1、7.1CHAPTER 7FUNDAMENTALS OF CONTROLTerminology . 7.1Types of Control Action 7.2Classification by Energy Source 7.4CONTROL COMPONENTS 7.4Controlled Devices. 7.4Sensors . 7.8Controllers . 7.10Auxiliary Control Devices 7.11COMMUNICATION NETWORKS FOR BUILDING AUTOMATION SYSTEMS 7.14Communication Protoco
2、ls 7.14OSI Network Model 7.14Network Structure. 7.14Specifying BAS Networks . 7.17Approaches to Interoperability. 7.17SPECIFYING BUILDING AUTOMATION SYSTEMS. 7.18COMMISSIONING 7.18Tuning. 7.18Codes and Standards. 7.20UTOMATIC HVAC control systems are designed to maintainA temperature, humidity, pres
3、sure, energy use, power, lighting lev-els, and safe levels of indoor contaminants. Automatic control primar-ily modulates actuators; stages modes of action; or sequences themechanical and electrical equipment on and off to satisfy load require-ments, provide safe equipment operation, and maintain sa
4、fe buildingcontaminant levels. Automatic control systems can use digital, pneu-matic, mechanical, electrical, and electronic control devices. Humanintervention often involves scheduling equipment operation and ad-justing control set points, but also includes tracking trends and pro-gramming control
5、logic algorithms to fulfill building needs.This chapter focuses on the fundamental concepts and devices nor-mally used by a control system designer. It covers (1) control funda-mentals, including terminology; (2) types of control components; (3)methods of connecting components to form various indivi
6、dual controlloops, subsystems, or networks; and (4) commissioning and opera-tion. Chapter 47 of the 2011 ASHRAE HandbookHVAC Applica-tions discusses the design of controls for specific HVAC applications.TERMINOLOGYAn open-loop control does not have a direct feedback linkbetween the value of the cont
7、rolled variable and the controller.Open-loop control anticipates the effect of an external variable onthe system and adjusts the set point to avoid excessive offset. Anexample is an outdoor thermostat arranged to control heat to a build-ing in proportion to the calculated load caused by changes in o
8、utdoortemperature. In essence, the designer presumes a fixed relationshipbetween outside air temperature and the buildings heat requirement,and specifies control action based on the outdoor air temperature.The actual space temperature has no effect on this controller.Because there is no feedback on
9、the controlled variable (space tem-perature), the control is an open loop.A closed-loop or feedback control measures actual changes inthe controlled variable and actuates the controlled device to bringabout a change. The corrective action may continue until the con-trolled variable is at setpoint or
10、 within a prescribed tolerance. Thisarrangement of having the controller respond to the value of the con-trolled variable is known as feedback.Every closed loop must contain a sensor, a controller, and a con-trolled device. Figure 1 illustrates the components of the typical con-trol loop. The sensor
11、 measures the controlled variable and transmitsto the controller a signal (pneumatic, electric, or electronic) havinga pressure, voltage, or current value related by a known function tothe value of the variable being measured. The controller comparesthis value with the set point and signals to the c
12、ontrolled device forcorrective action. A controller can be hardware or software. A hard-ware controller is an analog device (e.g., thermostat, humidistat,pressure control) that continuously receives and acts on data. A soft-ware controller is a digital device (e.g., digital algorithm) thatreceives a
13、nd acts on data on a sample-rate basis. The controlleddevice is typically a valve, damper, heating element, or variable-speed drive.The set point is the desired value of the controlled variable. Thecontroller seeks to maintain this set point. The controlled devicereacts to signals from the controlle
14、r to vary the control agent.The control agent is the medium manipulated by the controlleddevice. It may be air or gas flowing through a damper; gas, steam, orwater flowing through a valve; or an electric current.The process is the HVAC apparatus being controlled, such as acoil, fan, or humidifier. I
15、t reacts to the control agents output andeffects the change in the controlled variable.The controlled variable is the temperature, humidity, pressure,or other condition being controlled.A control loop can be represented in the form of a block dia-gram, in which each component is modeled and represen
16、ted in itsown block. Figure 2 is a block diagram of the control loop shown inFigure 1. Information flow from one component to the next isshown by lines between the blocks. The figure shows the set pointbeing compared to the controlled variable. The difference is theerror. If the error persists, it m
17、ay be called offset drift, deviation,droop, or steady-state error. The error is fed into the controller,which sends an output signal to the controlled device (in this case, avalve that can change the amount of steam flow through the coil ofFigure 1). The amount of steam flow is the input to the next
18、 block,which represents the process. From the process block comes thecontrolled variable, which is temperature. The controlled variable isThe preparation of this chapter is assigned to TC 1.4, Control Theory andApplication.Fig. 1 Example of Feedback Control: Discharge Air Temperature Control7.2 2013
19、 ASHRAE HandbookFundamentalssensed by the sensing element and fed to the controller as feedback,completing the loop.Control loop performance is greatly affected by time lags, whichare delay periods associated with seeing a control agent changereflected in the desired end-point condition. Time lags c
20、an causecontrol and modeling problems and should be understood and evalu-ated carefully. There are two types of time lags: first-order lags anddead time.First-order lags involve the time it takes for the change to beabsorbed by the system. If heat is supplied to a cold room, the roomheats up gradual
21、ly, even though heat may be applied at the maxi-mum rate. The time constant is the unit of measure used to describefirst-order lags and is defined as the time it takes for the controlledvariable of a first-order, linear system to reach 63.2% of its finalvalue when a step change in the input occurs.
22、Components withsmall time constants alter their output rapidly to reflect changes inthe input; components with a larger time constant are sluggish inresponding to input changes.Dead time (or time lag) is the time from when a change in thecontroller output is made to when the controlled variable exhi
23、bits ameasurable response. Dead time can occur in the control loop ofFigure 1 because of the transportation time of the air from the coilto the space. After a coil temperature changes, there is dead timewhile the supply air travels the distribution system and finallyreaches the sensor in the space.
24、The mass of air in the space furtherdelays the coil temperature changes effect on the controlled vari-able (space temperature). Dead time can also be caused by a slowsensor or a time lag in the signal from the controller when it firstbegins to affect the output of the process. Dead time is most ofte
25、nassociated with the time it takes to transport the media changed by thecontrol agent from one place to another. Dead time may also be inad-vertently added to a control loop by a digital controller with an exces-sive scan time. If the dead time is small, it may be ignored in thecontrol system model;
26、 if it is significant, it must be considered.Figure 1 depicts the mechanisms that create both first-order anddead-time lags, and Figure 3 shows the effect related to time. Deadtime is the time it takes warmer air resulting from a higher set pointto reach the space, followed by the first-order lag cr
27、eated by the wallon which the thermostat is mounted, and that of the temperaturesensor (all of which warm gradually rather than all at once). Thecontrol loop must be tuned to account for the combined effect ofeach time lag. Note that, in most HVAC systems, the first-order lagelement predominates.The
28、 gain of a transfer function is the amount the output of thecomponent changes for a given change of input under steady-stateconditions. If the element (valve, damper, and/or temperature/pres-sure differential) is linear, its gain remains constant. However, manycontrol components are nonlinear and ha
29、ve gains that depend on theoperating conditions. Figure 3 shows the response of the first-order-plus-dead-time process to a step change of the input signal. Notethat the process shows no reaction during dead time, followed by aresponse that resembles a first-order exponential.TYPES OF CONTROL ACTION
30、Control loops can be classified by the adjustability of the con-trolled device. A two-position controlled device has two operatingstates (e.g., open and closed), whereas a modulating controlled de-vice has a continuous range of operating states (e.g., 0 to 100% open).Two-Position ActionThe control d
31、evice shown in Figure 4 can be positioned only toa maximum or minimum state (i.e., on or off). Because two-positioncontrol is simple and inexpensive, it is used extensively for bothindustrial and commercial control. A typical home thermostat thatstarts and stops a furnace is an example.Controller di
32、fferential, as it applies to two-position controlaction, is the difference between a setting at which the controlleroperates to one position and a setting at which it operates to theother. Thermostat ratings usually refer to the differential (indegrees) that becomes apparent by raising and lowering
33、the dial set-ting. This differential is known as the manual differential of thethermostat. When the same thermostat is applied to an operatingsystem, the total change in temperature that occurs between a “turn-on” state and a “turn-off” state is usually different from the mechan-ical differential. T
34、he operating differential may be greater becauseof thermostat lag or hysteresis, or less because of heating or coolinganticipators built into the thermostat.Anticipation Applied to Two-Position Action. This commonvariation of strictly two-position action is often used on room ther-mostats to reduce
35、the operating differential. In heating thermostats,a heater element in the thermostat is energized during on periods,thus shortening the on time because the heater warms the thermostat(heat anticipation). The same anticipation action can be obtainedin cooling thermostats by energizing a heater therm
36、ostat at off peri-ods. In both cases, the percentage of on time is varied in proportionto the load, and the total cycle time remains relatively constant.Modulating ControlWith modulating control, the controllers output of the controllercan vary over its entire range. The following terms are used tod
37、escribe this type of control:Throttling range is the amount of change in the controlled vari-able required to cause the controller to move the controlled devicefrom one extreme to the other. It can be adjusted to meet jobrequirements. The throttling range is inversely proportional toproportional gai
38、nFig. 2 Block Diagram of Discharge Air Temperature ControlFig. 3 Process Subjected to Step Input Fig. 4 Two-Position ControlFundamentals of Control 7.3Control point is the actual value of the controlled variable atwhich the instrument is controlling. It varies within the controllersthrottling range
39、and changes with changing load on the system andother variables.Offset, or error signal, is the difference between the set point andactual control point under stable conditions. This is sometimescalled drift, deviation, droop, or steady-state error.In each of the following examples of modulating con
40、trol, there isa set of parameters that quantifies the controllers response. Thevalues of these parameters affect the control loops speed, stability,and accuracy. In every case, control loop performance depends onmatching (or tuning) the parameter values to the characteristics ofthe system under cont
41、rol.Proportional Control. In proportional control, the controlleddevice is positioned proportionally in response to changes in thecontrolled variable (Figure 5). A proportional controller can bedescribed mathematically byVp= Kpe + Vo(1)whereVp= controller outputKp= proportional gain parameter (inver
42、sely proportional to throttling range)e = error signal or offsetVo= offset adjustment parameterThe controller output is proportional to the difference between thesensed value, the controlled variable, and its set point. The con-trolled device is normally adjusted to be in the middle of its controlra
43、nge at set point by using an offset adjustment. This control is sim-ilar to that shown in Figure 5.Proportional plus Integral (PI) Control. PI control improveson simple proportional control by adding another component to thecontrol action that eliminates the offset typical of proportional con-trol (
44、Figure 6). Reset action may be described byVp= Kpe + Ki+ Vo(2)whereKi= integral gain parameter = timeThe second term in Equation (2) implies that the longer error e exists,the more the controller output changes in attempting to eliminate theerror. Proper selection of proportional and integral gain c
45、onstantsincreases stability and eliminates offset, giving greater control accu-racy.Proportional-Integral-Derivative (PID) Control. This is PIcontrol with a derivative term added to the controller. It varies withthe value of the derivative of the error. The equation for PID control isVp= Kpe + Ki+ V
46、o(3)whereKd= derivative gain parameter of controllerde/d = time derivative of errorAdding the derivative term gives some anticipatory action to thecontroller, which results in a faster response and greater stability.However, the derivative term also makes the controller more sensi-tive to noisy sign
47、als and harder to tune than a PI controller. MostHVAC control loops perform satisfactorily with PI control alone.Adaptive Control. An adaptive controller adjusts the parametersthat define its response as the dynamic characteristics of the processchange. If the controller is PID based, then it adjust
48、s feedback gains.An adaptive controller may be based on other feedback rules. Thekey is that it adjusts its parameters to match the characteristics of theprocess. When the process changes, the tuning parameters change tomatch it. Adaptive control is applied in HVAC systems because nor-mal variations
49、 in operating conditions affect the characteristics rel-evant to tuning. For instance, the extent to which zone dampers areopen or closed in a VAV system affects the way duct pressureresponds to fan speed, and entering fluid temperatures at a coil affectthe way the leaving temperature responds to the valve position.Fuzzy Logic. This type of control offers an alternative to tradi-tional control algorithms. A fuzzy logic controller uses a series of“if-then” rules that emulates the way a human operator might con-trol the process. Examples of fuzzy logic might includeIF room temperat
