ASHRAE FUNDAMENTALS SI CH 19-2017 Energy Estimating and Modeling Methods.pdf

《ASHRAE FUNDAMENTALS SI CH 19-2017 Energy Estimating and Modeling Methods.pdf》由会员分享，可在线阅读，更多相关《ASHRAE FUNDAMENTALS SI CH 19-2017 Energy Estimating and Modeling Methods.pdf（54页珍藏版）》请在麦多课文档分享上搜索。

1、19.1CHAPTER 19ENERGY ESTIMATING AND MODELING METHODSGeneral Considerations. 19.1Degree-Day and Bin Methods . 19.6Thermal Loads Modeling. 19.8HVAC Component Modeling . 19.15Low-Energy System Modeling 19.24Data-Driven Modeling . 19.27Model Calibration 19.34Validation and Testing . 19.37NERGY requireme

2、nts of HVAC systems directly affect a build-E ings operating cost and indirectly affect the environment. Thischapter discusses methods for estimating energy use for two pur-poses: modeling for building and HVAC system design and associ-ated design optimization (forward modeling), and modelingenergy

3、use of existing buildings for establishing baselines, calculat-ing retrofit savings, and implementing model predictive control(data-driven modeling) (Armstrong et al. 2006a; Gayeski et al.2012; Krarti 2010).1. GENERAL CONSIDERATIONS1.1 MODELS AND APPROACHESA mathematical model is a description of th

4、e behavior of a sys-tem. It is made up of three components (Beck and Arnold 1977):Input variables (statisticians call these regressor variables,whereas physicists call them forcing variables), which act on thesystem. There are two types: controllable by the experimenter(e.g., internal gains, thermos

5、tat settings), and uncontrollable (e.g.,climate).System structure and parameters/properties, which providethe necessary physical description of the system (e.g., thermalmass or mechanical properties of the elements).Output (response, or dependent) variables, which describe thereaction of the system

6、to the input variables. Energy use is often aresponse variable.The science of mathematical modeling as applied to physical sys-tems involves determining the third component of a system when theother two components are given or specified. There are two broadbut distinct approaches to modeling: forwar

7、d (classical) and datadriven (inverse). The choice of approach is dictated by the objectiveor purpose of the investigation (Rabl 1988).Forward (Classical) ApproachThe objective is to predict the output variables of a specifiedmodel with known structure and known parameters when subject tospecified i

8、nput variables. To ensure accuracy, models have tended tobecome increasingly detailed. This approach presumes knowledgenot only of the various natural phenomena affecting system behaviorbut also of the magnitude of various interactions (e.g., effective ther-mal mass, heat and mass transfer coefficie

9、nts). The main advantageof this approach is that the system need not be physically built topredict its behavior. The forward-modeling approach is ideal in thepreliminary design when design details are limited.Forward modeling of building energy use begins with a physicaldescription of the building s

10、ystem or component of interest. Forexample, building geometry, geographical location, physical charac-teristics (e.g., wall material and thickness), type of equipment andoperating schedules, type of HVAC system, building operatingschedules, plant equipment, etc., are specified. The peak and averagee

11、nergy use of such a building can then be predicted or simulated bythe forward-simulation model. The primary benefits of this methodare that it is based on sound engineering principles and has gainedwidespread acceptance by the design and professional community.Major simulation codes, such as DOE-2,

12、EnergyPlus, ESP-r, andTRNSYS are based on forward-simulation models.Although procedures for estimating energy requirements varyconsiderably in their degree of complexity, they all have three com-mon elements: calculation of (1) space load, (2) secondary equip-ment load and energy requirements, and (

13、3) primary equipmentenergy requirements. Here, secondary refers to equipment that dis-tributes the heating, cooling, or ventilating medium to conditionedspaces, whereas primary refers to central plant equipment that con-verts fuel or electric energy to heating or cooling effect.The space load is the

14、 amount of energy that must be added to orextracted from a space to maintain thermal comfort. The simplestprocedures assume that the energy required to maintain comfort isonly a function of the outdoor dry-bulb temperature. More detailedmethods consider humidity, solar effects, internal gains, heat

15、andmoisture storage in walls and interiors, and effects of wind on bothbuilding envelope heat transfer and infiltration. The section on Ther-mal Loads Modeling addresses some of these factors. ASHRAEStandard 183 and Chapters 17 and 18 discuss load calculation indetail.Although energy calculations ar

16、e similar to the heating and cool-ing design load calculations used to size equipment, they are not thesame. Energy calculations are based on average use and typicalweather conditions rather than on maximum use and worst-caseweather. Currently, most procedures are based on hourly profiles forclimati

17、c conditions and operational characteristics for a number oftypical days of the year or on 8760 hours of operation per year.The space load is converted to a load on the secondary equipment.This can be a simple estimate of duct or piping losses or gains, or acomplex hour-by-hour simulation of an air

18、system, such as variable-air-volume with outdoor-air cooling. This step must include calcu-lation of all forms of energy required by the secondary system (e.g.,electrical energy to operate fans and/or pumps, energy in heated orchilled water).The secondary equipment load is converted to the fuel and

19、energyrequired by the primary equipment and the peak demand on the util-ity system. It considers equipment efficiencies and part-load charac-teristics. It is often necessary to keep track of the different forms ofenergy, such as electrical, natural gas, and/or oil. In some cases,where calculations a

20、re required to ensure compliance with codes orstandards, these energies must be converted to source energy orresource consumed, as opposed to energy delivered to the buildingboundary.Previously, the steps were performed independently: each stepwas completed for the entire year and hourly results wer

21、e passed tothe next step. Current software usually performs all steps at eachtime interval, allowing effects such as insufficient plant capacity tobe reflected in room conditions.Often, energy calculations lead to an economic analysis to estab-lish the cost effectiveness of efficiency measures (as i

22、n ASHRAEThe preparation of this chapter is assigned to TC 4.7, Energy Calculations.19.2 2017 ASHRAE HandbookFundamentals (SI)Standard 90.1). Thus, thorough energy analysis provides intermedi-ate data, such as time of energy use and maximum demand, so utilitycharges can be accurately estimated. Altho

23、ugh not part of the energycalculations, capital equipment costs should also be estimated toassess the life-cycle costs of alternative efficiency measures.Data-Driven (Inverse) ApproachIn this approach, input and output variables are known and mea-sured, and the objective is to determine a mathematic

24、al description ofthe system and to estimate system parameters. In contrast to the for-ward approach, the data-driven approach is relevant only when thesystem has already been built and actual performance data are avail-able for model development, calibration (see the section on ModelCalibration), an

25、d/or identification. Two types of performance datacan be used: nonintrusive and intrusive. Intrusive data are gatheredunder conditions of predetermined or planned experiments on thesystem to elicit system response under a wider range of system per-formance than would occur under normal system operat

26、ion to allowmore accurate model identification. When constraints on systemoperation do not allow such tests to be performed, the model must beidentified from nonintrusive data obtained under normal operation.The data-driven model has to meet requirements very differentfrom the forward model. The dat

27、a-driven model can only contain arelatively small number of parameters because of the limited andoften repetitive information contained in the performance data. (Forexample, building operation from one day to the next is fairly repet-itive.) It is thus a much simpler model that contains fewer terms

28、rep-resentative of aggregated or macroscopic parameters (e.g., overallbuilding heat loss coefficient and time constants). Because modelparameters are deduced from actual building performance, it is muchmore likely to accurately capture as-built system performance, thusallowing more accurate predicti

29、on of future system behavior undercertain specific circumstances. Performance data collection andmodel formulation need to be appropriately tailored for the specificcircumstance, which often requires a higher level of user skill andexpertise. In general, data-driven models are less flexible than for

30、-ward models in evaluating energy implications of different designand operational alternatives, and so are not substitutes in this regard.To better understand the uses of data-driven models, considersome of the questions that a building professional may ask about anexisting building with known energ

31、y consumption (Rabl 1988):How does energy consumption compare with design predictions(and are any discrepancies caused by anomalous weather, unin-tended building operation, improper operation, as-built deficien-cy, etc.)?How would consumption change if thermostat settings, ventila-tion rates, or ind

32、oor lighting levels were changed?How much energy could be saved by retrofits to the building shell,changes to air handler operation from constant volume (CV) tovariable air volume (VAV), or changes in the various controlsettings?If retrofits are implemented, can one verify that the savings aredue to

33、 the retrofit and not to other causes (e.g., weather)?How can one detect faults in HVAC equipment and optimize con-trol and operation?All these questions are better addressed by the data-drivenapproach. The forward approach could also be used, for example, bygoing back to the blueprints of the build

34、ing and of the HVAC sys-tem, and repeating the analysis performed at the design stage usingactual building schedules and operating modes, but this is tediousand labor intensive, and materials and equipment often perform dif-ferently in reality than as specified. Tuning the forward-simulationmodel is

35、 often problematic, although it is an option (see the sectionon Model Calibration).1.2 OVERALL MODELING STRATEGIESIn developing a simulation model for building energy prediction,two basic issues must be considered: (1) modeling components orsubsystems and (2) overall modeling strategy. Modeling comp

36、o-nents, discussed in the sections on Thermal Loads Modeling andHVAC Component Modeling, results in sets of equations describingthe individual components. The overall modeling strategy refers tothe sequence and procedures used to solve these equations. Theaccuracy of results and the computer resourc

37、es required to achievethese results depend on the modeling strategy.Early building energy programs, including some still in com-mon use, execute load models for every space for every hour of thesimulation period. (Most models of this type use 1 h as the timestep, which excludes any information on ph

38、enomena occurring in ashorter time span.) Next, the program runs models for every sec-ondary system, one at a time, for every hour of the simulation.Finally, the plant simulation model is executed again for the entireperiod.This procedure is shown in Figure 1. Solid lines represent datapassed from o

39、ne model to the next; dashed lines represent informa-tion, usually provided by the user, about one model passed to thepreceding model. For example, the system information may consistof a piecewise-linear function of zone temperature that gives systemcapacity.Because of this loads-systems-plants sequ

40、ence, certain phenom-ena cannot be modeled precisely. For example, if the heat balancemethod for computing loads is used, and some component in thesystem simulation model cannot meet the load, the program canonly report the current load. In actuality, the space temperatureshould readjust until the l

41、oad matches equipment capacity, but thiscannot be modeled, because loads have been precalculated andfixed. If the weighting-factor method is used for loads, this problemis partially overcome, because loads are continually readjustedduring the system simulation. However, the weighting-factor tech-niq

42、ue is based on linear mathematics, and wide departures of roomtemperatures from those used during execution of the load programcan introduce errors.A similar problem arises in plant simulation. For example, in anactual building, as load on the central plant varies, the supplychilled-water temperatur

43、e also varies. This variation in turn affectsthe capacity of secondary system equipment. In an actual building,when the central plant becomes overloaded, space temperaturesshould rise to reduce load. However, with this separate load-system-plant approach, this condition cannot occur; thus, only the

44、overloadcondition can be reported. These are some of the penalties associ-ated with decoupling the load, system, and plant models.More recent building energy programs use an alternative strat-egy, in which all calculations are performed at each time step. Here,the load, system, and plant equations a

45、re solved simultaneously ateach time interval. With this strategy, unmet loads and imbalancescannot occur; conditions at the plant are immediately reflected tothe secondary system and then to the load model, forcing them toreadjust to the instantaneous conditions throughout the building.The results

46、of this modeling strategy are superior, although themagnitude and importance of the improvement are case specific.Fig. 1 Overall Modeling StrategyEnergy Estimating and Modeling Methods 19.3The principal disadvantage of this approach, and the reason thatit was not widely used in the past, is that it

47、demands more computingresources. However, most current desktop computers can now runprograms using the alternative approach in a reasonable amount oftime. Programs that, to one degree or another, implement simulta-neous solution of the loads, system, and plant models have beendeveloped by ESRU (2016

48、), Klein et al. (1994), Park et al. (1985),Taylor et al. (1990, 1991), and U.S. Department of Energy (1996-2016). Some of these programs can simulate the loads, systems, andplants using subhourly time steps.An economic model, as shown in Figure 1, calculates energycosts (and sometimes capital costs)

49、 based on the estimated requiredinput energy. Thus, the simulation model calculates energy use andcost for any given input weather and internal loads. By applying thismodel (i.e., determining output for given inputs) at each hour (orother suitable interval), the hour-by-hour energy consumption andcost can be determined. Maintaining running sums of these quanti-ties yields monthly or annual energy usage and costs.Because detailed models are computationally intensive, severalsimplified methods have been developed, including the degree-day,bin, and correlation methods. See the section on Degree