Study On The Intelligent Ventilation Strategy Based On Big Data Technology

Food is the basic material for people to live on,and it is a strategic resource for every country to reserve,because it is related to social harmony and stability.Grain storage safety is an important part of food security,and grain warehouse ventilation is a necessary part of grain storage safety.Intelligent ventilation is a powerful measure to ensure the safety of grain storage.In order to avoid inefficient ventilation and even harmful ventilation,many intelligent ventilation strategy models are studied by experts.The primary goal of this paper is on the basis of predecessors' research by combining grain situation information associations before and after the fact to find a more practical and intelligent ventilation strategy model of higher accuracy,in this article,through the activation function can improve the S-P-LSTM intelligent ventilation strategy model and verify its validity.Other previous grain situation data acquisition frequency is too low,the training data set and test data set is too small,not enough trained model precision directly affects the accuracy of the ventilation strategy model.The reason is that the large amount of data cannot be stored in single machine,and the processing data of single machine can only be sampled.This paper presents a system architecture based on the large grain situation data,design a set of perfect data processing system for grain situation,solve the problem of mass storage,and focuses on the data cleaning process and improve overall system performance.The whole system process consists of four steps: data collection,data cleaning,data processing and data visualization.Finally,the accuracy and efficiency of s-p-lstm model are improved by using the food information system.The main research contents are as follows:(1)analyzed the deficiencies of the current intelligent ventilation system,and systematically studied the knowledge system of big data technology,such as Hadoop,HDFS,MapReduce,Hive,Spark and other key technologies.In addition,an architecture system based on big data is designed to improve the accuracy of intelligent ventilation decision.This paper implements the system and measures the model from operating time and acceleration ratio,and proves its feasibility and superiority.(2)the theoretical basis of the BP neural network is studied and the LSTM basic theory of the neural network,the BP algorithm and a brief induction of BPTT algorithm and made a detailed derivation LSTM neural network algorithm.(3)in view of the shortcomings of the intelligent ventilation decision model of the previous grain warehouse,this paper proposes an intelligent decision-making model s-p-lstm based on the improved activation function s-p-relu.And then use MATLAB to simulate the model and the simulation results are analyzed,and found that this model can well solve the problem of granary intelligent ventilation decision,finally by deploying to big grain situation data system platforms proves the superiority of the system.