Noise Reduction in Neonatal ICU Paul Baker, Rogelio Grijalva, Elton Leung, Roberto Rios-Rios University of California Merced Mechanical Engineering Capstone Design Spring 2012 Neonatal Acoustic Solutions Mission Statement Neonatal Acoustic Solution’s purpose, in cooperation with Children’s Hospital Central California (CHCC), is to find methods to reduce noise levels to 55 dB in the neonatal ICU through the application of existing technologies, creation of new products, and alteration of patterns of behavior with a focus on efficiency and cost effectiveness, while minimizing the impact on hospital staff. Theory Introduction Problem: Excessive noise in the CHCC Neonatal ICU – OSHA standard mandates 55dB Max. Background: Medical research has demonstrated the importance of noise regulation in a hospital setting. Noise can be harmful to the development of fragile neonates and has been shown to increase recovery times. A previous study done by the hospital showed an average noise level of 66-71dB. However, there were three main concerns in our designs: they need to be safe for the babies, financially viable, and should not prevent the hospital staff from performing their duties. Investigation: Course of Action: Our initial task was to measure the sound levels for ourselves. Some data from that experiment is to the right. We discovered that there were two main sources of noise: equipment and staff. The equipment produced low frequency periodic noise whereas the staff were responsible for intermittent wide band bursts. These represent very different types of noise pollution and demanded separate solutions. We decided upon a three pronged solution: two targeted and one general. To combat the periodic machine noise generated by the ventilators, oscillators, and pumps we developed an Active Noise Reduction device. In response to the people noise we devised an Active Feedback Sign which gives the staff immediate feedback about their volume. Lastly, to cover everything our targeted solutions missed, we recommended changing out the ceiling for better Acoustic Panels, which affect all frequencies and provide overall noise reduction in the room. Room #2605 Room Ambient: Oscillator/Ventilator: Diaphragm Valve: Suction Regulator A: Suction Regulator B: Room #2604 Room Ambient: Ventilator: Room #2606 Ventilator HVAC High (dBc) 67.1 81 87.4 67.1 74.8 63.9 81.2 74.3 68.7 Low (dBc) 62.4 78.8 84.1 62 64.8 57.4 63.4 72.4 64.6 Wave Theory Active Noise Reduction (ANR) Active Feedback Signage Sound is the vibration of the air, and when air vibrates it creates areas of high and low pressure. This pressure wave can be described by this equation: Active noise cancellation works on the principle of destructive interference. When two waves with the same phase but opposite amplitudes overlap, the peaks of one interfere with the troughs of the other, which leads to a reduction in total amplitude. We found that noise made by hospital staff was not continuous, but did spike up to unacceptable levels during shift changes and procedures that involved more than a single nurse. While the staff is trained to keep their voices down, this clearly is not effective when they are focused on other things. The goal of the sign is to alert the staff when their voices are getting too loud. When two or more waves coincide the interaction between them and resulting combined pressure wave can be described as: While this is helpful, for our purposes the most important equation is this one: This equation represents the resulting amplitude when two waves interfere with each other. Since we want to reduce noise we need to either absorb sound or generate a waveform which, when added to the existing noise, causes the left side of this equation to go toward zero. This concept is referred to as destructive interference. Source: ecalculator.com In our research we found that if you are able to keep the canceling source within ¼ of a wavelength of the noise source you can achieve global cancellation without all the complicated sensors, microphones and logic circuits that 3D cancellation normally requires. For our project the frequency we needed to cancel out was in the 1-1000 Hz range, and at 1000 Hz, ¼ of the wavelength is approximately 3.5 inches, so as long as unit remains close to the source it should affect the entire room. Our ANR circuit is comprised of 3 main parts: the input, inverter, and amplifier. The input is a unidirectional microphone connected to a unity gain operational amplifier acting as a pre-amp. This boosts the signal level of the microphone such that the rest of the circuit can use it (microphones usually have very low voltages). The second stage is an inverter, which in our circuit is again an operational amplifier though in this case it is utilizing its inverting input. This chip takes the incoming microphone signal and flips its phase 180°. This means that anywhere there used to be a peak there is now a trough, and vice versa. The last stage is a low noise audio amplifier with 20200dB gain. This section gives the signal enough power to operate a speaker. The specific components are noted in the circuit diagram below. ANR Circuit The concept is identical to the highway speed signs that tell you your current speed as you pass. While you know that the speed limit is 65, that does not mean you always travel at that speed. Sometimes you end up going faster without noticing, and the sign reminds you to slow down. Similarly, our signs are designed to alert the staff when their voices are nearing the 55 dB level. This prevents undue noise spikes, while our other two solutions actively reduce the ambient noise level. Active Signage Circuit Pre-amp Low Gain Amplifier Signal Inverter Display Driver Active Signage Model COMSOL Room Model ANR Device Model Speaker Audio Amplifier Analysis COMSOL Model Noise Reduction Calculation To determine the improvement that could be expected from swapping out the ceiling tiles, we performed a noise reduction calculation. The concept behind this is a comparison between the total absorptive area before and after the change. The relationship is below: NR = log Aa/Ab Material NRC Linoleum 0-0.05 Plywood 0.1-0.15 Drywall 0.05-0.2 Acoustic Tile 0.5-0.75 Ecophon The procedure is to calculate the total surface area in a given room (we used a typical ward with 6 beds). For each surface multiply the area by the noise reduction coefficient (NRC) of the material. The sum of these products represent the effective absorptive area Ab. The second step is to assume a change in the room composition. In our case this was a change in the NRC value of the ceiling tiles (0.5 to 0.95). Following the same procedure with the new ceiling generates Aa, from which we get the noise reduction in dB. Glass 0.95 .05-.10 Case NR (dB) Best 2.45 Average 1.45 Overall performance will be analyzed by a COMSOL model of a sample room. We will place equivalent point noise sources and accurate material properties into the model and run it using the acoustics module, which generates a representation of existing sound level pressures in the room. Our research indicated that while ANR on average reduces noise levels by around 10dB for low frequency noise (less than 1500 Hz). Using this 10dB as a starting point we will reduce the output of the point sources by that amount and change the NRC value of the ceiling. After running the analysis again this will give us a new graph which shows the approximate reduction in overall room noise as a result of our designs. Discussion Our aim was to reduce noise levels to 55dB. With the ANR averaging 10dB and the new ceiling providing approximately 2dB, the overall noise reduction is expected to be around 12dB. Given a starting level of 67dB this means we achieve our goal. That said, though we accomplished a great deal this semester, there is much that could be done to improve the design of both our circuits. The ANR circuitry works best against low frequency sound, but currently it is forced to process all noise. The addition of a low pass filter following the pre-amp would allow us to exclude frequencies outside our ideal range. This would decrease feedback potential, unintentional high frequency amplification, and could slightly decrease overall power consumption in the circuit. Similarly, even though the sign is only intended to target human noise, it currently registers all surrounding noise, which means alarms may cause it to generate false positives. An idea that we did not have time to implement was an adjustable band pass filter. It would allow the user to select a range of frequencies in which the alarms operate and exclude those noise from registering on the LED display. This is useful because if the sign is going into the red due to alarms the staff may begin to ignore it, which defeats its purpose. While COMSOL provides a decent representation of the acoustics of our room, there are more specialized pieces of software that would give a more accurate picture. Odeon is one of these, and though we wanted to use it for this project it did not fit in our budget this semester. Future work could include this analysis as well. Pre-amp Signal Inverter