ASHRAE OR-05-13-1-2005 Detecting Critical Supply Duct Pressure《供应管道压力的关键检测》.pdf

1、OR-05-1 3-1 Detecting Critical Supply Duct Pressure Clifford C. Federspiel, PhD Associate Member ASHRAE ABSTRACT Fan energy use in variable-air-volume (VAV) systems can be reduced by resetting the supply ductpressure. The standard way to reset ductpressure is by controlling the most open termi- nal

2、damper to a nearly open position. This strategu is rarely used because of a variety of issues including sensing limita- tions, network bandwidth, and stability This paper describes the development of a new method of determining the critical supply ductpressure for VAV systems. The method relies on a

3、 short, simple functional test and a data processing technique that is based on a simple model of the system behaviol: The method can be implemented during normal system operation, and it could be automated. Thesystem model includes the eflect of duct leakage, which ofers thepotential for dual use a

4、s a duct leakage diagnostic. Results from experiments on a laboratory- scale system demonstrate good accuracy for determining crit- ical pressure and moderate accuracy for determining duct leakage. Results from experiments on two commercial air- handling units demonstrate that the method is practica

5、l and that it ofers the potential for large energy savings. INTRODUCTION Every year the US consumes 0.75 quadrillion Btus of primary energy to move air in buildings (DOE 2000). This paper describes technology to reduce that energy consump- tion. The focus is on variable-air-volume (VAV) heating, ven

6、tilating, and air-conditioning (HVAC) systems, which condition 29% of the floor space in commercial buildings (EIA 1999). The standard way to control VAV fan systems is to regu- late the static pressure in the main supply duct. This strategy ensures that zone terminals have enough pressure to operat

7、e properly, but it is inefficient because the pressure setpoint will be higher than necessary all of the time. Considerable energy savings can be achieved if the supply duct pressure is reduced at part load. Lorenzetti and Norford (1994) showed that fan energy consumption in VAV HVAC systems could b

8、e reduced by 19% to 42% with static pressure reset (SPR). They also showed that SPR results in a 20% lower pressure at design load conditions. Federspiel (2003) found that reducing the pressure in response to reduced supply airflow could reduce fan power consumption by 26% and cooling power by 17% (

9、due to reduced leakage). This strategy is called static pressure adjust- ment from volume flow (SAV). The standard way to reset the pressure is to use a feedback loop that regulates the most open terminal damper to a nearly open position (e.g., 90% open) by adjusting the static pressure setpoint. Th

10、ere are a number of variants of this method. Static pressure reset (SPR) has been in existence for more than 15 years, and it is now required by ASHRAE Standard 90.1 when the VAV terminals have digital controls, but it is still not widely used. SPR is not widely used for the following reasons: 1. SP

11、R requires a networked digital control system. 2. SPR requires digital controls on zone terminal units. 3. Some SPR strategies require terminal damper position sensors. 4. SPR adds to the complexity of control software. 5. SPR strategies that use feedback are difficult to tune. Even today, many new

12、systems do not have digital termi- nal controls, and it is very uncommon for terminals to have position sensors on the terminal dampers. However, in legacy systems, it is much less common for systems to meet the control and sensing infrastructure requirements of standard Clifford Federspiel is princ

13、ipal at Federspiel Controls, LLC, El Cerrito, Calif. 02005 ASHRAE. 957 SPR strategies. Engineers have solved this problem by invent- ing ad hoc resetting strategies that reset static pressure based on some measurable quantity that is related to the load. For example, static pressure may be reset bas

14、ed on time, outdoor temperature, flow, or a combination of these. Ad hoc resetting has the advantage of not requiring digital terminal controls. They also cannot destabilize the static pressure loop because they do not involve feedback. However, they must still be configured, and there is no way to

15、do this today except through trial and error or good engineering judgment. This paper involved the development of a technique that can be used to configure ad hoc SPR strategies so that they yield nearly optimal performance. We developed a new method of determining the critical supply duct pressure

16、for VAV systems. The new method only requires measurement of supply duct static pressure and supply airflow rate. It relies on a short, simple functional test and a simple model of the system behavior. This functional test could be implemented during normal system operation, and it could be automate

17、d. The system model includes the effect of duct leakage, which offers the potential for dual use as a duct leakage diagnostic. This is important because duct leakage is a significant contrib- utor to inefficiency of air-handling equipment. Xu et al. (2002) found that the air leakage ratios in five l

18、arge HVAC systems were as high as one-third of the fan-supplied airflow. Results from experiments on a laboratory-scale system demonstrate 6. (“mi Skew - q2 go o3 ZJ 05 -a -J(Ji n a- good accuracy for determining critical pressure and moderate accuracy for determining duct leakage. Results from expe

19、ri- ments on two working VAV air-handling units demonstrate the potential for energy savings. The functional test and the corre- sponding analytical method for determining critical pressure and leakage are called “infer critical information about termi- nals (InCITE). InCITE can form the basis for c

20、onfiguring ad hoc SPR strategies so that they deliver nearly optimal perfor- mance. METHODS Test Stand We designed and constructed a test stand that has a vari- able-speed fan and a main duct supplying four VAV terminals. Each terminal duct terminated with a commonly used diffuser. We designed the m

21、ain supply duct so that we could introduce a pressure drop between the terminal duct branch points that emulated the frictional and minor losses in a long duct without actually requiring a long duct. The test stand included a computer-based data acquisition system, a pitot tube array to measure supp

22、ly airflow, and a supply duct pressure sensor. Figure 1 shows a schematic diagram ofthe test stand duct- work downstream of the supply fan. The fan is powered by a 3 hp (2.25 kW), three-phase motor. The motor is driven by a three-phase inverter so that the speed can be adjusted contin- I I Note: Rou

23、nd duck are located ai the veftical centerline of ihe main duct sections. I section 10. (25 4cm) sleeve 10 (25 4 Cm) VAV box x Figure 1 A schematic of the test stand ductwork downstream of the supply fan. 958 LZ cm) deve w(rn.3cm VAV bol ASHRAE Transactions: Symposia uously. The duct upstream of the

24、 supply fan is 15 by 15 in. (38.1 by 38.1 cm), and it contains the pitot tube array used to measure supply flow. We placed screens covered with strips of tape in the main supply duct between each terminal duct branch. The tape was applied to achieve pressure drops comparable to a system with a much

25、longer supply duct. We cut eight holes in the main supply duct to emulate leakage. Each hole was square-shaped with an area of 4 in.2. The four terminal ducts included two 6 in. (1 5 cm), pres- sure-independent VAV terminals, one 8 in. (20 cm), pressure- independent VAV terminal, and one 10 in. (25

26、cm), pressure- independent VAV terminal. The 6 in. and 8 in. VAV terminals were manufactured by one company, while the 10 in. VAV terminal was manufactured by another. One of the 6 in. VAV terminals included a hot-water reheat coil that was used for its pressure-drop characteristic. The four termina

27、l ducts were each terminated with a single, round diffuser. We installed an array of five pitot tubes in the duct upstream of the fan. Each pitot tube was equipped with a sepa- rate pressure sensor. The same kind of pressure sensors were used to measure velocity through each VAV terminal. We used th

28、e flow pickups installed in the VAV terminals to measure the flow. We cali- brated each flow pickup using a pitot tube traverse with ten points spaced according to the Log-Tchebychev method (ASHRAE 1988; Klaassen and House 2001). We installed a floating actuator with position feedback on each VAV te

29、rminal. The position sensors are linear-taper potentiometers. The test stand was controlled with a desktop computer. A data acquisition card with a 12-bit A/D converter was used to measure sensor voltages. An application was written to perform all of the experiments automatically using a commer- cia

30、lly available data acquisition software package. Experiments We conducted a large set of laboratory experiments and two field experiments. Laboratory Experiments. We conducted 60 experi- ments consisting of four sensor locations, three leakage rates, and five rates of change of the terminal flow rat

31、es. The supply duct static pressure sensor locations were near the fan, one- third of the way down the supply duct, two-thirds of the way down the supply duct, and at the end of the supply duct. Leak- age areas cut into the supply duct for each of the three leakage rates corresponded to O in.2, 16 i

32、n.2 (103 cm2), and 32 in.2 (206 cm2). The leakage areas were evenly distributed among the four sections of the supply duct. The rates of change were derived from trend logs from a building in Oakland, CA. Figure 2 shows the trend of supply airflow rates for that build- ing. For the slowly increasing

33、 rates of change, we used the average rate of change in the middle of the day as the loads and supply flow increased in response to increasing outdoor temperature and increasing load. We set the average flows for each terminal to the average of the minimum and maximum flows for the same-sized termin

34、als from the Oakland building. At the beginning of each test, the data acquisition appli- cation performed a zero calibration of every sensor, and a span calibration of the position sensors. The application then proceeded to control the supply duct pressure and terminal velocities to the setpoints f

35、or the specific test. Each set of setpoints was held for two minutes. Time-stamped data were recorded every 250 milliseconds throughout every test. Field Experiments. We conducted experiments on two working air-handling units in two different buildings at a university campus. We refer to these units

36、 as AHU-3 and AHU-6. AHU-3 is a 60,000 cfm (28 m3/s) air-handling unit that serves the west side of four floors. This unit is normally operated at a duct static pressure of 1.5 in. W.C. (325 Pa). The air-handling unit controls include a supply flow station, which we used for the experiment. The term

37、inal controls are pres- sure-independent and electro-pneumatic. Pneumatic control is used to regulate the terminal flow. An electric-to-pneumatic (EP) transducer is used to command the setpoint of the flow loop based on the output of the temperature control loop. The temperature control loops are di

38、gital. From the operator work- station, space temperature in each zone and the command to the EP transducer could be read and archived. We archived fan speed, supply flow, duct pressure, duct pressure setpoint, and the EP command and space temperature for the zones on each floor that were closest to

39、 the top and bottom of the range of the EP transducer span. We changed the supply duct pressure setpoint in increments of 0.1 in. W.C. (25 Pa), starting at 1.5 in. W.C. (325 Pa). We started the test at 8:30 a.m. and concluded it at 12:30 p.m. 50000 , 250 45000 40000 35000 30000 25000 20000 15000 100

40、00 5000 O 200 cn 8 100 -p o E 150 E O. cn a 50 O Figure 2 The trend of supply airflow rates for a building in Oakland. CA. ASHRAE Transactions: Symposia 959 AHU-6 is a 12,000 cfm (5.7 m3/s) air-handling unit that serves a computer room and adjacent offices. AHU-6 is normally operated at 0.6 in. W.C.

41、 (150 Pa). This unit does not have a supply flow station, so we installed a pitot tube in the supply duct and recorded the velocity pressure with a laptop computer. All of the terminal controls are pneumatic. The only temperature that could be read and archived was the return temperature. We archive

42、d fan speed, supply flow, supply duct pressure, supply duct pressure setpoint, and return tempera- ture. We started the test at 9 a.m. andconcludedit at 12:30 p.m. Analysis We determined the pressure-flow exponent that produced the highest accuracy. We assessed how accurately the proposed innovation

43、 detects the critical pressure and also how accurately it estimated the leakage flow rate. We determined the measurement noise and filtering requirements necessary to achieve the desired accuracy. The detector is based on a dual-mode model of a VAV air- handling system. The two modes are “controllin

44、g” and “starved.” The supply fan in most VAV air-handling systems is used to regulate the static pressure at apoint in the supply duct. The static pressure should be sufficiently high that all termi- nals served by the air-handling unit get enough air to meet the load. If it is too high, then even t

45、he most open VAV terminal will be throttling considerably, and energy will be wasted. Lorenzetti and Norford (1994) showed that in constant-pres- sure systems, the pressure setpoint is usually too high even for meeting on-peak load conditions. The critical supply duct pressure occurs when the most o

46、pen VAV terminal is 100% open and meeting the load. When the duct pressure is high enough that all of the terminals are meeting the load, the system is operating in the controlling mode. When one or more terminal dampers are 100% open and not meeting the load, the system is in the starved mode. The

47、lowest supply duct pressure that keeps all the terminals in control is called the critical pres- sure. The controlling model contains three terms. The first is a constant term that represents the cumulative flow rate through the dampers at the beginning of the functional test used to determine the c

48、ritical pressure. The second is a term to account for duct leakage, which can be very significant in some systems. The third is a time-dependent term that accounts for the fact that the loads, and therefore the supply flow, may change over the course of the functional test if it is conducted during

49、normal operation. The mathematical equations for the two models are described in detail in Federspiel (2004). When the pressure drops below the critical pressure, the pressure-flow model described above must be changed. In the starved mode, we model the “constant” term (first term described above) as a quadratic function of pressure. There is no physical basis for this model, but we find that it works well. We used the following test procedure to collect data: 1. Start at a sufficiently high pressure. 2. Wait for terminals to reach equilibrium. 3. Take a reading of supply flow and static