Method, device and system for fault detection

A fault detection device for collecting interactive data generated during a working process of an Internet of things system, comprising:

• an acquisition module, configured for collecting the interactive data generated during the working process of a target Internet of Things system and generating data to be detected based on the interactive data;
• a detection module, configured for detecting whether the data to be detected is fault data or not, and performing an emotion analysis on the data to be detected which is judged as the fault data to obtain a corresponding emotion label;
• an output module, configured for extracting a corresponding fault type from a preset fault emotion mapping table based on the emotion label and outputting the fault type;
• and
• a model constructing module, wherein the model constructing module comprises
• a sample acquisition unit, configured for collecting the interactive data generated during the working process of the Internet of things system, and generating a plurality of pieces of sample data based on the interactive data;
• a first marking unit, configured for marking each of the pieces of sample data with a fault label to generate corresponding first training data, the fault label being normal or abnormal;
• a second marking unit, configured for extracting sample data with abnormal fault labels, marking the sample data with emotion labels and generating corresponding second training data;
• a first constructing unit, configured for training based on the first training data to obtain the detection model, the detection model being configured for identifying whether the input data to be detected is fault data or not;
• a second constructing unit, configured for training based on the second training data to obtain an emotion analysis model, the emotion analysis model being configured for identifying an emotion type of the input data to be detected;
• and
• a model output unit, configured for generating the fault detection model based on the detection model and the emotion analysis model.

A fault detection system, comprising:

• an Internet of Things system;
• an Internet of Things gateway coupled with an Internet of Things system signal and configured for collecting interactive data generated during a working process of the Internet of Things system;
• a cloud device coupled with the Internet of Things gateway signal, configured for receiving the interactive data collected by the Internet of Things gateway, and further for executing the fault detection method, wherein the fault detection method comprises:
• acquiring interactive data generated during the working process of a target Internet of Things system and generating data to be detected based on the interactive data;
• detecting whether the data to be detected is fault data or not, and performing an emotion analysis on the data to be detected which is judged as the fault data to obtain a corresponding emotion label;
• and
• extracting a corresponding fault type from a preset fault emotion mapping table based on the emotion label and outputting the fault type;
• wherein the detecting whether the data to be detected is the fault data is made based on a pre-built fault detection model, and the emotion analysis is performed on the data to be detected which is judged as the fault data to obtain the corresponding emotion label, and a method for constructing the fault detection model comprises following steps:
• collecting the interactive data generated during the working process of the Internet of things system, and generating a plurality of pieces of sample data based on the interactive data;
• marking each of the pieces of sample data with a fault label to generate corresponding first training data, the fault label being normal or abnormal;
• extracting sample data with abnormal fault labels, marking the sample data with emotion labels and generating corresponding second training data;
• training based on the first training data to obtain the detection model, the detection model being configured for identifying whether the input data to be detected is fault data or not;
• training based on the second training data to obtain an emotion analysis model, the emotion analysis model being configured for identifying an emotion type of the input data to be detected;
• and
• generating the fault detection model based on the detection model and the emotion analysis model.