基于TRIZ理论的“环境监测预报系统的设计应用”研究

Research on the Design and Application of Environmental Monitoring and Forecasting System Based on TRIZ Theory

通过建立深度学习模型,应用多层限制玻尔兹曼机(RBM)和一层反向传播(BP)网络组成多层深度信念网络,应用TRIZ理论分析,更好地提炼出空气浓度的趋势,从而更加准确无误地进行监测系统预报和监控。通过模拟人类大脑的神经连接结构,将数据原有空间的特征转换到模型特征空间中进行训练,自动地学习得到层次化的特征表示,从而提高预报性能。近年的实践表明,环境检测系统已成为解决环境空气监管的主要技术措施和手段,为城市主动监管和治理环境空气质量提供了强有力的工具,是实现国家治理雾霾的一项重要措施。

In this paper,a deep learning model is established,a multi-layer deep belief network is composed of RBM and BP,TRIZ theory is applied to analyze the trend of air concentration,so as to forecast and monitor the monitoring system more accurately.By simulating the neural connection structure of human brain,the feature representation of data in the original space is transformed into a new feature space with semantic features,and the hierarchical feature representation is automatically learned to improve the prediction performance.Practice in these years shows that environmental detection system has become the main technical measures and means to solve the problem of environmental air supervision.It provides a powerful tool for the city to actively monitor and control the environmental air quality and an important measure for the country to control the haze.