Skip to main content

Reduction of noise emission on machine tools

Risultati finali

The stand-alone active damping device (ADD) has been designed to introduce significant damping in all observable structural modes from 20Hz to 2kHz. The device consists of: - A sealed electro-mechanical actuator to be attached to the structure to be damped. The actuator also includes a vibration sensor and its conditioning electronics. - A remote electronic box, including the actuator power electronics and the controller, as well as performance monitoring interfaces. In-situ integration is simple; up to 15% damping has been achieved on industrial production machines on first attempt without preliminary knowledge of the machine dynamics.
A new optimisation algorithm to locate actuators and sensors on the machine frame, based on physical criteria, has been developed. These criteria can be: ability to excite frame modes by the actuator, observability of the frame modes by the sensor, stability of the control transfer function, etc¿ This method has been demonstrated to be more efficient and produce better results than for example genetic algorithms.
Based on previous experiments developed un simple systems a structure borne noise active control has been implemented. The main principles of the control system are: - Piezo actuators parallel to the main plane of the structure in order to introduce moments. - Acceleration Feedback Control. - Active modal damping with second order filters. The system is potentially applicable in large, rigid system in with a great percentage of free contours.
At the beginning of the project, one press was selected as representative of the presses made by Fagor. Actually, the selected machine was ¿FAGOR¿ MODEL SE4-1250-5000-2300 single action for cutting and drawing sheet. The machine is available at Fagor workshop, because it is a testing press used to test and improve process and machine parameters, usually for specific customers¿ demands. This machine was also selected because of its actual location and its availability along the project. This way, it could be possible to check the improvements in a particular machine before and after the project. Once the machine was selected and the acoustic and dynamic behaviour were measured in first steps of the project, some conclusions were drawn. Based on these conclusions, the optimisation strategy was defined. And along the project, the idea is to refit this press machine according to the defined optimisation options, applying active and passive control systems for airborne noise and for structure borne noise. This way, the result will be a noiseless press. But, more important than the modified press, Fagor will be able to design new presses with the new devices for active noise control, as well as with the best damping and absorbent materials for passive noise control. In fact, along the project noise sources and their transmission paths will be analysed and means for noise reduction will be studied and this knowledge will be used by Fagor in new press designs.
In the other hand, the piezo technology is used to design the piezo actuators to integrate in the structure. This specific design allows studying the behaviour of the piezo actuator on large structure to reduce the noise. To complete the study, Cedrat Technologies develops standard controllers to reduce the noise on a small demonstrator base on a simple beam. The results of this study is a know how on the control of vibration using the piezo technology like the behaviour of the performance with the different parameters from the piezo actuator.
A new high power linear amplifier has been developed for piezoelectric dampers. The new technology used in the prototype permits to multiply by 10 the available electrical power of standard power supplies available on the market and dedicated to piezoelectric actuators and dampers. The direct consequence is a larger bandwidth for the active vibration and noise control of structures. The second one is the possibility to supply larger piezoelectric actuators or several actuators in parallel. Know-how and the technology are now available at Cedrat Technologies for the industrialisation of a new product, which will permit to access to a new market. It will also directly increase the sell of standard piezoelectric actuators.
To generate large force with large stroke, a largest Piezo actuato was designed: The APA1000XL includes the patented mechanical shell to magnify the stroke. The principal characteristics are a stroke of 1000µm and a force up to 800Newtons. This piezo actuator can be used in large structure from the positioning to dynamic applications like the control of vibration.
The market on passive damping material is very diverse. That means there exist a huge number of providers for such damping materials with a lot of different product each. Within the NOISELESS project, different passive damping materials have been analysed with respect to their dynamic and acoustic effectiveness. Attention was laid onto the fact, to test a wide range of different product classes. Those passive damping materials have been tested on a 1:1-scale laboratory demonstrator under constant conditions in an anechoic room. Like this it was possible to determine the best passive damping material for different application purposes. For instance in the case of the press, different damping material was used for the front cover panels and the panels of the columns.
The new prototype of punching machine developed within Noiseless will be less noise producing because of some modifications of the design and because of some extra equipment that will be integrated in the machine. It is now hard to predict the acoustic level that will be reached with such prototype. Acoustic simulation is not performant enough to allow predicting the improvement in the sound level that could be achieved with all the modifications included. The limitations concerning power and natural frequency of inertial actuator has limited the performance of the active systems implemented in the prototype. Nevertheless, some characteristic details of the future punching machine can be advanced at that point: - It will equip linear motors for both axis X and Y. This will suppress the noise produced by the transmission, nowadays composed by rotary motor, belt, bearings and ballscrew. Even if this noise is not so harmful as the noise produced by the punching unit, it can be quite disturbing when the process includes fast movements. - The current linear guides will be substituted by similar linear guides, but including caged balls, that are supposed to reduce the noise emission. - It will equip active dampers to reduce the structural noise. - It will use re-designed punches and stations that reduce noise emission. - It will be able to work at different speeds of punching unit, and the speed will be controlled during the motion of the actuator. This will allow producing non-constant ramps that will reduce noise by saving time. Apart of economical impact in the partners involved in the development of such machines, the potential impact is very important in terms of physic and psychic health of the workers. The current state of development of the improvements described is the one carried out in the prototype of the machine, that is: - The machine has an axis driven with a linear motor and guided by re-circulating caged balls. The measurement to be carried out will show the performance of such systems. - The active mass damper of 900N has been installed and set-up. The presence of its first mode of vibration at 900Hz, that is, within the frequency range to be damped, has limited the effect of such strategy of the combination of the damper and the iterative learning controller (ILC). - The machine works with newly developed tools and tools holders. - The machine already can work at different constant speeds. It is not able to vary the speed during the stroke of the actuator. This means that the noise is reduced, as well as productivity.
The main objective of the project from Fagor point of view is to improve their machines in the future, reducing their noise level. This way, all the information, all the analysis along the project will allow designing new machines with a reduced noise level. Actually, as a result of the project, Fagor will be able to identify noise sources and their transmission paths on their presses. As well as this, Fagor will know means for noise reduction. This information could be used for new developments. As well as this, the information could be organised as design guidelines for silent machines.
Simulation of the vibroacoustic behaviour of the scaled demonstrator including a structure borne noise control system. The noise control system considers control filters, sensors and actuators.
The principle of the active mass damper is now well kwon. It has been successfully applied in the NOISELESS project to the structural noise control. Usual active mass dampers are electromagnetic ones, based on the moving coil (voice-coil) principle. The main advantage is the very low resonance frequency of such device. Their problem is that they produce quite low force. The use of the piezoelectric actuators permits to generate higher force, but standard direct piezoelectric actuators lead to too stiff dampers with a too high resonance frequency. They cannot damp low frequency vibration and noise. The use of Amplified Piezoelectric actuators developed by Cedrat Technologies permits to have quite low resonant frequency proof-mass actuator with enough force to be useful in structural control. But this active damper has a very high quality factor that impose to work at frequencies at least 2 times higher than the resonance frequency of the actuator in order to keep the control loop stable. The performances of the developed piezo active damper have been considerably improved by the addition of a passive damper in parallel with the piezo element. Then the quality factor of the active dampers is lower ant it can be used near its resonance frequency.
Design guidelines will be extracted from this research. From the measurement sessions performed in task 2 and the investigation of the small-scale the punch machine, construction guidelines how to build machines with well behaved dynamics will result. Also how to develop active systems (such as actuators and controllers) specifically for large and heavy machines subject to extreme high dynamic excitation will result.
In order to identify the correlation between noise and wear in punching tools an extended experimentation has been carried out. The wear has been analysed with optical and confocal techniques and comparing the tool profile with different level of wear. The method can be applied by tool manufacturers o analyse the correlation between noise and wear.
With the help of the dynamic and acoustic measurements a complete overview of the aspects in huge machine tools, especially at presses could be achieved. Not only vibration transmission within the structure of the press was investigated, but also the effect by changing specific properties of the machine tool, like rpm of the drive train, counter balance system etc. In addition a Modal Analysis was performed to identify the modal vibrational behaviour of the press as a whole. Therefore, not only local effects have been analysed but the complete interaction of the press parts. These results performed with the help of several measurements campaigns at the FAGOR workshop, give also a basis for the complete-Finite-Element Model, which WZL performed from 2-D-Drawings, which FAGOR provided WZL. By the help of this FE-model some effects could be investigated and it was the first step in building a complete Multi-Body-Simulation of the press structure and the press drive train.
A new innovative control algorithm has been developed wherein non-causal filters have been introduced, capable to control large complex structures subjected to repetitive excitation events. The control algorithm is an iterative learning algorithm suitable to control repetitive events. It uses information from previous events to enhance its performance when controlling new upcoming events. The innovation, which has been added by our research, is the introduction of non-causal learning filters in the controller. These are filters which work with future information and have attractive properties to stabilise control transfer functions. In a conventional feedback controller, information from the future is not available and these filters cannot be used. The iterative learning controller is able to reconstruct ¿future¿ information from the past events. The controller needs a trigger signal, which announces the next event. The non-causal filters cause the controller to respond to the event in advance. The response and the event need to be synchronised.
Our evaluations are based on the results of the tests carried out by Mondragon University during this year. The tests have been made on three types of tools with the same dimensions ( 40x5 square tools, working with the appropriate 40x5 + 0,9-die and a 4-mm-sheet). The only difference was represented by the type of sharpening: the first tool (tool H1) has a sharpening with only one inclination degree and direction; the secon tool (H3) is provided with a sharpening which creates a convex angle on the cutting surface, whereas in the third tool (H2) this angle is concave. In our opinion, 1000-strokes-sessions are not sufficient to evaluate the increase of noise caused by the tool wear. The tests show which tool is the noisiest (H1) and they also confirm that the punch with angles of inclination on the cutting surface is less resistant to the impact. The phonometric results show that H1 has a gap of 4,4 decibel from H2 and of 5 decibel from H3, therefore is quite far from 80 decibel. The other two solutions are far better, infact their results are much lower than the other one. In our previous presentation we carried out a test on a tool with a double concave sharpening (therefore with more angles of inclination on the cutting surface); the results were extremely satisfying. In this case, the cutting session was quite "noise-free" and it didn't cause any relevant deformation of the metal sheet. From the results we've noticed that the noise decreases of 0,8 decibel in H1 and H2 and of 0,2 decibel in H3. That can be explained by the fact that the cutting session occurs sharply and the sharpened tools produces a "dry" sound. On the other side when the tool is no longer fully sharpened the metal sheet is more subject to be deformed. As for the tool wear, we believe that 1000 strokes is a too short session as the cutting surface of the tool does not undergo any relevant wear.
The research carried out in tasks 2.5 and 3.2 allowed identifying relations between different process parameters, mainly tool sharpening and punching speed, and acoustic noise emission. This know-how is very useful in order to be implemented in future machines.
A small-scale punching machine has been developed and realised. The vibration and noise radiation properties of the original machine have been measured during task 2. Based on these results, the small-scale machine has been developed such that the vibration and noise radiation properties of the original machine are also present. The machine was developed using numerical simulations and experimentally verified. The resulting small-scale machine permits to develop new noise reduction technologies in an efficient and economic way without interfering with other activities planned on the original machine. These new techniques can be developed using the numerical models as well experimentally on the small-scale machine. Once these new technologies are developed and tested, they can be transferred to the original machine.