طراحي، ساخت و آزمون سامانه پايش وضعيت جعبه دنده با كمك پردازش سيگنال هاي صوتي

Design, development and test of the gearbox condition monitoring system using sound signal processing

یكی از روش هایی كه امروزه برای به حداقل رساندن هزینه نگهداری و تعمیرات تجهیزات صنعتی دوار به‌كار می رود، پایش وضعیت به كمك تحلیل صدا می باشد. در پژوهش حاضر به منظور بررسی روند تشخیص عیب و پایش وضعیت جعبه دنده ابتدا سامانه ای طراحی و سپس ساخته شد. صدای حاصل از چرخ دنده های سالم و معیوب به‌وسیله صداسنج اخذ و در رایانه جهت پردازش ذخیره گردید. اندازه گیری صدا در سه سرعت پینیون 750، 1050 و 1500 دور بر دقیقه برای چرخ دنده های سالم، عیب شكستگی یك دندانه و ساییدگی یك دندانه انجام شد. مقایسه سیگنال های صوتی پردازش شده از چرخ دنده های جعبه دنده در شرایط سالم و معیوب نشان دهنده فركانس درگیری چرخ دنده، هارمونیك های آن و نیز تغییرات در اثر عیوب ایجاد شده بود. نتایج حاصل از تحلیل طیف فركانسی سیگنال ها در سرعت 1500 دور بر دقیقه پینیون، عیب شكستگی یك دندانه در فركانس درگیری چرخ‌دنده ها را به ترتیب برابر 500، 350 و 250 هرتز نشان داد كه با افزایش دامنه فركانس مذكور ظاهر شد. عیب سائیدگی دندانه ها در یك چرخ دنده به‌صورت ایجاد باندهای جانبی به فاصله مساوی در اطراف فركانس درگیری چرخدنده در طیف فركانسی سیگنال در سرعت های 1500 و 1050 دور بر دقیقه پینیون كاملاً قابل تشخیص بود. بررسی طیف فركانسی سیگنال صوتی حاصل از چرخ دنده، بیانگر توانمندی این روش در پایش وضعیت جعبه دنده با دقت بالا در زمان حداقل می باشد.

<strong >Introduction </strong > One of the ways used for minimizing the cost of maintenance and repairs of rotating industrial equipment is condition monitoring using acoustic analysis. One of the most important problems which always have been under consideration in industrial equipment application is confidence possibility. Each dynamic, electrical, hydraulic or thermal system has certain characteristics which show the normal condition of the machine during function. Any changes of the characteristics can be a signal of a problem in the machine. The aim of condition monitoring is system condition determination using measurements of the signals of characteristics and using this information for system impairment prognostication. There are a lot of ways for condition monitoring of different systems, but sound analysis is accepted and used extensively as a method for condition investigation of rotating machines. The aim of this research is the design and construction of considered gearbox and using of obtaining data in frequency and time spectrum in order to analyze the sound and diagnosis. <strong >Materials and Methods </strong > This research was conducted at the department of mechanical biosystem workshop at Aboureihan College at Tehran University in February 15th.2015. In this research, in order to investigate the trend of diagnosis and gearbox condition, a system was designed and then constructed. The sound of correct and damaged gearbox was investigated by audiometer and stored in computer for data analysis. Sound measurement was done in three pinions speed of 749, 1050 and 1496 rpm and for correct gearboxes, damage of the fracture of a tooth and a tooth wear. Gearbox design and construction: In order to conduct the research, a gearbox with simple gearwheels was designed according to current needs. Then mentioned gearbox and its accessories were modeled in CATIA V5-R20 software and then the system was constructed. Gearbox is a machine that is used for mechanical power transition from a productive source of power to a consumer, for torque meeting and for rotating speed needed for the consumer. In fact, gearbox is an interfere between power source and power consumer which produces a flexible communication between power source and power consumer. Needing to a gearbox as a machine which can generate harmony as an interface is unavoidable due to lack of harmony of torque and rotating speed of production source of power. So necessary calculations in order to attain to technical characteristics of gearwheels, bearings, shaft dimensions and other accessories of gearbox were done. This gearbox is from kinds of simple gearwheel which its input and output shaft are parallel to each other. Main accessories of gearbox are: 1.crust 2.shaft 3.gearwheel 4.thorn 5.bearing 6.cover. All of the design parameters were calculated and considered in designing of all of the accessories of gearbox. Electromotor rotating calibration: For this aim, a light-contact telemeter in model of Lutron was used as contact. Acoustic module of electro motor: A module was constructed in order to prevent from sound waves interaction resulting from an electromotor function with waves of gearbox function. Three layers of sound absorbent including common felt with 1mm width, polyethylene foam with 15 mm width and shoulder foam egg with 35 mm width were used for the module insulation. Material used for the body of this module was MDF. Based on field measurement, level of electromotor sound decrement using the acoustic module was 20dB. Investigated malfunctions in this research are relevant to gearwheel with one tooth fracture, one worn tooth and one tooth fracture and other worn tooth. Collection and storage of acoustic data: In this research, an audiometer in model of HT-157 made in Italy in order to obtain acoustic data and a laptop with a model of Lenovo-G550 for data storage and processing was used. Cool Edit Pro 2.0 software was used for data processing. Data storage was in PCM format and MATLAB R2014a software used for data processing. Data processing: Signal processing method in the frequency domain is used in order to reveal the defects. Fast Fourier Transform: Fast Fourier Transform FFT for application in electronic equipment specially analyzers have great importance. In this condition, sampling number is chosen exponentially as 2N which decreases the calculation volume significantly. Determination of defect kind of gearwheel using frequency spectrum analysis: In mentioned gearwheel, errors were generated synthetically. Defect kind of these errors was generated in separate gearwheels in order to investigate the defects more precisely and a gearwheel was considered as control gearwheel. Despite of this, the sound of all of the gearwheels in correct condition was stored. <strong >Results and Discussion </strong > Comparison of processed acoustic signals from gearwheels of gearbox in two correct and incorrect conditions was indicative of gearwheel involvement, frequency, their harmony and the changes resulted from defects. Gearwheel defect detection tests showed that at the speeds of 1496, 1050 and 749 rpm, investigated defects are recognizable with a comparison of the frequency spectrum of obtained signals in correct and incorrect conditions and according to the involvement frequency of gearwheel, its harmony and sided spectrum. Results of the frequency spectrum of signal analysis in speed of 1496 rpm pinion showed the defect of one tooth fracture in involvement frequency of gearwheels by 489, 350 and 249 Hz respectively which became apparent with a mentioned frequency domain increment. A worn tooth defect in a gearwheel was completely determinable as sided bands with equal distance around gearwheel involvement frequency in the signal frequency determination of the speeds of 1496 and 105 rpm pinion, but became a bit harder in less speeds. Investigation of frequency spectrum of acoustic signal resulted from gearwheel, is indicative of the ability of this method in gearbox condition investigation with high precision and minimum time. So the gearbox condition investigation is reached by investigation of the frequency spectrum of acoustic signal resulted from gearwheel. <strong >Conclusion </strong > In current research, acquisitive signals resulted from produced sound waves of constructed gearwheel were used for investigation and diagnosis. Recorded signal in time domain and processed frequency and exploited characteristics of signal in frequency domain for diagnosis were analyzed. Obtained results of this research can be summarized as follow: 1. Precision level in the diagnosis decreased by increasing in pinion speed. 2. There will be a decrement in gearwheel diagnosis after defects integration and signal behavior won’t be completely similar to the defect as individual. 3. Proper placement of audiometer is effective in diagnosis trend. 4. In frequency spectrum of obtained signals, particle velocity level is more efficient in diagnosis than the sound pressure level.