電氣工程專業(yè)畢業(yè)設計需要翻譯的英語文獻.pdf

Evaluation of Energy Effi ciency in Lighting Systems using Sensor Networks Declan T Delaney Gregory M P O Hare and Antonio G Ruzzelli CLARITY Centre for Sensor Web Technologies University College Dublin declan delaney gregory ohare ruzzelli ucd ie Bob Privacy Networked Embedded Systems Group Swedish Institute of Computer Science bob example se Abstract In modern energy aware buildings lighting control sys tems are put in place so to maximise the energy effi ciency of the lighting system without effecting the comfort of the oc cupant In many cases this involves utilising a set of presence sensors with actuators to determine when to turn on off or dim lighting when it is deemed necessary Such systems are installed using standard tuning values statically fi xed by the system installer This can cause ineffi ciencies and en ergy wastage as the control system is never optimised to its surrounding environment In this paper we investigate a Wireless Sensor Network WSN as a viable tool that can help in analysing and evaluating the energy effi ciency of an existing lighting control system in a low cost and portable solution We introduce LightWiSe LIGHTting evaluation through WIreless SEnsors a wireless tool which aims to evaluate lighting control systems in existing offi ce buildings LightWiSe determines points in the control system that ex hibit energy wastage and to highlight areas that can be opti mised to gain a greater effi ciency in the system It will also evaluate the effective energy saving to be obtained by replac ing the control system with a more judicious energy saving solution During a test performed in an offi ce space with a number of different lighting control systems we could high light a number of areas to reduce waste and save energy Our fi ndings show that each system tested can be optimised to achieve greater effi ciency LightWiSe can highlight savings in the region of 50 to 70 that are achievable through opti mising the current control system or installing an alternative Categories and Subject Descriptors H 4 wireless sensor networks embedded energy effi ciency buildings lighting Applications Miscella neous D 2 8 sensor networking Metrics complexity measures performance measures Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profi t or commercial advantage and that copies bear this notice and the full citation on the fi rst page To copy otherwise to republish to post on servers or to redistribute to lists requires prior specifi c permission and or a fee BuildSys 09 November 3 2009 Berkeley CA USA Copyright 2009 ACM 978 1 60558 824 7 5 00 Keywords sensor network energy effi ciency building decision support lighting 1Introduction Buildings are responsible for up to 40 of energy use as offi cial US statistics report 6 It is evaluated that 39 of all offi ce building electricity costs in the US are due to light ing 5 An estimated 30 of this can be classifi ed as wasted energy providing more effi cient lighting control systems that reduce the amount of wasted energy in offi ce buildings can provide a major contribution to lowering overall energy consumption Many new structures incorporate building management systems such as intelligent lighting or heating controls to improve occupants comfort and to save energy However a large number of existing buildings still have traditional light ing and heating installations This market is more responsive to quantifi able benefi ts and building owners individuals or companies are prone to invest in a modern lighting con trol system if the system promises them signifi cant energy saving in the monthly bill Currently energy consultants an installer can provide only a rough estimation of the energy saving through simulation and experience from previous in stallations The estimation error is large as consumption is tightly coupled to occupants activity building utilisation and its structural characteristics Control system installers are usually faced with the problem how can energy usage be re duced without impacting negatively on human experience One solution is to look at the way we use energy and de termine where energy is wasted We can then save energy by reducing this waste Previous tests performed in an of fi ce building over a seven month period 7 show a need to evaluate the performance of existing control systems and ad vise on an energy saving solution The fi ndings prove that a signifi cant amount of energy saving 46 can be gained by correctly installing a system into its environment or in stalling an alternative system Wireless Sensor Network WSN technologies have in troduced a cheap and portable method of measuring our sur rounding environment By using a WSN we can better rep resent the nature of a changing environment and use this to better control factors within the environment This is partic ularity relevant in buildings where certain factors light level temperature humidity are constantly controlled in a closed 61 environment In this paper we will investigate how a WSN can be used to evaluate a lighting control system in terms of energy effi ciency In stark contrast to existing approaches WSN is an infrastructure less and portable system that can be used to observe gather and interpret real sensor data in situ while the building is in everyday use taking into ac count the building occupancy work styles preferences etc The aim is to use LightWiSe to evaluate lighting control sys tems in existing offi ce buildings The rest of this paper is organised as follows Section 2 describes traditional lighting control systems and where energy waste occurs with these systems It also introduces energy saving concepts used in lighting systems to reduce energy waste Based on this sec tion 3 introduces a WSN approach to evaluating a control system and describes what LightWiSe aims to achieve Sec tion 4 describes the the design of the system and presents a set of design considerations when using the LightWiSe sys tem This section will also discuss how the data is modelled and used Section 5 describes the experiments what envi ronmental variables are considered hardware used how the data was gathered and presents the results set after data mod elling Section 6 concludes this paper and introduces further work 2Lighting Control Systems This section describes the most common lighting control techniques and concepts used in buildings In order to un derstand how a sensor network approach can be appropriate in evaluating light control systems it is important to under stand how existing light control systems operate In general energy loss for a lighting control system can be determined when the lighting system is illuminating a space that is not being used at that given time or is illuminating a space that has suffi cient ambient lighting for the needed tasks to be per formed The Chartered Institution of Building Services En gineers CIBSE code for lighting recommends lighting level to be maintained at 500lux for an offi ce environment 4 A brief description the most common lighting control systems is as follows Manual Switch This lighting control system is composed of a user switch to turn on off one or more lights The user decides to turn on or off a light or set of lights as to there preference Occupancy detecting The occupancy detecting lighting control systems with Passive Infra Red PIR sensors are popular within large buildings and offi ces When the PIR sensor detects movement in a space the con trol system activates the lighting system Occupancy sensors use timers and timing models when activating lighting Savings can be made by optimising the timers to suit the environment Ambient light compensated The ambient light compen sated control system uses a concept know as daylight harvesting throughtheuseoflightsensors Thecontrol system uses the data from the light sensors and the PIR sensors to make decisions with the lighting system We now describe two main energy saving concepts used for optimising the operations of lighting control systems 2 1Energy Saving Concepts Controlled spaces A controlled space is an area where a light or set of lights that can independently be turned on or off under a certain condition This condition may be evaluated over a single sensor or set of sensors By partitioning a large offi ce space into controlled spaces the control system can activate the lighting system in independent areas pinpointing certain areas where illu mination is needed Sensors can have a one to one or a many to one mapping with a controlled space Daylight harvesting the daylight harvesting system 3 monitors the amount of natural light present in a space The objective is to identify portions of the building where there is suffi cient natural light The control sys tem can then adjust the lighting to illuminate only the spaces receiving inadequate natural light 3Sensor Network Approach Evaluating existing lighting control systems is key to identifying and address sources of energy wastage In stark contrast to existing approaches based on simulations or ex pensive infrastructures the solution proposed uses a WSN approach The use of a WSN will give a high level of fl exi bility and portability Using a WSN it is possible to achieve quick low cost and non invasive deployment of the testing system The WSN approach is scalable and adaptable to fi t a wide range of spaces and offi ce topologies As nodes are added to the network in an ad hoc fashion and can be placed almost anywhere in a building they provide a fl exible frame work to produce a complete model of the system and en vironment The WSN is used to collect key environmental data from the structure under evaluation from which our rec ommendations are drawn The LightWiSe system is used to determine points in the control system that exhibit energy wastage and to highlight areas that can be optimised to gain a greater effi ciency in the system while considering the effects on the human experience We use a comprehensive approach that takes into account 1 human activity 2 Ambient light ing and 3 Artifi cial lighting when evaluating the control system By analysing this data we can determine when and where in the building energy wastage occurs Recorded data is sent wirelessly to a central gateway for storage and analy sis The LightWiSe system aims to evaluate the energy effi ciency of a lighting control system while highlighting ar eas that can optimised for greater energy effi ciency and to quantify the amount of energy saving that is possible by re placing the control system with an alternative solution To achieve this the LightWiSe system must measure exactly how much energy is wasted in the lighting system It should determine in what way the offi ce space is used from the data Optimising the control system to its environment and saving on wasted energy can be achieved by using this information From the data the LightWiSe system should also discover how and why energy ineffi ciencies occur in the control sys tem 4System Design Many factors should be taken into consideration when de signing an evaluation tool The tool must be able to adapt to 62 varying environments so that it can evaluate a variety of sys tems in varying situations An effective tool must be easy to deploy and provide a simple yet comprehensive results set to allow full and quick analysis A model of the evaluated system must be constructed from its fundamental properties and evaluated on its ability to perform its task 4 1Deployment Considerations There are a number of factors that must be taken into con sideration during deployment which can effect the results de rived from the WSN Network layout The sensor positioning and orientation is crucialforderivingusefuldatafromthesensornetwork Careful consideration needs to be given to to an appro priate position for each type of sensor The network layout should is to mirror the perception of the envi ronment by the people using it We use this as the central paradigm for our model to deduce where each sensor type is placed We can determine two types of used spaces workstations and walkways in an offi ce environment Light sensors are placed on used surfaces desk wall to determine a comprehensive model of the perceived light level in the offi ce Occupancy sensors are placed to monitor any movement within a controlled space Data gathering There are some trade off s to be consid ered for data collection when deploying the WSN Each node samples its sensors periodically to return environ mental data Sampling the sensors has an energy cost overheadonthenode Ontheotherhandtheaccuracyof the model of the environment is increased with a higher sampling rate We also need the node to remain active for the duration of the testing period This results in a trade off between model accuracy and node lifetime A similar trade off occurs between data throughput in the network and node lifetime as propagating messages through the network has an energy cost overhead on the sending and receiving nodes Consideration must be taken to avoid congestion caused by heavy through put in the network Although energy and congestion is sues are not the focus of this paper the system installer should take this into account when deploying the Light WiSe system The resulting data set from the network must be easy to analysis A large data set can be com putationally challenging to analyse while a smaller set can provide less accurate results The sampling rate and the throughput will affect the accuracy of the model of the environment so then a trade off exists between ob taining an acceptable model of the environment against node lifetime network congestion and ease of analysis Based on the described considerations the following sec tion describes how data are gathered from the sensor net work and then modelled to obtain meaningful information of energy wastage of the lighting control system in place Ulti matelysuchinformationisusedtoproviderecommendations for enhancing a light control system installation 4 2Data Modelling In order to capture environmental parameters of the build ing LightWiSe uses two common sensing devices 1 A light detector used for detecting ambient light and luminar ies state 2 A PIR sensor to detect people presence The resulting dataset allows the creation of a model of the envi ronment and shows how effi ciently the system runs within the environment The model will describe the light level throughout the offi ce the occupancy within offi ce and the state of the lights in the offi ce over a period of time Acquir ing of the result set from the data follows the model shown in Figure 1 The algorithm defi nes two independent coun ters namely Ambient light Waste ALW counter and Tim ing Model Waste TMW Counter The ALW counter will increment when both the ambient light exceeds the prede fi ned threshold and the luminary state is active This records the energy wasted due to ambient light The TWC counter is used to record wasted energy when the timing model is not calibrated for its environment Figure 1 Process fl ow for modelling data Data collected by the WSN defi nes three environment variables ambient light occupancy and luminary state Am bient light and the luminary state data are used to evaluate the ambient light waste in the system For each data point the Ambient light is checked against the threshold of 500lux When the ambient light level is under the threshold no waste due to ambient light occurs at this point When the ambi ent light level is over the threshold the luminary state for that datapointischecked ActiveLuminarystateatthispointwill correspond to energy waste Measuring the energy waste due to the lighting control system timing model also requires oc cupancy data The occupancy data is passed through a low pass fi lter to stabilize the data received from the occupancy detector The resultant is an envelope over the occupancy data that can be tuned for varying activity and user comfort levels The occupancy fi lter process is shown in Figure 2 The Occupancy fi lter algorithm describes a low pass fi lter with two counters namely No Motion Counter NMC and Occupancy Counter OC The NMC increments every data 63 point that shows no movement and is reset when movement is detected The OC increments when movement is detected The key threshold in the fi lter varies over time of day and the controlled space according to the amount of activity ex pected in the space Expected activity for a space is obtained through empirical measurement Each point in the data set then contains the luminary state and fi ltered occupancy state Waste due to timing model is then recorded when luminary state is active and fi ltered occupancy is zero The energy waste can then be expressed as a type of loss over a period of time The energy loss is calculated by multiplying lighting power in a controlled space by the calculated wasted time the lights were active in the controlled space The energy loss within the system can then be approximated for a monthly or yearly basis Figure 2 Process fl ow of data in fi lter 5Experimentations A test environment was established in a real world of fi ce scenario to validate fi ndings of the LightWiSe system We chose a site that is representative of many offi ce envi ronments The test site is shown in Figure 3 Testing took place in three separate spaces a large open plan offi ce a small individual offi ce and a corridor and present two types of control systems presence detecting control system and manual switch control system The layout of occupancy sensors are shown in Figure 3 The building is situated on latitude 53 degrees 20 minutes north with windows facing north Northwest this provides an even distribution of light throughout the offi ce In the open plan offi ce we used four occupancy detecting sensors with two light sensors to detect ambient ligh