Patent Application
20080085018
The invention provides a means to clean the intake air filter without removing it from the vehicle, thus reducing the amount of maintenance required and extending the service life of the filter and consequently reducing the operating cost of the vehicle.
The filter cleaning action is accomplished with the use of a sonic horn; a device that generates low frequency, high-energy sound that lifts the dust from the surface of the filter and through a vibrating motion transports it away from the filter. This invention provides a means to apply very small, low energy usage horns to mobile equipment.
Through sonic horn design and critical placement of the horn, or multiple horns on or in the air cleaner housing allows the use of small sized horns that can provide adequate energy to clean the filter surface with low operating energy requirements and without generating objectionable noise levels. This cleaning action can be assisted with the release of compressed air to the inside of the filter surface, which will blow the sonically released dust away from the filter. The use of the compressed air is not necessary to achieve adequate cleaning action but can be used to augment the effectiveness of the sonic horn.
The design of the sonic horns will allow for operating power to be supplied either as low voltage direct current from the vehicle's electrical system or a horn that operates on compressed air can be chosen. Many commercial, industrial and military vehicles have a source of compressed air on the vehicle.
At low frequency sonic horn is part of the air filter. The placement of the sonic horn is important. The sonic horn use low frequencies to vibrate the dust from the surface of the air filter. Gravity is used to remove the sonically released dust from the filter. Compressed air can be used to augment the effectiveness of the method of cleaning the air filter, by blowing away the sonically released dust.
In the preferred embodiment, the wheeled or tracked vehicle is stopped during the cleaning process. Alternatively, the sonic horn could be activated while the vehicle is moving.
In the preferred embodiment, the sonic horn is activated manually. In an alternative embodiment, the sonic horn is activated automatically, such as after a specified amount of time has passed.
In the preferred embodiment, a low energy usage sonic horn is used.
In order to use acoustic methods to loosen particulate from a filter, the acoustic system must be capable of producing extreme sound levels, inducing large acoustic velocities, on the surface of the filter. Given the size and cost constraints of a military vehicle filter, the very large, low frequency transducers used in industrial and agricultural filter systems cannot be used. To produce very high sound levels from a small transducer, higher frequencies must be used and natural acoustic enhancement of the sound field through acoustic resonance of the canister/filter needs to be utilized. To begin the design process a good knowledge of the resonance characteristics (i.e. the acoustic modes) of the canister/filter is required. In particular, knowledge of the resonance frequencies, pressure mode shapes, and the acoustic velocities are required.
When the resonance frequencies and mode shapes are known, the frequency of the transducers, required transducer strength, and the transducer locations can be determined and a more complete simulation and construction of a prototype can begin.
A simplified 3-D acoustic finite element model (FEM) of the HMMWV air cleaner has been developed. The model consists of a rigid canister with the air inlet extension as shown in FIGS. 2,3 below. Inside the canister, the space has been separated into three radial regions. The outer and inner are air and the center region is the air filter, modeled as a highly damped dispersive media. The outlet tube is modeled but that is not seen in the figures below. Also shown in the figure is the FEM mesh. This system was analyzed to determine the resonance frequencies and shapes of the acoustic modes. These acoustic modes are the natural resonance conditions for the canister/filter system. They are the frequencies where resonant enhancement of the transducer output will occur.
The computations were analyzed to determine the resonance frequency, the acoustic pressure variations and the acoustic velocity vectors of each mode. Of particular interest are modes that have acoustic pressure variations and acoustic velocities that would induce motion of the particulate on the filter surface in a manner that would dislodge them. Longitudinal modes with variations along the length of the canister and azimuthal modes with variations around the circumference of the canister will have acoustic velocities parallel to the surface of the filter. These may be the best modes for loosening particulate. Sloshing modes with side to side variations are another mode. Radial modes will have radial pressure variations and acoustic velocities. These will alternately pull the particulate off the filter surface and then push the particulate back on.
The plots that follow are an analysis of the first ten acoustic modes of the canister. They range in frequency from 670 Hz to 1344 Hz. The modes with frequencies above 1000 Hz are more interesting because the small, efficient, robust, and inexpensive transducers required to excite these modes are more readily available at the higher frequencies. The analysis shows a cluster of three modes around 1290 Hz, two of which are radial and one of which is longitudinal. A tone at 1292 Hz would probably excite all three modes well with a very complex resulting motion.
Transducer excitation can be varied to determine how much to excite each mode and if the resulting multimodal excitation would be an enhancement to particulate motion or a detriment.
The two modes at 670 Hz
The two modes at 945 Hz
The two modes at 1159 Hz
The mode at 1288 Hz
The mode at 1292 Hz
The mode at 1296 Hz
The mode at 1344 Hz
Using COMSOL (formerly called FEMLAB) software a simple circular duct was modeled. See
Next, a model of the HMMWV filter canister with a filter was developed independently. The independently developed model and the earlier model predict the same resonance frequencies.
There was a cluster of modes with resonance frequencies in the 1500 to 1600 Hz range. This is a frequency range that can be strongly driven with off-the-shelf sound reinforcement horn drivers for testing purposes. Additionally, the frequency is high enough that a high efficiency, compact piezoelectric driver could be designed to drive the resonance in a production product.
A model was developed which included a transducer. The transducer was modeled as a surface with a fixed acceleration. This is a very good approximation to an electrodynamic or piezoelectric loudspeaker. The transducer selected for the model produces about one watt of acoustic power. This is a large acoustic power output, but certainly an output achievable with off-the-shelf sound reinforcement drivers.
Sound pressure levels for excitation of several of the modes were found during natural frequency analysis. Mode shapes were mapped out for a particular choice of driver location.
The initial simulations were done with a random placement of the transducer. The first position was on the side of the canister about ⅗ of the length from one end. This t is very near to a node for strong axial (end-to-end) resonances. The maximum sound pressure level obtained was less than 130 dB.
The transducer was placed at one end of the canister so as to maximally excite the axial (along the length of the canister) modes.
Tests were conducted with the transducer positioned between the filter and the outside canister
Tests were conducted with the transducer positioned between the filter and the outside canister wall, with the transducer completely under the filter, and with the transducer half way under the filter and half way between the filter in the canister wall. It may be expected that the position between the canister wall and the filter would produce the highest output. The position between the filter and the canister was indeed the best.
With one acoustic watt from the driver, sound levels in excess of 150 dB were excited. At a frequency of about 1550 Hz, a very strong axial mode in conjunction with a strong circumferential mode was excited. There are strong pressure gradients along the length and around the side of the filter—this is the type of excitation that will best dislodge a particulate from the filter. A radial mode will have gradient away from the filter wall, but that is followed by a gradient toward the filter wall so it is not clear that a radial mode would help re-move particulate—it may just impact it further onto the filter.
It was noticed that an end position of the speaker could be a problem for testing as the outlet port might interfere with installation of a large horn driver. The position of the driver was subsequently moved back to the side of the canister, but very near the bottom. There was little change in the response. The new location of the transducer can be clearly seen in the lower right corner of
Simulations of the speaker being excited at 1552 Hz show the expected excitation of an axial mode as well as a circumferential mode.
Plots of the acoustic pressure gradient are shown in
The response output may be further improved by further testing other transducer positions. Additional transducers may be added in appropriate locations and with appropriate phasing to excite only preferred modes.
A canister is disclosed with a slide mount for the loudspeaker. Using a loudspeaker, an audio amplifier, and a laptop with a simple tone generator software as a source, it is possible to loudly excite the canister with pure tones. It is possible to excite a very strong mode near 1220 HZ BUT NOT NEAR 1550 Hz. Dust is released when playing a loud tone in the 122-Hz range.
Another cluster of modes existed just above 1200 Hz. It may be noted that several of these were radial modes.
The modes near 1200 Hz were excited quite a bit more than those around 1550 Hz. Another set of strong modes were found near 1700 Hz. It was noted that the speaker source was actually located nearly 1.5 inches from the end of the canister rather than at the very bottom as was modeled in COMSOL. Dust is released at 1220 Hz but not at 1550 Hz. A frequency chirp from 1200-2000 Hz did an even better job of releasing dirt than a single 1220 Hz tone.
The model was updated with the speaker position moved to match the experimental setup. It was found that the new location, the 1200 Hz cluster of modes was indeed excited more than the 1550 Hz cluster. Excitation of longitudinal modes was highly sensitive to the position of the transducer-far more than radial or axial modes. A canister may be provided with the speaker mounted as close to one end as possible.
Actual particle motion may be modeled using COMSOL. Particle motion appears to be modeled only with true transient simulation rather than steady state simulation. It was deemed impractical to try to model a steady state mid frequency acoustic system using a transient simulation.
A new canister was developed that removed the inner tube as well as a new filter with a more shallow pleat and a material that should release particulate easier. The new canister also has the transducer mounted very close to the end of the can.
The COMSOL model was updated to match the current canister configuration and is shown in
The major resonance frequencies match between the real canister measurements and the model predictions.
One embodiment more accurately predicts the actual resonance frequencies better than the previous model. A new speaker with higher output capability was used in the experiments. A number of frequency sweep ranges were tested and it was found that sweeps from 1200-2000 Hz and 1400-2400 Hz both were able to loosen significant amounts of dust.
The radial modes are actually very useful. It is now thought that the frequency sweep works so well because it actually excites the particles in a way such that a radial mode will pull the particulate off of the filter and then the longitudinal and axial modes will pull the airborne dust away from the filter.
High output piezo buzzers and sirens are disclosed. Piezo alarms with two tones or with a chirp are fairly inexpensive and robust. They appear to be ideal candidates for a prototype transducer.
The COMSOL model for a system with two transducers show some interesting results.
An example of the excited mode shapes is shown in
Multitone and chirp excitation signals are disclosed, as well as an actual measurement of the amount of dust removed from the filter. Sweeps in the 1800 Hz to 2400 Hz range are disclosed.
A second transducer may be mounted to the canister. A production piezo buzzer or siren is disclosed for installation in/on a canister.
