Research Of Key Technologies Of Semantic Sensor Web For Monitoring Building Structure Safety

With the development of the Internet of Things(Io T), all kinds of monitoring systems have been connected to the Internet constantly. The modern city is becoming more intelligent with a variety of interconnected data. The rapid development of city construction and the emerging danger in old buildings in the city provide an application field for the Io T. The monitoring system for the safety of the building structure is a specific application of the Io T. More and more buildings and bridges in the city have been deployed the monitoring system. The monitoring for the safety of the building will be lasted for many years, even the whole life of the building. Moreover, the mechanical features of a building cannot be reflected by only a few sensors or one type of sensors. These objective factors lead to a result that the monitoring system for the SBS has the characteristics of long service time, complex monitoring objects, and diversified data types. The big problems appear in sensor management, data management, operational efficiency, and system intelligence. The Semantic Sensor Web(SSW) associates the semantic technology with the sensors in the Io T together, provides effective data representation and management and builds the domain ontology as a domain specific knowledge expression. The main contents of this paper include:1) The monitoring system for the safety of the building structure belongs to an application system with multi-disciplinary and multi-field. After some research about technical specification for monitoring building, the composition of data acquisition instrument(DAI) and the sensor technical indicator, the ontologies for monitored building, DAI and sensor are designed to meet the needs of these different applications. With these ontologies, knowledge reasoning can be applied in system design, maintenance and fault diagnosis.2) The characteristics of data composition and organization are analyzed. On the premise of remaining the original data organization unchanged, the related resource is organized in different levels including the raw data, semantic information and multi-source knowledge. The raw data can be labeled to form the sematic information which is inherited from some ontology classes. Based on the fusion of multi-source knowledge, some rules are proposed for detecting the fault sensor and processing the sensing data.3)The more sensing data appears on the Web and this increases the burden on the engineer. After the analysis of the traditional data process, a meathod for processing the sensor data with combining the Web services is proposed. On one hand the knowledge on the Web can be used sufficiently for judging the rationality of data; on the other hand, combining Web service can realize the algorithm for processing the sensing data automatically. In order to avoid errors caused by the combination of asynchronous and parallel services, Pi calculus is used for the formalized expression and verifying the validity.4) A testing system is designed for monitoiring the safety of the building structure, and the functional verification is carried out by the real data from the field. The test results show that by combining the knowledge and the sevices, the errors in the sensing data can be detected out, which can be used for inferring the instrument failure. The combination of Web service also can be used to reduce the data size and provide the user with more accurate data.The main innovations of this paper are summarized as following:1) The monitoring for the safety of the building structure has the characteristic of inter-discipline. In order to overcome this problem, more suitable ontologies are developed to represent and organize the domain knowledge. There is no good solution for managing the error or missing information in the traditional monitoring system. In this paper based on the ontologies, automatic classification and logical detecting can be carried out and the intelligent level of the system is improved.2) The semantization method is proposed for the transformation of the different data to the semantic information, aiming at the data diversity in monitoiring building structure safety which includes the real-time data produced by the sensor, the devices properties data and the monitoring site information. The heterogeneity of multi-source data leads to a result that the sharing and the interoperability of data can not be realized. A differentiated semantization meathod makes the multi-source data fused in the domain knowledge and form different semantic resources.3) Aiming at the magnanimity, complexity and diversity, a meathod is proposed to process the sensor data by combing the Web services. For the manless monitoring field the sensing data is not only the reaction of the monitord objects, but also the sign of system operating state. Based on domain knowledge and Web service resources, this paper realizes the fault detection and the automated processing about the sensing data and combines different sorts of atomic Web services for detecting outlier, processing the sensor data and labeling data.