5results about How to "Good predictive accuracy" patented technology

The utility model discloses a portable intelligent substation microstation equipment inspection device, including front end equipment, back end equipment and cloud equipment. Front end equipment is handheld inspection terminal, and integrated main control board, camera, display screen, fingerprint identification, communication interface etc. parts are used for information collection and interaction, and cloud equipment includes equipment information database, fault log record library, solution library, and is connected with front and back end equipment through information interaction module, realizes data storage and scheme calling, and back end equipment includes video processing, state monitoring, fault prediction, personnel scheduling and information transmission etc. module, forms the automation inspection and fault processing process. The device has the advantages of flexible deployment, accurate identification, efficient response, and can improve the intelligent degree of inspection and the substation operation and maintenance efficiency.

The present application relates to the technical field of scraper conveyor, and particularly relates to a chain breakage fault pre-judgment device and method of a scraper conveyor and the scraper conveyor, the device comprising: a controller, a first plug plate and a first magnetic sensor fixed to the first plug plate; a conveying groove of the scraper conveyor is provided with a first opening, the first plug plate is inserted into the first opening and located at a side of the chain close to a chain channel, and the first magnetic sensor is fixed to the first plug plate and opposite to one side of the chain; in the process of running of the chain of the scraper conveyor, the first magnetic sensor collects a first magnetic characteristic signal of a chain link passing through the first magnetic sensor; the controller is connected with the first magnetic sensor, the damage state of the chain link is analyzed according to the first magnetic characteristic signal, and whether the chain link is a chain breakage fault risk point is pre-judged based on the damage state. In this way, the present application pre-judges the chain damage of the scraper conveyor in advance, so as to realize the early warning and preventive maintenance of the chain breakage fault, and the operation reliability and safety of the scraper conveyor are improved.

This invention discloses a numerical method for transparent characterization of stress fields in mining engineering, belonging to the field of numerical simulation. The method includes: constructing a standardized numerical computation environment and determining the target processing area; establishing a three-dimensional rock mass mesh model and performing mechanical equilibrium solutions, simultaneously acquiring stress component and elastic strain energy data to achieve stress-energy dual-dimensional characterization; based on preset target center coordinates, real-time analysis and direct storage of spatial distance and angle information of each mesh point; writing node coordinates, displacement, stress, energy, and element connection relationships into a unified ASCII universal format file to generate a cross-platform readable data file; and finally, data verification and output. This invention overcomes the limitations of traditional methods, such as single characterization dimension, reliance on third-party processing for spatial analysis, and difficulties in cross-platform data interaction, significantly improving the efficiency and accuracy of stability analysis in mining engineering.

The present application relates to a kind of antibody viscosity prediction method and its application, the method includes: (1) extracting physical and chemical characteristics from the variable region sequence and structure of antibody;The physical and chemical characteristics include net charge, hydrophobicity, DeepSCM space charge chart score, space aggregation attribute based on DeepSP prediction, local SCM based on DeepSP prediction, surface hydrophobic region, positive charge region, negative charge region, variable region charge symmetry parameter and length parameter;(2) the features and viscosity value extracted in step (1) are used to train viscosity prediction model;(3) the viscosity prediction model obtained in step (2) is used for prediction;The viscosity judgment method is: greater than 20 cP, high viscosity, less than or equal to 20 cP, low high viscosity.The present application extracts 43 physical and chemical characteristics from the variable region sequence and structure of antibody, and the accuracy of detection is greatly improved compared with the existing prediction method by the viscosity prediction model obtained by training.

The invention discloses an anti-aging peptide prediction model and a construction method and application thereof. The construction method comprises the following steps: acquiring a training data set containing positive and negative samples; extracting a plurality of initial features from the peptide sequence; based on a weighted fusion evaluation and greedy search strategy of four models of a random forest, a gradient boosting decision tree, a lightweight gradient elevator and XGBoost, screening out an optimal feature combination composed of CKSAAGP, k-mer, AAindex, secondary structure and tertiary structure features; and by taking the optimal feature combination as input, constructing and training a neural network model which combines a bidirectional long-short-term memory network (BiLSTM) and a global predictor. According to the method, high-precision and high-efficiency prediction of the anti-aging peptide is realized through the optimized feature combination and the model structure.