Zach-2010-Monitoring for Simulation Validation-182.Pdf
Total Page：16
File Type：pdf, Size：1020Kb
MONITORING FOR SIMULATION VALIDATION Robert Zach and Ardeshir Mahdavi Department of Building Physics and Building Ecology, Vienna University of Technology, Austria building data streams are not exploited. Such benefits ABSTRACT include: One of the key problems in building simulation is to i) Energy optimization through improved determine the accuracy of a simulation model. Due to management of technical building systems. the complexity of a building, a comprehensive and exhaustive mathematical proof is usually not ii) Increased awareness of building users possible. Therefore, an appropriate way to validate a regarding their impact on buildings’ energy building model is to compare simulation results with use. measurements obtained from real buildings. Such iii) Early detection (and treatment) of comparisons not only allow for the validation of deficiencies and malfunctions in energy simulation models used in the context of building systems and devices, thus effectively design support, but also provide calibrated simulation supporting a preventive maintenance models to be applied in the context of real-time regime. simulation-assisted building systems control. iv) Successive building performance INTRODUCTION improvement and optimization via the analyses of dynamically updated building This paper deals with the monitoring infrastructure energy and performance data bases. necessary to validate building simulation models and implement simulation-based control strategies v) Long-term accumulation of empirical (Mahdavi et al. 2009, Orehounig et al. 2010). information on buildings' energy and Required sensors are discussed and technologies for environmental performance toward different domains are compared and assessed. improving the design, construction, and Possible network infrastructures to collect the operation of existing and new buildings. measured data are discussed. To fully validate Further, to validate simulation models and to building simulation models, a multi-layered implement simulation-based control strategies, real- architecture for concurrent energy, performance, and time sensor-data is required. Therefore a monitoring occupancy monitoring is defined. network is essential. Two buildings in Vienna, Austria are equipped with multi-layered monitoring systems. One is a new APPROACH building, which already provides some monitoring Research efforts involve the following steps. First, a infrastructure. This paper illustrates how this network based monitoring in two buildings of the building can be enhanced with a comprehensive Vienna University of Technology will be realized to monitoring system, while reusing the existing obtain real-world data. Then simulation models are infrastructure as much as possible. The second validated and improved based on the monitored data. building was built more than 100 years ago and Finally, the models are used to execute simulation- provides no reusable building automation based control strategies in real-world scenarios. infrastructure. Therefore, an independent system The present paper focuses on the required design and must be installed for monitoring. implementation of the monitoring infrastructure. Motivation Summary There is currently a paucity of systematic and First, an overview of relevant technologies for comprehensive implementations of monitoring monitoring systems is presented. To structure infrastructures in buildings (Raftery et al. 2010, J. technologies a four layer model is used. Then, two O'Donnell 2009, Neumann and Jacob 2008). Thus, prototypical implementations of the proposed multi- the critical benefits that could result from the layered monitoring system are described in detail. integrated and concurrent analysis of multiple The paper concludes with discussion and future sensor for heating, ventilation, and air-conditioning outlook. (HVAC) can work with longer intervals (i.e., minutes). MONITORING NETWORK TECHNOLOGIES Fieldbus level Usually building communication networks are The function of the fieldbus level is to transfer the described with the three layer model defined in ISO measured data streams to the automation layer, which 2004. This model is appropriate to describe network then acts as a backbone. Common service parameters communication strategies, but lacks the coverage of for networks are throughput, reliability, security, sensor/actor technologies. It does not deal with the scope, real-time, and power use. These challenge of getting the information of different characteristics can be used to describe a fieldbus, but physical domains into an electronical signal and their it is not always possible to directly compare them different requirements regarding fieldbus networks. based on these properties. For example, KNX has a To fully cover monitoring strategies, an additional much smaller throughput than LonTalk on the wire. layer describing sensor/actor technologies is added as But, depending on the system design and the shown in Figure 1. grouping of devices in different KNX lines, the overall network load can be less than in a comparable LonTalk system. To get a starting point for technology decisions, current wired and wireless fieldbuses and their field of applications are listed in Table 2 and Table 3 (Daniels 2003, KNX 2004, LON 2010). Different monitoring strategies can be used to get the measured data from the sensor to the management level. The sensor can send measurements event triggered when some predefined conditions occur, or periodically with a fixed interval, or the Direct Figure 1 Four layer model of a generic monitoring Digital Controller (DDC) can poll the stations. system Depending on the data stream, different strategies fit Physical level best. For example, data from a presence detector is most accurate when an event based strategy is used The physical level addresses pertinent sensory while an electrical meter is usually polled with a devices and technologies that are required toward an periodic interval to obtain a temporal view. An event efficient, dynamic, and scalable acquisition of the based strategy can also be used to reduce power required data. Table 1 provides an overview of consumption of battery or self-powered devices. relevant data streams together with their required sensor technologies. Table 1 Table 2 Data streams and required sensor technologies Wired fieldbuses and their field of applications DATA STREAM SENSOR TECHNOLOGIES FIELDBUS FIELD OF APPLICATIONS i Energy use (Sub-)Meters for electricity, KNX/TP General purpose fieldbus. Used for lights, gas, oil, water, etc. blinds and HVAC systems. KNX is the ii Indoor Temperature, humidity, CO2, successor of the European Installation environment VOC, illuminance Bus (EIB) and is therefore mostly used in iii Outdoor Temperature, wind, rain, solar the European Union. environment radiation LonTalk General purpose fieldbus. Used for lights, iv Occupants’ Motion/presence detectors, blinds and HVAC systems. presence, actions, number of people, location M-Bus Used for metering devices (electrical and feedback sensing meter, heat meter, flow meter, etc.). v Environmental Window and door states, blinds DALI Used for controlling lights in isolated control systems applications. states vi Economical Documents, reports, bills information Different sensor technologies do often require different fieldbus characteristics. For example, a presence-detector to turn on the light needs to send its information in a fraction of seconds while a CO2 Table 3 Management level Wireless fieldbuses and their field of applications The management layer handles the data storage, the FIELDBUS FIELD OF APPLICATIONS visualization and the further processing of the data ZigBee Supports the dynamic creation of meshed streams. Possible technologies for historical storage networks which increases reliability and and abstract data representation are OPC (Data scope. ZigBee is based on the IEEE 802.15.4 standard and can work in the 2,4 Access - DA, Historical Data Access - HDA, Unified GHz and 868 MHz ISM band. Most Architecture - UA), BACnet/Web-Services (WS), devices use the 2,4 GHz band which is oBIX and custom database designs. often crowded when, for example, wireless OPC DA is highly used to provide a common local area networks (WLAN) are used too. interface to different automation and fieldbus ZigBee is used as a general purpose networks in the management layer. So called OPC fieldbus. DA servers abstract the sensors and actors as data EnOcean Is optimized for low power consumption, which therefore allows the construction of points. The data of the OPC servers can then be self-powered sensor/actor devices. accessed with OPC clients, which can be a user EnOcean uses the 868 MHz ISM band with interface or any other processing application. OPC amplitude modulation optimized for short DA server provide only live data and run only on packet transmission time and low power windows operating systems. To provide historical consumption. This increases throughput data access the OPC HDA standard or a custom and reliability. EnOcean is mainly used for database is usually used (Iwanitz and Lange 2002, self-powered sensors and simple actuators. OPC 2010). KNX/RF Is the wireless version of KNX. Only a few devices are available on the market at the To overcome the restriction of running OPC DA and time. OPC HDA server on windows only and to integrate Z-Wave Is designed for small systems in the field of all OPC sub-standards (DA, HDA, etc.), the OPC home automation. UA standard was created. It provides high potential,