1、2010 ASHRAE 553This paper is based on findings resulting from ASHRAE Research Project RP-1322.ABSTRACT The goal of this study was to investigate the effects of noise from building mechanical systems with time-varying fluctua-tions on human task performance and perception, and to deter-mine how well
2、current indoor noise rating methods account for this performance and perception. Six different noise conditions with varying degrees of time-varying fluctuations, many focused in the low frequency rumble region, were reproduced in an office-like setting. Thirty participants were asked to complete ty
3、ping, grammatical reasoning, and math tasks plus subjective questionnaires, while being exposed for approxi-mately one hour to each noise condition. Results show that the noise conditions with higher sound levels (greater than 50 dBA) combined with excessive low frequency rumble as well as those wit
4、h larger timescale fluctuations (i.e., a heat pump cycling on and off every 30 seconds) were generally perceived to be more annoying than the other signals tested, although statistically significant negative relationships to task perfor-mance were not found. Other findings are (1) that the noise cha
5、racteristics most closely correlated to higher annoyance/distraction responses in this study were higher ratings of loud-ness followed by roar, rumble, and changes in time; and (2) that perception of more low frequency rumble in particular was significantly linked to reduced performance on cognitive
6、ly demanding tasks. As for the ability of current indoor noise rating systems to match human performance or perception, none of the indoor noise rating methods evaluated were signif-icantly correlated to task performance, but aspects of subjec-tive perception such as loudness ratings were statistica
7、lly related. Spectral quality ratings included with some noise rating methodologies were inconsistent with subjective perception, but other metrics such as RNC, L1 L99 LF ave, and LCeq LAeq, were strongly correlated to rumble perception. The authors use the results to suggest a framework for an idea
8、l indoor noise rating method, but further research is required towards quantifying specific guidelines for accept-able degrees of time-varying fluctuations and tonalness.INTRODUCTIONMechanical systems responsible for heating, ventilation and air-conditioning are sources of background noise in build-
9、ings. Acceptable noise level guidelines have been suggested using a number of indoor noise rating methodologies proposed over the past 60 years, such as Noise Criteria (NC), Room Criteria (RC) and Room Criteria Mark II (RC-Mark II) (ASHRAE 2007). There is some debate about which noise rating system
10、should be advocated by ASHRAE, as the vari-ous methodologies do not always give the same assessment. Furthermore, experience in the field suggests that these noise rating systems do not account well for time-varying fluctua-tions that can occur with modern mechanical systems. The fluctuations may be
11、 due to ill-designed systems that demon-strate surging and excessive low frequency rumble, or may be on a larger timescale where the systems settings change over time, such as variable air volume systems or systems switch-ing on and off. This research project investigates the effects of noise with t
12、ime-varying fluctuations on human performance and perception, and correlates these findings with current indoor noise rating methods. An earlier phase of the work focused on the effects of noise with varying degrees of tones, Human Performance and Perception-BasedEvaluations of Indoor Noise Criteria
13、 forRating Mechanical System Noise withTime-Varying FluctuationsLily M. Wang, PhD, PE Cathleen C. NovakMember ASHRAELily M. Wang is an associate professor in the Durham School of Architectural Engineering and Construction, University of NebraskaLincoln, Omaha, NE. Cathleen C. Novak is a consultant w
14、ith PMK Consultants, Dallas, TX.AB-10-019 (RP-1322)2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions (2010, Vol. 116, Part 2). For personal use only. Additional reproduction, distribution, or transmission in eithe
15、r print or digital form is not permitted without ASHRAEs prior written permission.554 ASHRAE Transactionsanother problem produced by modern mechanical systems (Ryherd and Wang 2010). The results have collectively been used to outline a framework that an ideal indoor noise rating method should follow
16、, as described later in this paper.Much research has been conducted regarding the effects of noise on human perception and performance; reviews of such work may be found in Kryter (1985), Jones and Broad-bent (1998), and the accompanying paper by Ryherd and Wang (2010). One consistent finding from t
17、he previous work is that while sound level is certainly an important factor, spec-tral characteristics of the noise also affect human perception and performance. In particular, noise with excessive low frequency energy or rumble has been shown in the lab and in the field to result in greater annoyan
18、ce than equivalently loud signals without rumble (Berglund et al. 1996, Leventhall 2003, Persson et al. 1985, Persson and Bjrkman 1988, Pers-son Waye et al. 2001, Persson Waye and Rylander 2001). Bradley (1994) reported an investigation in which subjects adjusted the level of an amplitude-modulated
19、signal (simulat-ing rumble) to be equivalently annoying to a reference signal with a neutral spectrum. He found that both level and the modulation frequency of the stimulus could negatively impact the perceived annoyance. Noise conditions with rumble can also result in degradation in task performanc
20、e, as shown by a number of researchers, although many of these studies compared only a few signals at a time (Kyriakides and Leven-thall 1977, Landstrm et al. 1991, Holmberg et al. 1993, Pers-son Waye et al. 1997, 2001). To quantify the degree of low frequency content and possibly predict the result
21、ing annoyance, Broner and Leven-thall (1983) proposed using the difference between the C-weighted equivalent sound pressure level and the A-weighted equivalent sound pressure level, LCeq LAeq(often referred to as dBC dBA), since the A-weighting curve corrects more severely for low frequency componen
22、ts than the C-weighting curve. They suggested that values of LCeq LAeqgreater than 20 dB would signify a low frequency noise problem. Holm-berg et al. (1996) correlated a number of metrics that were easily calculated by a sound level meter to the perception of annoyance from low frequency noise, inc
23、luding LCeq LAeq, and found that this particular metric did differentiate between annoying and non-annoying cases. Holmberg et al. (1997) later suggested that a value of 15 dB or greater could indicate the potential for low frequency noise problems. Kjellberg et al. (1997) conducted office surveys a
24、nd suggested that LCeq LAeqmay be limited as a predictor of annoyance, particularly at lower overall noise levels when the low frequency content was not as perceptible. A metric that has been proposed for quantifying more specifically the degree of time-varying fluctuations in a signal is the differ
25、ence between two statistical sound level measures, such as L10 L90, where L10 is the sound level exceeded 10% of the time and L90 is the sound level exceeded 90% of the time (Blazier and Ebbing 1992). More recently, Mann et al. (2007) utilized another variation, L1 L99, to quantify time-varying fluc
26、tuations during their ASHRAE 1219-RP project which sought to quantify duct rumble noise resulting from various aerodynamic system effects at the discharge of a centrifugal fan.The most recent version of ANSI Standard S12.2 “Crite-ria for Evaluating Room Noise” (2008) includes a methodol-ogy known as
27、 Room Noise Criteria (RNC), originally proposed by Schomer (2000), that explicitly attempts to coun-ter the indoor noise rating methods NC and RCs deficiencies in dealing with low frequency fluctuations. RNC requires calculations of Lmax Leq, energy averaged for the octave bands from 16 Hz to 63 Hz,
28、 as well as at the 125 Hz octave band, to provide an indication of whether low frequency fluc-tuations are a problem. In cases where surging or low frequency fluctuations are indicated, the RNC method essen-tially calculates a penalty to add to the levels in the lowest frequency bands, and then a ta
29、ngency method is applied to determine the final rating; otherwise, it defaults to the NC method described in the standard. The octave band in which the tangency is met is to be reported with the RNC value. Schomer and Bradley (2000) applied the RNC ratings to the findings from Bradleys previous stud
30、y on annoyance due to amplitude-modulated signals (1994) and found that the RNC methodology was validated by those results.There has been some work on other random (or aperi-odic) time-varying fluctuations, such as those that may occur on longer timescales than low frequency rumble (e.g. systems swi
31、tching on and off) or involving different content (e.g. office noise including speech, equipment noise, etc.). Eschenbrenner (1971) compared the effects of continuous periodic and aperiodic noise, and found that the aperiodic noise reduced the performance times on a visual tracking task, although ex
32、posure times were brief in this study. In Weinsteins 1977 study, subjects completed a proofreading task while listening either to a recording of radio news or in quiet, and the results were mixed; detection of grammatical errors decreased in noise, but speed and detection of spelling errors were not
33、 significantly affected. Recently, Witterseh et al. (2004) investigated human perception and performance over a three-hour period in an open-office type environment due to various combinations of three thermal and two acous-tic conditions. They found that the office noise condition (55 dBA) which in
34、cluded a great deal of aperiodic content resulted in increased fatigue and decreased performance in comparison to a quiet noise condition (35 dBA). The current investigation is focused on how both human perception and performance are affected by noise produced from mechanical systems in buildings th
35、at feature time-vary-ing fluctuations, primarily in the low frequency region or on a larger timescale. The project involved systematically expos-ing participants to six different noise signals over a period of one hour each and gauging their performance on three types of tasks (typing, grammatical r
36、easoning and math tests) and their perception via subjective questionnaires. The results have then been related to commonly used indoor noise rating systems, 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions (2010
37、, Vol. 116, Part 2). For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission.2010 ASHRAE 555suggested within the ASHRAE Applications Handbook, in an effort to improve those methods. METHOD
38、OLOGYThe protocol described in this section for this phase of research is similar to one used for an earlier phase of testing, presented in an accompanying paper (Ryherd and Wang 2010). As the authors believe that readers may not necessarily access both papers, some of the same methodology is discus
39、sed in both manuscripts.Thirty test subjects (15 males and 15 females) from the University of Nebraska community were recruited to partici-pate in this study, ranging in age from 19 to 61 with a mean of 22.6 years. All participants first underwent a series of pre-test screens to gauge the subjects v
40、ision, hearing, and typing skills. The minimum requirements to participate in the study were as follows: normal vision as verified by a Keystone Opthalmic Telebinocular, hearing thresholds below 25 dB hearing level in octave bands from 125 Hz to 8 kHz, and a minimum typing speed of 20 wpm. None of t
41、hese participants had participated in the earlier phase of the ASHRAE 1322-RP project regarding tonal noise conditions (Ryherd and Wang 2010).Testing was conducted in a 906 ft (25.7 m) indoor envi-ronmental test chamber at the University of Nebraska, outfit-ted as a typical office with two desks, ca
42、rpet, gypsum board walls, and acoustical ceiling tile. The test chambers envelope has a high sound transmission class of STC 47, and its interior acoustic condition demonstrates low background noise level of RC 26(H) (or an equivalent A-weighted sound level of 35 dBA) and a low reverberation time of
43、 0.25 sec at 500 Hz. During all tests, the test chamber was thermally controlled to maintain a temperature of 72F (22C). Overhead fluorescent lighting provided an constant average illuminance of 71 foot-candles (764 lux) at the work plane. The sound in the test chamber was the only environmental cha
44、racteristic that changed between test sessions, with the signals being presented in an inconspicuous manner over two loudspeakers: (i) an Armstrong i-ceiling loudspeaker which has the same appearance as the other ceiling tiles in the room, and (ii) a JBL Northridge E250P subwoofer, disguised to rese
45、mble an endta-ble in the corner of the room. The test administrator and vari-ous equipment (e.g. the hard drive to the test computers and other audio gear) were located in a control room, adjacent to the chamber. A repeated measures test design was used in which each subject was exposed to the same
46、six noise conditions, each for a period of 55 minutes at a time. This length of exposure time was selected due to the results from a previous phase of the ASHRAE 1322-RP project (Ryherd and Wang 2007). Partic-ipants were asked to come for their six listening sessions at approximately the same timesl
47、ot on different days. For each session, the test subjects spent the first 25 minutes adapting to the noise condition and completing a test on paper, developed from material taken from the verbal portion of the Graduate Record Examination (GRE). Unbeknownst to the subject, this material was not to be
48、 marked but was simply to keep the subject mentally alert during the adaptation period. The next 15 minutes consisted first of three skill tests, administered on a computer using SkillCheck software: typing, grammatical reasoning, and math. The typing test was allotted five minutes, and involved typ
49、ing a passage from a piece of paper with the mouse disabled. The reasoning task was allotted two minutes, and included 20 questions in which subjects indicated whether a statement regarding a presented sequence of letters was true or false. The math test was allotted seven minutes, and included 11 problems involving the four basic functions with integers, fractions, and decimals, presented either mathematically or as a word problem. Partic-ipants were provided with pencil and paper but no calculator. Results for the typing test were output as an adjusted
