The system is composed of 4 layers, and the overview structure is presented in the figure 1. The principle of the design stems from the idea “awareness, transmission, perception and usage”, a concept widely applicated in the in the field of public safety .
- Data Acquisition Layer On the bottom layer of the system, large amounts of cameras and sensors are deployed to monitor situation of specific areas and capture the running parameters of equipment according to the identification and prioritization of risk factors (as presented in ). To be more specific, the PTZ (Pan/Tile/Zoom) cameras are installed in most ordinary area, while the cameras set in the boiler room or other areas of high incidence of fire are required with quality of heat resistance. In addition, with the function parameters embedded in BIM of hospitals, the covering scope of every camera can be present in 3D view, which contributes to the supplement and adjustment of cameras to ensure the concerned areas under monitoring. In terms of the equipment, the sensors, related to the pressure and temperature of liquid oxygen tank, the air quality in the tube of air condition, and the smoke in the air etc., are deployed. The selection of parameter requiring detection results from the causal connection between the deviation of these parameters and occurrence of emergencies. In the hospitals we surveyed, information of fire prevention, crowd gathering areas, and critical equipments including boilers, elevators, and HVAC (heating, ventilating and air-conditioning) etc., are already well detected. In comparison, sensors relevant to the states and security of gas tube networks, building structure, and neighboring spaces need to be monitored with corresponding sensing devices. Apart from automatic sensing, there are also some risk factors requiring manual inspection. For example, whether the firefighting access is unobstructed, or the preparedness of relevant staff members in terms of emergency procedures, and so on. Such information need to be collected by specific staff members and recorded in mobile terminals and transferred to system database for supervision and analysis.
- Transmission Layer The data comes from the front-end sensors and cameras need to be transferred to the part of data process and analysis and command center, and the notification to the staff and disposal instruction must be distributed to the mobile terminals. A multitude of communication means are employed for data exchange and communication. Existing local area network (LAN) carries internal system data including sensing data and mobile terminals (operating within the precinct of the hospital), while the wide area network (WAN) is used for exchanging data with external departments, such as meteorology office, fire safety agency, and so on, because some sever weather is considered as emergencies and uncontrolled fires require the fire departments getting involved. Under the extreme circumstance (e.g. emergency response with paralyzed communication infrastructure), amateur radios support the communication and command of internal staff members, and a series of special crisis communication means, including satellite phone and microwave communication, support the information exchanging and command transferring among hospitals and some external professional disposal sectors.
- Processing Layer To facilitate the automatic extraction and integration of large volume of heterogeneous data generated in the sensing and operation process, a database model is constructed as shown in figure 2. The model consists of 3 modules: •Real-time monitoring and controlling, distributes and controls the access permission, modes, and frequency, and exception handling, formulating a reliable data access mechanism; •Data extraction and format convention: the core of the model, examines the correctness of data received from the first module, automatically attracts heterogeneous data through multi-class and multithreading methods and different attracting rules, and establishes the uniform format via data format conversion in accordance with the requirements of database storage; •Real-time storage: maintains the processed data and serves as a basic database for decision-making support.
- The processing layer also hosts GIS and BIM support software, and assessment model base. The GIS system provides the information of the space scene of hospitals, and an abundant set of tools for calculation and simulation. The BIM system enables a complete digital modelling of hospital buildings, especially the interior information of them, such as space layout, room details as well as facilitates and infrastructure. The integration of the front-end sensing and monitoring data, and GIS and BIM systems, provides the foundation of risk assessment, which in turn enables the generation of incident forecast, impact prediction, alarm level and corresponding measures. D.Application Layer The layer presents the function and application the system will provide in the users interface, including: •Safety information management: users have the access to look up and alter the risk factors and related indexes; •Risks monitoring: dynamic grid management and the risk factor monitoring; •Assessment and alarming: dynamic states assessment and alarm dispatching; •Comprehensive situation awareness and joint response decision support.
The realization of the system function requires the advancement in some fields: 1) Risk factors classification and multi-source heterogeneous data collecting and accessing: section 2.1 and 2.3 provide the solution and description; 2) Precise and accurate positioning and visualization of full risk factor: the function is much dependent on the deep integration of GIS, BIM and IoT. The detecting data and location of sensors provide the inputs for calculating and simulating tools of 3-dimentional GIS, while BIM presents the spatial settings and completes some constraint conditions. Utilizing the algorithm and tools embedded in the assessment model, the function of full risk factor visualization also covers correlation analysis, key performance indexes (KIP) analysis and so on. 3) Comprehensive situation awareness and joint response: the process takes account not only the risks sources themselves, but also the situation of neighboring space, which contributes to more considerate disposal procedures and resources mobilization, and detailedly illustrated in the two cases in section 4
The two use cases are the potential or expected use of the risk management system, which are preconceived ideas or now simulated in other software, such as ArcGIS, whose related function will be integrated into the system.
This case is based on ArcGIS, part involved function of which will be incorporate into risk management system. In this simulated scene, the pressure sensors detect the abnormality in pipeline parameters, and then the assessment system estimates the potential influence scope and the severity in the emergency. As show in figure 3, the graph of networks and the data of flaw and pressure can be present with different colors. Moreover, the prioritized protection areas are located based on the data of buildings and sensors embedded beforehand in BIM and GIS. The predicted consequences of the leakage itself along with the states of the neighboring space, determine the level of alarm, disposal procedures and the response team and resource to mobilize. For example, if there are mass gathering or important buildings, such as laboratories, examination rooms and so on, in the neighboring space, the alarm level is certainly higher than the situation that the perimeter zone is open area or similar states. In addition, the amount of leakage is undoubtedly positive correlated to the alarm level. Meanwhile the disposal forces are called together, according to the command issued on the mobile terminals, and then the state of the scene and the corresponding disposal instruction will be distributed with them. The graph of pipe networks, related parameters of the pipe, like flow and pressure etc., the neighboring facilities requiring attention and analogous information are included. When the emergency is handled, the states of all regions come back to the default state, with regular color presented. If the situation gets worse, for example, the leakage cannot be controlled or other secondary incidents come up, another disposal force is going to evacuate the people nearby. What’s more, there is a coordinate response mechanism to call on the more professional disposal team out of hospitals on the basis of the pre-arranged plan stored in the database.
The case of fire in a laboratory is a preconceived idea based on the function and strategy of the system. The sensors in the laboratory detect the smoke, and the critical parameter data, like the density and components of smog, are transferred to the analysis sector, through IoT monitoring gateway. The camera view and the simulated visual scene are presented in the central control room immediately. Meanwhile the system will estimate the existing combustible materials and explosive substance in terms of their destructive power, and the potential of being affected, considering distance to the incident location and other factors. The situation near the laboratory, including the gathering crowd, the high value and significant assets and equipments, the function of neighboring rooms and so on, is also to be taken into account, which is much dependent on the preset data in BIM and GIS. With such situation awareness, the system is able to ascertain the level of threat posed. As shown in the figure 5 and 6, the location of the fire is near the operating room, which contributes to improve warning level. The disposal force will receive the information in the emergency scene and corresponding precept. Site disposal and evacuation will be conducted according to the pre-arranged plan, for instance the patients in the operating room will be protected separately and wait for rescue, while most people will be evacuated. When the fire is put out, the system will display the states of the hospital to be normal. Moreover, the cause of fire and the process of the disposal are taken as the reference to adjust the pre-plan and daily management
CONCLUSION AND DISCUSSION
There is some informatoinized response mechanism in terms of some specific emergencies, such as fires, parameters deviation of boilers and pressure pipelines, human attack and so on, but they are distributed in different management system. The risk management system proposed in the paper aims construct one comprehensive system cover all possible emergencies in hospitals, with integration and adjustment of the existing ones and supply other parts. The risk management system will promote the ability of hospitals in risk detection, analysis, visualization and response decision-making based on the integration of IoT, BIM and GIS. The system architecture and function design were described respectively, and its application was illustrated through application cases in this paper. In the viewpoint of public safety, the assessment and management of these risks consist of identification and classification, real-time monitoring and estimation, consequence prediction and alertness and coordinate response. Undoubtedly, the informatoinized systematic management of the risk elements contributes to achieve the goal. The difficulties of later work concentrate on the recognition algorithm with regard to the different emergencies, and the standard of different alarm levels. The economic efficiency also needs to be taken into consideration, since the disposal measures for some emergencies can meet the safety requirement. In terms of these emergencies, our work is to make their record of the state of risk factors informatoinized.