Student Projects
Contact Information
Department: Industrial Engineering
Institute: Jazan University, Jazan
Team Members: Ihab M El-Sayed
Project Information
Abstract:
The maintenance of the machines running at high speed in an industrial environment is of great important and need sufficient care for its proper and satisfactory working. In the present paper an idea for on-line condition monitoring for machines are suggested. In this paper piezoelectric sensors will be used to gather the Data ( vibration signal) in a textile spinning machine to monitor it on-line and analyze the vibration in the real time in order to send an email to the maintenance department or an SMS to a group of stakeholders or relevant people in the maintenance department as an alert or reporting a failure in the system when the vibration reaches a preset or a predefined certain level.
A Labview 8.6 Package with Signal Express 3.0 from National Instruments will be used in this application. Also ComapctRio as a Hardware for real time controller will be utilized in order to interface the textile machine with the accelerometer (piezoelectric sensor) and the internet or the GSM network. It is a very good e-service application for the Textile industry field which could be applied in many other industrial fields like Petroleum , food, Plastic , Paper and any other field the stopping time of which will cause a high cost or losses in the resources. The proposed system will lead to a revolutionary enhancement in e-maintenance and many other e-service applications.
1- INTRODUCTION
Linking Industrial Engineering Domain with emphasis on Small and Medium enterprises SMEs and eServices is now growing up in the industrial field. One of the main issues in the SMEs is the maintenance of the machines. The production line in any factory consists from one or multiple machines. These machines need to be well maintained in order to keep the production to the required level of quantity, quality and safely. Most of the machines are expressed to different level of vibration. The vibration level depends on many factors and affect many parameters in the machine parts including stiffness. For these reasons, Vibration analysis and its application in maintenance and reliability of systems become an important issue.
The maintenance problem is directly related to the production and productivity. An established system for maintenance can improve the productivity and increase the production rate. SMEs that has exporting activity need to apply modern Industrial Management system like Lean Manufacturing and Six Sigma. Total Quality Management (TQM) consider the main maintenance issue. The well established enterprises need to have a Quick Response (QR) system of maintenance. In which, the machine stopping time must kept as minimum as possible. This could be achieved by installing a good maintenance system capable of quick response to any failure to the system or machine, or a preventive maintenance system must be installed in the factory. In other words the machine must be repaired Just In Time (JIT) when the failure has occurred. This means the shutdown time will be minimum, the productivity will be maximum and profit will be maximum as well.
In order to achieve all these requirements in the industry, innovative and creative solutions must be established. By utilizing the revolution in communication technology including the increasing number of users who have access to the world wide web (www) through Ethernet connections. Such creative ideas and solution might be achieved. Also, the increasing number of mobile user made the Global System Mobile (GSM) and its application are growing fast as well. This also could be utilized in contributing in the creation of innovative maintenance ideas.
This paper discusses the innovative way of joining the above three main domains together (Industrial Engineering (IE)+ SMEs+GSM and Ethernet) in the beneficiary of the Textile industry as a special but could be applied in all other similar situations fields of industry.
There are many SMEs in Egypt many of them working in the textile industry. They are located in certain number of area all over the country but concentrated in two famous region the first one in Gharbia Governorate ( El- Mahala El-Kobra) and the second in Cairo Governorate (Shoubra El-Kheimah).Textile Weaving factories in Greater Cairo (Shoubra El-Kheimah) and in the main city of Textiles ( El-Mahala El-Kobra) the majority of them are considered as a Small or Medium Enterprises (SMEs).
Most of them do not have a highly qualified maintenance engineers. Many of them have a limited number of machines, whereby the breakdown of one of them might affect the whole production. Most of the machines are sophisticated and need extensive training courses for maintenance. The machines are made mostly in western countries , spare parts are costly. Spare parts are not on the shelves. Shipping and manufacturing might take up to 6 weeks for delivery for some rare and also major parts. In other words Inventory is not proper. Economy of developed countries with high number of SMEs will need such projects.
The idea is using a sophisticated system for the analysis of the vibration and the estimation of non linear parameter could be utilized in the early detection of faulty parts and failure of mechanical systems and machines. By installing a piezoelectric sensor to monitor the vibration on-line and record this vibration signals on a computer for further analysis will assure early detection and avoidance of system failure. In addition to that it will be very valuable if e-service were utilized efficiently to have a very quick response from the relevant department. The system will serve as a remote sensing for early detecting of failure and faulty parts to be replaced. The proposed system will help in reducing the shutdown time and of course reducing cost and increasing profit.
2 - Material and Method
A piezoelectric sensor (Accelerometer) were installed on the textile machine to gather the vibration signals of the entire system due to the running with high speed. In Previous work for the authors, they concluded the Volterra kernels extraction was proposed for the estimation of polynomial parameters, which can also be used for parameter estimation of ball bearing if the cubic term was considered instead of the quadratic non- linear terms in frequency domain. In the previous paper simulation and modeling part was carried out . Here, in this paper we are extending the work beyond the parameters estimation to condition monitoring of the machine part or the machine as whole. The system is very economic when integrated to the current and existing systems, the sensors will not be costly.
Figure 1 shows a schematic line diagram for the vibration Monitoring System. The author has already carried out the experiment on a textile machine by introducing faults to the system and without faults as well. Figure 2 is a real photograph for the machine used in the experiment.
In the schematic line diagram no. 1, 2 are representing the two sensors ( accelerometers) to measure the vibration signals in different positions. Figure 3, 4 are the photograph of the accelerometer.
The two sensors were connected to an analogue / digital convertor card (A/D) Card, which is shown in figure 5. Figure 6 and Figure 7 shows the way of fixing the sensor to the machine.
Fig 2.
Fig 3.
Fig. 4
Fig 5. PCI 4461 from National Instruments
Fig 6.
Fig 7.
2-1-Hardware
Wired and Wireless technologies has been widely used for measurement and automation, National Instruments and LabView has introduced it over the past several years. Wi-Fi and 802.11 b/g: The IEEE 802.11 standard encompasses a series of specifications for wireless LAN technology. Bluetooth (802.1a): A consortium of companies composed of Ericsson, IBM, Intel, Nokia, and Toshiba defined this standard. The main intent of this wireless communication standard was to make it easier for devices to communicate over short distances – typically less than 10 m. GPRS, GSM: The General Packet Radio Service (GPRS) is a nonvoice service intended for information to be sent and received across a mobile telephone network.
With GPRS, data can you can send or receive data immediately as it is produced, as long as the radio signal is available. Unlike traditional land lines, this system does not require establishing a connection – it is always connected. This is an advantage for applications where time and quick reaction to events is crucial. GPRS overlays a packet-based air interface on the existing circuit switched Global System for Mobile Communications (GSM) network. Wireless modems and proprietary networks:
There are many vendors that offer industrial-grade modems specifically designed for rugged environments with extreme temperature ranges and high-shock and -vibration conditions. Ethernet-Based Distributed I/O Systems with National Instruments LabVIEW The growth of Ethernet as an industry standard has greatly improved the capability and integration of networked measurements, remote monitoring, and distributed control applications. In both manufacturing and test, engineers continue to find new ways to take advantage of this PC-based technology to build individual test stations or machines. These measurement and control systems use Ethernet to connect to management, data, and quality systems. As companies start to integrate disparate hardware with the common Ethernet link, they find that having flexible, open software becomes key.
Fig 8
Fig 9.
Fig 10
Fig 11.
Fig 12
3- COMPUTER SIMULATION
For the purpose of the condition monitoring we have considered Single Degree of Freedom (SDOF) System with quadratic nonlinearity. The excitation to the system is considered to be random in order to include all frequency ranges. The detailed mathematical analyses can be seen in Appendix. The procedure is illustrated through numerical simulation of the response for the non-dimensional equation (4). The forcing function in the equation is a normalized zero mean random force, f( ). The excitation force is simulated through random number generating subroutines and is normalized with respect to their maximum values. The typical sample of the input force is shown in Figure-13.
Owing to the statistical nature of the problem, the procedure is illustrated for various sets of nondimensional nonlinear stiffness parameters and linear damping terms. The following case studies have been designed to study the influence of parameters and errors involved. The response is computationally simulated, for various sets of numerical values of the parameters in equation (4). The governing equation is then numerically solved using a fourth order Runge-Kutta method, to obtain the response. The response is fed as input to the parameter estimation algorithm. The frequency response functions (FRFs) are extracted from the response and consequently parameter estimation is carried out. The estimated parameters are compared with those originally used for the simulation of the response and accuracy of the estimates and errors involved are studied.
Owing to the statistical nature of the problem, the procedure is illustrated for various sets stiffness and damping parameters. The following case studies (Table-1) has been designed to study the influence of various combinations of direct and coupled parameters in response simulation and their utilization in parameter identification using the algorithm developed.
4- RESULTS AND DISCUSSION
First of all the power spectrum of the random force averaged over the ensemble of 2000 samples of Figure-13 is obtained. The non-dimensional response has been numerically generated in the non-dimensional time range 0-2048. The ensemble is constructed from 2000 different samples of the simulated random force. The response spectra obtained are fed as input to the parameter estimation algorithm. The first order Volterra kernel i.e. Figure-14 is computed using equation mentioned above. The kernel exhibits peak response corresponding to its critical frequency. The linear parameters are obtained using curve fit technique from( ) 1 H . The linear parameters thus estimated are n = 0.1592 (cycles/ ) and 0.0107. The exact values of the above coefficients are those used as input for numerical simulation of the response. While, the stiffness term can to be estimated with a very good degree of accuracy, the error in the damping estimates is higher. The second order estimates are shown in Figure 14 and 15. The nonlinear estimate have been carried out in the frequency range as shown in Figure -16.
These linear and nonlinear estimates are compared with the exact value used for simulation and the error is obtained. This error can be monitored online for safe and smooth running condition of the machine considered. In other words we can say that we can monitor the machine on-line just with the information of the signals picked up from the sensors.
Fig. 13 Typical input sample
Fig. 14 Estimate of First order kernel transform
Fig. 15 Measured second order kernel transforms
Fig. 16 Second order kernel Transform
Fig. 17 Estimate of quadratic non-linear parameter
5- CONCLUSION
The eService techniques for the condition monitoring of a machine running at high speed is proposed. The linear and nonlinear parameters were obtained by the algorithm developed and the same is being watched for their smooth running. The Volterra kernels extraction is used for the estimation of polynomial type of stiffness parameters. The analysis presented is in nondimensional form and can be suitably used to design experiments. Curve fit technique is employed in order to estimate the linear parameters.
The effects of various types of disturbances have been analyzed using simulated data. It has been found that large sample size is preferable for obtaining more accurate results. The present work to systems with more than one degree of freedom can be explored.
ACKNOWLEDGMENT
The Authors wish to express their sincere thanks to whoever participated in this project and especially the Textile Institute in Aachen – Aachen University – Germany.
REFERENCES
1. Volterra, V.,“Theory of Functionals and of Integral and Integro-Differential Equations”, Dover
Publications, Inc., New York, 1959,.
2. Wiener, N., 1958, “Non-linear Problems in Random Theory”, The Massachusetts Institute of
Technology and John Wiley and Sons, New York.
3. Bendat, J. S., “Nonlinear System Analysis and Identification from Random Data”, New York: John
Wiley, 1990.
4. Schetzen, M., “Measurement of the Kernels of a Nonlinear System of Finite Order,” International
Journal of Control, 1(3), pp. 251-263, 1965.
5. French, A.S., and Butz, E.Z., 1973, “Measuring the Wiener Kernels of Non-linear System using the
Fast Fourier Transform Algorithm”, International Journal of Control, 17, No.3, pp.529-539.
6. Odiari, E.A., and Ewins, D.J., 1992, “Parameter Identification for Non-linear Rotor-Stator
Systems- the Volterra/Wiener based Approach”, Institution of Mechanical Engineers, Vibration in
Rotating Machinary Conference Proceedings, pp.-193-202.
7. Stephen Boyd, S.; Tang, Y.S., et al, 1983, “Measuring Volterra Kernels”, IEEE Transaction on
Circuits and Systems, CAS-30, pp.571-577
8. Ewen, J.E., and Wiener, D.D., 1980, “Identification of Weakly Non-linear Systems Using Input and
Output Measurements”, IEEE transaction on Circuits and Systems, CAS-27, No.12, pp.1255-1261.
9. Schetzen, M., 1980, “The Volterra and Wiener Theories of Nonlinear Systems”, John Wiley, New
York.
10. E.C.Levi, “Complex curve fitting”, IRE Transaction on Automatic Control AC-4, pp. 37-44, 1959.
11. Khan A. A. and Khan A. H., 2003, “Estimation of Linearised Fluid Film Coefficients In A Rotor
Bearing System Subjected To Random Excitation”, Proceedings of NaCOMM-2003, pp. 380-386.
12. National Instruments, CompactRio Manual, www.ni.com,
13. RWTH, ITA Aachen, Germany 2008 Research Report
