1、4760 Coupling of Airflow and Pollutant Dispersion Models with Evacuation Planning Algorithms for Building System Controls J.S. Zhang C.K. Mohan Member ASHRAE K. Mehrotra S. Wang ABSTRACT A mathematical formulation was developed including an objective function that minimizes the cumulative exposure o
2、f occupants to pollutants under the constraints of pollutant dispersion pattern, evacuation paths, and their capacities. The formulation included airflow andpollutant dispersion models and optimization algorithms for evacuation planning. The formulation was applied to two example cases: (I) a 6-zone
3、 testbed assuming a constant pollutant release at the HVAC intake and (2) a 73-zone case representing ajoor section of 22,000 fi within a four-story building assuming an instant pollutant release inside onezone of the building. Assuming that a simulatedpollutant (SFJ was released at the HVAC intake
4、or inside a particular zone of the building, simulations were conducted for (1) a normal mode of HVAC operation, (2) an intuitive pollutant dispersion control mode, and (3) an intel- ligent control mode in which the feedback from the evacuation planning model was used. Pollutant dispersion control s
5、trat- egies included closing selected dampers, turning of general air supply fan, pressurizing the corridor zone and pathway zone adjacent to a designated “safe haven ”zone, turning on a backup exhaustfan, depressurizing contaminated zones, and activating decontamination devices. Simulation results
6、show that (I) an evacuation plan obtained based on the predicted pollutant dispersion patterns can have signijcant advantage over the intuitive “shortest path to the exit” approach and (2) the pollutant dispersion control strategies simulated are efec- tive in reducing the occupants exposure. The in
7、telligent control methodology proposed can minimize the exposure of humans to pollutants indoors, subject to computational and cost constraints inherent in the real-time nature of theproblem. P. Varshney C. Isik Z. Gao R. Rajagopalan INTRODUCTION Indoor environmental quality (IEQ) can have significa
8、nt impact on human health, comfort, satisfaction, and productiv- ity. Fisk (2001) estimated that improving IEQ could poten- tially result in annual productivity gains of up to $250 billion per year in the US alone. While the importance of improving IEQ cannot be overemphasized, it is equally importa
9、nt that the IEQ goals be achieved in an energy-efficient and cost-effective manner. This would require that the building environmental systems be optimally operated and controlled based on heat- inglcooling and purification requirements, variable utility rates, and emergency response requirements (e
10、.g., rapid evac- uation in case of fire). The need for developing buildings immune to potential chemical and biological agent (CBA) attacks further escalated the urgency of developing an intelligent building environmen- tal system (i-BES) that can achieve desired IEQ goals while minimizing energy co
11、nsumption and cost and the risk to occu- pants in case of an emergency. Such an i-BES would require reliable sensors distributed in and around the building; communication networks that transmit the sensed data to local or central information processors, which, in turn, devise opti- mal control strat
12、egies or emergency response plans; and control devices, such as fans, airflow dampers, and air purifi- ers, that implement the control strategies or safety officers who execute the emergency response plans. In order to devise optimal control strategies or emergency response plans, the i-BES needs to
13、 be able to predict in real time the outcome of environmental conditions and its impact (on occupant exposure to pollutant) if certain controls are actuated. The objective of this paper is to introduce an i-BES model that couples an airflow and pollutant dispersion model J.S. Zhang is an associate p
14、rofessor and S. Wang and Z. Gao are graduate research assistants, Department of Mechanical, Aerospace and Manufacturing Engineering, and C. K. Mohan, P. Varshney, and K. Mehrotra are professors, C. Isik is an associate professor, and R. Raja- gopalan is a graduate research assistant, Department of E
15、lectrical Engineering and Computer Science, Syracuse University. 196 02005 ASHRAE. with an evacuation planning algorithm. Through computer simulations for two example cases, we illustrate how the proposed model would work to obtain optimized control strat- egies for reducing the exposure of occupant
16、s to the pollutant during evacuation, assuming that a chemical is released at an HVAC intake or inside a building. A SYSTEM MODEL FOR i-BES Optimization Framework for Intelligent Control The goals of the intelligent control system can be formu- lated as multiobjective optimization (Andersson 2000; C
17、oello et al. 2002) tasks in which the system must focus on multiple distinct objectives: 1. Safety of occupants 2. 3. Personal comfort of occupants 4. Energy consumption Each of these objectives may conflict with others-for instance, attempting to maximize conflicting temperature- level comfort requ
18、irements of different occupants can be expected to result in increased energy consumption if not well optimized at the system level. There are also some intangible costs that must be considered in the final system evaluation- for instance, the implementation of some control and evacua- tion activiti
19、es may result in inconvenience to occupants, which should be avoided if possible without compromising other considerations. Several methods have been proposed in the literature to address multiobjective optimization tasks. These methods first require constructing a quantifiable model for each object
20、ive and then selecting an approach to handle multiple objectives. The simplest multiobjective optimization technique involves constructing a linear combination of the quantified expres- sions describing each objective, where the weights for each objective are obtained from problem-specific expertise
21、 or util- ity-theoretic considerations. Nonlinear combining functions are instead preferred in some applications, e.g., when personal comfort considerations such as temperature preference are to be balanced against energy cost. Another multiobjective opti- mization strategy is a Pareto approach, whe
22、re the goal of the optimization algorithm is to obtain a large collection of ?nondominated? solutions, Le., candidate solutions that are not worse than any other solution with respect to all objectives. One Pareto-optimal solution may be better than a second Pareto-optimal solution with respect to o
23、ne objective but worse with respect to a different objective. The second and fourth objectives can be quantified using standard economic analyses. The third objective, personal comfort of occupants, requires utility-theoretic analyses, transforming subjective preferences into numerically quanti- fia
24、ble evaluations. We focus on formulating and separately optimizing the first objective, viz., occupant safety, in the rest System development, implementation, operational and maintenance costs of this paper, particularly in the context of potential chemical and biological contamination of the indoor
25、 building environ- ment. In the case of a pollutant release, even when some zones in a building cannot be decontaminated rapidly, human occupants can be moved to other zones within or outside the building, thereby helping to achieve the primary goal of minimizing damage to human health. This is the
26、activity we refer to as ?evacuation,? although a more precise term would be ?people movement control.? It should be noted that building evacuation planing has been an active research area for many years, especially in the field of fire safety and smoke control. Gwynne et al. (1999) reviewed the meth
27、odologies used in 22 evacuation models. These models cover different applications, including optimization, simulation, and risk assessment, and differ in the methods of building enclosure representation, population perspectives, and behavioral perspective. They have been used to plan evacuations or
28、analyze people?s behavior in case of fire emergencies (e.g., Thompson and Marchant 1995; Ashe and Shields 1999; Benthorn and Frantzich 1999). These studies are focused on the effects of people attributes (e.g., location, response to alarms and signs, moving speed), exit routes, and smoke movement. T
29、he proposed i-BES model differs from these previous models in that it proposes an approach for the dynamic optimization of the evacuation plan based on the monitored and predicted pollutant dispersion patterns as well as the people attributes and exit routes. An i-BES Model for Optimal Evacuation Pl
30、anning A BES can be viewed as a nested multi-scale system involving a microenvironment around a person, an individual room environment, and the whole building environment. Vari- ous control strategies can be applied at each scale for achiev- ing optimal IEQ, high energy efficiency, and adequate buil
31、ding security. As controls are applied from the whole building down to the personal scale, the BES can satisfy the environmental and safety needs of more diverse individual occupants who may have very different preferences for envi- ronmental conditions and, hence, increase the occupants? satisfacti
32、on to higher degrees than the currently adopted industrial standards for ventilation, IAQ, and thermal comfort (ASHRAE 2001,1981). As shown in Figure 1, the proposed i-BES includes distributed sensors that monitor environmental conditions in the built environment and its surrounding outdoor environm
33、ent. The data will provide the real-time initial and boundary conditions needed for the prediction of the airflow and pollutant dispersion in the building. The predicted pollutant dispersion patterns (i.e., concentrations over time in various zones/rooms in the building) are fed into an evacuation p
34、lanning algorithm that suggests an optimal evacuation plan that is conditional to the predicted pollutant dispersion patterns. The evacuation planning algorithm considers existing occupant locations, available evacuation paths and their capacities, in addition to the pollutant dispersion pattern, AS
35、HRAE Transactions: Research 197 Figure 1 An i-BES model for optimal evacuation planning. Figure 2 Model-based predictiveflorecasting control methodology. and may suggest control operations that would change the pollutant dispersion pattern if implemented. These control operations will then be simula
36、ted by the pollutant dispersion model. This iterative process would continue until a global optimum evacuation plan is devised, which would be implemented to modify the BES conditions. This monitoring, prediction, optimizing, and controlling process would be performed by the i-BES on a continuous an
37、d dynamic basis. The predictive-control algorithm (Figure 1) is essential for the BES control because there can be significant delays between the action of the system controller and actual change in the environmental conditions across multiple scales. A traditional error-driven controller is not ade
38、quate, and a model-based predictive/forecasting control methodology should be used (Figure 2). The predictive control algorithms rely on rapid predictions of the airflow and pollutant disper- sion patterns in buildings. The pollutant dispersion patterns are affected by the airflow around the buildin
39、g, through the building envelope, interzonal airflows, air distribution inside a single roomizone, and the microenvironment around each human occupant, each of which has different spatial and temporal scales. Currently, no sufficiently fast model exists that can predict all the different scales of t
40、ransport processes in detail. However, several studies have been conducted to combine a computational fluid dynamics (CFD) model with a multizone airflow model to develop methods for predicting the airflow and pollutant transports in a whole building while being able to resolve in sufficient detail
41、the pollutant distribu- tion in individual room scales (e.g., Schaelin et al. 1993; Musser 2001; Zhai et al. 2003). In the following, we focus our discussions on the predic- tions of pollutant dispersion patterns in a multizone building and how the prediction can be used to optimize the evacuation p
42、lan in case of pollutant releases, assuming that each zone has uniform pollutant concentrations. The purpose is to investi- gate and illustrate the feasibility and potential of the proposed i-BES control methodology and identify the key areas for progress. w Building Level 4 4 Zone RoomlZone Zone 1
43、N Levei Figure 3 Schematic of airflow andpollutant transport in a multizone building. Mathematical Formulation Consider a BES with Nzones, including the ambient zone (Figure 3). Each zone has a defined air volume (V,) in which perfect air mixing is assumed. Controlled variables of the BES are pressu
44、re (Pi), temperature (Ti), and pollutant concentration (Ci) for each zone (i). The source emission rate in zone i is designated as si, which can represent typical indoor emission sources or accidental or intentional releases of chemical or biological pollutants. Pollutants could also be released in
45、the outdoor air supply intake. Each zone is serviced by a general (building level) heating, ventilating, and air-conditioning (HVAC) system with supply and return airflow rates of Qsi and e, respectively. The building HVAC system has an outdoor airflow rate of eo, exhaust flow rate ofQE, and return
46、flow rate of QR with a filter efficiency of vi. Each zone may be equipped with a local air supply and exhaust unit for local ventilation and pressure control, which may operate at given supply and exhaust flow rates of qsi and qei, respectively. They may also have a local air purification unit opera
47、ting at a flow rate of qi and contaminant removal efficiency of vi. Airflow rate controls at the building and roodzone level and air filters or purifiers are possible means of controlling the pollutant dispersion and concentrations in the building, and each has 198 ASHRAE Transactions: Research asso
48、ciated initial and operating costs. Sensors can be installed in each zone to provide timely detection of potential pollut- ants, also with an associated cost. An i-BES would be designed to minimize the exposure of occupants within the cost constraints to satisfi the requirements for emergency respon
49、ses, or it would be designed to minimize costs while maintaining acceptable exposure of occupants to pollutants under normal (routine) operations. Assuming that there are mi occupants in zone i at time zero when a pollutant is released at the outdoor air intake or one or more zones, what would be the optimal control strategy and evacuation plan if the goal is to minimize the exposure of the occupants to the pollutant? The answer to this question depends on the pollutant concentration and number of occupants in each zone and possible evacuation paths and their capacities. The mathemat- ical
