Abstract: The invention discloses a method, a device and a medium for predicting a flue dust concentration, which calculate a flue dust emission amount of each batch of coal fed into a furnace based on hourly coal consumption of a unit as a label value of a prediction model, generate a general rule between data of the coal fed into the furnace and a corresponding flue dust emission amount through training the prediction model, accurately identify a relationship between material and the flue dust emission amount, reduce workloads of manual accounting and verification, and provide a reference for CEMS flue dust monitoring data. At the same time, using an Adam algorithm to optimize a BPNN allows for automatic adjustment of a learning rate for each parameter, enabling fast and efficient training of the prediction model.

Abstract: A method for extracting significant texture features of a B-ultrasonic image and application thereof discloses a channel attention mechanism network, i.e. a context activation residual network, which is designed to effectively model the B-ultrasonic liver fibrosis texture information, and which uses the global context information to strengthen important texture features and suppress useless texture features, such that the deep residual network can capture more significant texture information in the B-ultrasonic images. The process can be mainly divided into two phases: training and testing. During the training phase, the context activation residual network may be trained by using the B-ultrasonic image blocks as input and the pathological results of liver biopsy as labels. During the testing phase, the B-ultrasonic image blocks may be input into the trained non-invasive liver fibrosis diagnosis model to obtain the liver fibrosis staging result for each ultrasonic image.

Abstract: The invention discloses a method and its device for preparing a magnetic core with amorphous ribbon. The magnetic core is prepared with amorphous ribbon, the size of the amorphous ribbon is controlled according to the target requirements, and the magnetic core with required size and shape is prepared according to the target requirements; the single-roller rapid quenching technology with online automatic segmentation and automatic storage capability is used for preparation, which can control the length, width and thickness of the amorphous ribbon according to the target requirements; the amorphous ribbon segmented by single-roller rapid quenching technology is used to spray and cool down one by one, and then air-dry, transfer, spray adhesive and online store it one by one; the stored amorphous ribbon is reshaped, compressed and heat-treated successively, and then demoulded to prepare a magnetic core.

Abstract: The invention discloses a method and its device for preparing a magnetic core with amorphous ribbon. The magnetic core is prepared with amorphous ribbon, the size of the amorphous ribbon is controlled according to the target requirements, and the magnetic core with required size and shape is prepared according to the target requirements; the single-roller rapid quenching technology with online automatic segmentation and automatic storage capability is used for preparation, which can control the length, width and thickness of the amorphous ribbon according to the target requirements; the amorphous ribbon segmented by single-roller rapid quenching technology is used to spray and cool down one by one, and then air-dry, transfer, spray adhesive and online store it one by one; the stored amorphous ribbon is reshaped, compressed and heat-treated successively, and then demoulded to prepare a magnetic core.

Abstract: A method for extracting significant texture features of a B-ultrasonic image and application thereof discloses a channel attention mechanism network, i.e. a context activation residual network, which is designed to effectively model the B-ultrasonic liver fibrosis texture information, and which uses the global context information to strengthen important texture features and suppress useless texture features, such that the deep residual network can capture more significant texture information in the B-ultrasonic images. The process can be mainly divided into two phases: training and testing. During the training phase, the context activation residual network may he trained by using the B-ultrasonic image blocks as input and the pathological results of liver biopsy as labels. During the testing phase, the B-ultrasonic image blocks may be input into the trained non-invasive liver fibrosis diagnosis model to obtain the liver fibrosis staging result for each ultrasonic image.