Research On Intelligent Classification Technology And Equipment For “wet Without Dry” Community Living Garbage

Since the reform and opening-up,the economic and social development in China has been advancing rapidly,resulting in a significant increase in the amount of household waste.In many cities,there exist "garbage besieging the city" predicaments,making the community-based garbage classification and disposal an urgent key issue for urban management.According to the standard classification,household waste mainly consists of four categories: dry waste,wet waste,recyclable waste,and hazardous waste.Among these categories,wet waste accounts for a relatively high proportion,exceeding 50% in some regions.This research aims to address the demand for community-based waste classification and disposal,particularly the issue of "wet without dry".We have developed intelligent classification recognition technology and equipment that can identify,classify,and cut bags without manual supervision,thereby replacing traditional manual equipment.The specific research contents of this study are presented.According to the requirements of the community,users,and the classification facility system,an intelligent integrated equipment for wet waste classification in communities with no dry waste has been developed.This equipment adopts a ground-to-underground layout and has a "wet waste recognition module in the absence of dry waste" as its core.Through the optimization of mechanical structure and electrical control system design of the appearance framework and internal functional modules,functions such as identifying,sorting,bag-breaking,and compost resource utilization have been achieved.In addition,a static load simulation analysis of the mechanical device was carried out to calculate the maximum load that the frame can withstand and the critical point of plastic deformation under the influence of potential static loads,ensuring the stability and safety of the structure.Based on the STM32F103ZET6 core controller,the control system wiring schematic was drawn,and an experimental prototype platform was built to test and evaluate the accuracy of the weight and overflow module sensors,which met the application scenario requirements.The maximum average deviation of the weighing sensor was 0.08 kg,and the maximum variance was 0.03596kg;the maximum accuracy error of the overflow sensor was 1.56%.Through the Alibaba Cloud Io T platform transparent transmission experiment,the bidirectional data network communication between the equipment and the cloud platform was stable and reliable,and the transmission delay was less than 1s.Explored the intelligent waste classification recognition technology for "wet-only" waste using X-ray transmission image data,with a focus on identifying glass,metal,ceramics,and other typical-shaped waste contaminants mixed in wet waste.A dataset containing 800 X-ray images of "wet-only" waste was compiled and data augmentation was performed from the perspectives of shape,color,and pixels.The YOLOv5 s and YOLOv6 m models of deep learning convolutional neural networks were trained using the data,and evaluated using evaluation metrics such as confusion matrix,precision,recall,m AP,and F1-score.The best models obtained were YOLOv5 s and YOLOv6 m,with identification accuracies of 99.5% and 99.3%,respectively.The test sets were evaluated for accuracy distribution,omission rate,and false positive rate,and the overall identification accuracy for "wet-only" waste was greater than 95%.Specifically,the identification accuracies for YOLOv5 s and YOLOv6 m models were 96.25% and 97.5%,respectively,meeting the purity requirements for insitu composting of wet waste.Based on the requirement of the core function of automatic bag-breaking integrated equipment for wet garbage without dry garbage,a wet garbage automatic bag-breaking device was designed.The automatic bag-breaking process was divided into four movement steps: piercing,breaking,separation,and bagging.The mechanical structure and control system of the automatic bagbreaking device were optimized for pure bagging and bag-breaking of wet garbage after identification and sorting through transmission recognition modules.The electrical control system schematic was designed and drawn based on the STM32F103C8T6 controller,and a physical prototype of the "wet without dry" automatic bag-breaking device was built.The bag-breaking effect and stability of the automatic bag-breaking device on four types of mainstream household garbage bags were investigated under different experimental conditions of the same content but different thickness and different content but same weight.The optimal bag-breaking effect for all four types of garbage bags was achieved in the range of 2 threads in thickness and 2kg in weight,with a maximum removal rate of 98.5%.