit, and monitor whether the data transmission unit receives the corresponding collection emergency data. If the collection emergency data is not received by the data transmission unit, the corresponding collection emergency data is transmitted to any data transmission unit whose operating load is lower than the threshold. 3. The big data real-time analysis and processing device based on deep learning according to claim 2, characterized in that the monitoring module is also signal-connected to the data analysis unit and the data processing unit; The monitoring module is used to monitor whether the first transmission data corresponding to the collected emergency data is received by the corresponding data analysis unit, and if the first transmission data corresponding to the collected emergency data is not received by the corresponding data analysis unit, the first transmission data corresponding to the collected emergency data is transmitted to any data analysis unit whose operating load is lower than a threshold; The monitoring module is used to monitor whether the first analysis data corresponding to the collected emergency data is received by the data processing unit. If the first analysis data corresponding to the collected emergency data is not received by the corresponding data processing unit, the first analysis data corresponding to the collected emergency data is transmitted to any data processing unit whose operating load is lower than the threshold. 4. According to the deep learning-based big data real-time analysis and processing device of claim 3, it is characterized in that the monitoring module is used to monitor multiple data identified as collected emergency data within a preset time interval, and the monitoring module is used to obtain the urgency level of the collected emergency data in the collection identification unit corresponding to the collected emergency data, and generate multiple data transmission path information corresponding to the collected emergency data in sequence according to the urgency level. 5. The big data real-time analysis and processing device based on deep learning according to claim 2 is characterized in that the monitoring module is also used to monitor the progress of the collected emergency data being transmitted by the transmission identification unit, the progress of the first transmission data corresponding to the collected emergency data being analyzed by the data analysis module, and the progress of the first analysis data corresponding to the collected emergency data being processed by the data processing module, and respectively generate the first progress information, the second progress information and the third progress information; The monitoring module is used to receive a first processing speed threshold, and obtain the speed at which the collected emergency data, the first transmission data corresponding to the collected emergency data, or the first analysis data corresponding to the collected emergency data are transmitted, analyzed or processed based on the first progress information, the second progress information and the third progress information, and send a prompt signal when the transmission, analysis or processing speed is lower than the first processing speed threshold. 6. The big data real-time analysis and processing device based on deep learning according to claim 5 is characterized in that the monitoring module is also used to receive a second processing speed threshold, and generate a switching signal when the speed at which the emergency data, the first transmission data corresponding to the emergency data, or the first analysis data corresponding to the emergency data are transmitted, analyzed, or processed is lower than the second processing speed threshold; The data transmission unit, the data analysis unit and the data processing unit are used to monitor the switching signal, and when the corresponding switching signal is monitored, stop data transmission, analysis or processing, and extract the unprocessed collected emergency data, the first transmission data corresponding to the collected emergency data, or the first analysis data corresponding to the collected emergency data, and transmit the unprocessed collected emergency data, the first transmission data corresponding to the collected emergency data, or the first analysis data corresponding to the collected emergency data to other data transmission units, data analysis units or data processing units whose operating load is lower than the threshold. 7. The big data real-time analysis and processing device based on deep learning according to claim 2 is characterized in that it also includes a state acquisition module, the state acquisition module is connected to the data transmission unit, the data analysis unit and the data processing unit by signal, and the state acquisition module is used to collect the operation load of each data transmission unit, the data analysis unit and the data processing unit; The data transmission unit, the data analysis unit and the data processing unit are all used to feed back the data processing requirements to the status acquisition module when the operating load is lower than the threshold. 8. The big data real-time analysis and processing device based on deep learning according to claim 7 is characterized in that the monitoring module is used to temporarily store the collected emergency data when the operating load of each data transmission unit is not less than a threshold value, and switch between the data transmission units to transmit the collected emergency data until the operating load of the data transmission unit is less than the threshold value or the collected emergency data is received by the data transmission unit. 9. The big data real-time analysis and processing device based on deep learning according to claim 6, characterized in that the data acquisition unit is used to add a timestamp to the first collected data identified as collected urgent data; The data transmission unit is used to calculate the delay between the time when the first collected data marked as collected urgent data is actually transmitted and the timestamp, and the monitoring module is used to determine whether the data transmission speed is lower than the first processing speed threshold and the second processing speed threshold according to the delay.