2results about How to "Low computing resource requirements" patented technology

The invention belongs to the technical field of power electronic system control, and particularly relates to a bridge arm energy balance control method of a modular multi-level converter, which comprises the following steps: generating a direct current bus current reference value for representing the direct current side energy regulation requirement of the modular multi-level converter; in combination with the energy state of each bridge leg, generating a bridge leg current reference value corresponding to each bridge leg by using a bridge leg current reference value distribution strategy; generating mutually independent bridge leg common-mode voltage modulation signals of the bridge legs; based on a common-mode modulation signal energy rectification distribution strategy, performing energy rectification type distribution in the bridge legs, and generating common-mode voltage modulation signals corresponding to the upper bridge arm and the lower bridge arm; the differential mode voltage modulation signals are combined with differential mode voltage modulation signals generated by alternating current side control to generate bridge arm voltage modulation signals corresponding to the bridge arms; trigger pulses of all sub-modules in the bridge arm of the modular multilevel converter are generated through a modulation method, and controlled adjustment of the energy of the bridge arm is achieved.

A chlorophyll visualization method and system based on hyperspectral ensemble learning belongs to the field of instrumentation technology. The method involves measuring chlorophyll physicochemical indices from modeled one-dimensional spectral data and augmenting the data with Gaussian noise and distortion. The data is then divided into a modeling set and a prediction set in a 7:3 ratio, ensuring the modeling set includes the prediction set. The divided modeling set is used to train and optimize XGBOOST, PLS, and RIDGE models respectively, yielding three optimal models. These models are then weighted and merged to obtain an ensemble model. After iterative optimization, the optimal weighted ensemble model is obtained. The spectral data of the region of interest (ROI) of the hyperspectral data to be measured is input into the optimal weighted ensemble model for prediction, and the predicted results and corresponding spatial locations are saved. Chlorophyll information is overlaid on the corresponding chlorophyll pixels in the three-dimensional image of the hyperspectral data to be measured, and a heatmap is generated. This method exhibits strong generalization ability and achieves high accuracy even with small sample sizes.