Deep Kong Zuan cuts condition to monitor with the analysis

  • Time:
  • Click:111
  • source:BAGANZ CNC Machining
Kong Zuan is cut is to be in close undertake below condition, cannot observe directly the cutting circumstance of cutting tool. Current, can by experience, in order to hear voice, look it is normal that outside cutting bits, observation to pressure is expressed and feel drill pipe vibration to wait, view phenomenon comes to judgement getting to cut a course. Deep aperture broach is divided outside wearing away normally, compare as a result of the long way of place bore big, machining complex tigidity is poor, broach easy generation is oscillatory, cut bits to be caused easily jam, consequently constant regular meeting makes deep Kong Zuan produces random damaged. Because deep Kong Zuan cuts a course very flabby calm, need a kind of deep Kong Zuan to cut the system that monitor very, real time is monitored auger the condition in cutting a course and concerned information, undertake signal is handled in real time, identify the unusual position in giving treatment course, adopt a change in time to get cut dosage or change the step such as cutting tool will eliminate unusual position, give when necessary show call the police, in order to avoid incident of hair make trouble, cause huge pecuniary loss, can raise cutting tool durability greatly thereby, make sure deep aperture gets cut on the rails. Because operating mode of deep aperture treatment is mixed, the physical origin that changes to causing these condition still lacks thorough understanding, still measure deep Kong Zuan to cut the method of the process and system without more perfect inspect at present. Traditional only factor monitoring and only factor model analyse the limitation with methodological very large presence, the impact of random element is too big, reflect true condition of the system correctly very hard. The article basically studies deep Kong Zuan cuts a process to monitor plan and signal analysis. The deep Kong Zuan that designed a kind of many sensor is cut force and auger cut oscillatory signal to gather device, a variety of signal that in using technology of much information confluence to cut a course to deep Kong Zuan, collect undertake analysis and processing, reach accurate judgement example, offer criterion with control to get those who cut a course to monitor. The plan that monitor and device cutting contained a large number of information in cutting process in force, almost when all cutting breakdown happen, cutting force signal all has corresponding change. But, the diagnostic signal that place of different cutting treatment means shows is endless also and identical, rely on this kind of onefold criterion to be judged merely cutting tool wears away or damaged, it is certain to can be caused sign up for by accident. Accordingly, the much sensor that the characteristic that considers to cut a system according to deep Kong Zuan originally designed a kind to be used at deep Kong Zuan to cut is measured force and measure brace up device, have strong, trends answers versatility to be mixed quickly installation is used convenient wait for an advantage. Graph 1 much sensor measures force and measure brace up much sensor of device sketch map measures force and measure brace up if the principle signal of device pursues show 1 times: Pass mix of sensor of type of meet an emergency in move plant accelerometer, join is on tailstock of deep aperture drilling machine; Sending feeling unit is general, the drill pipe that applies to all and different diameter (the aperture) that applies to all and different diameter namely; Hiring band is the card bind round that originallies closely to go up in drill pipe, its action is axial of the susceptive on drill pipe force and torque are delivered send feeling unit, bag bind round is not current, want with drill pipe form a complete set. The deep Kong Zuan that passes move plant institute to be collected cuts force and oscillatory signal, collect through data implement convey to analyse processing in the computer, get the accurate result that monitor, the real time that can be used at getting to cut a course is controlled. The experiment studies this experiment considered to collect deep Kong Jia in industry procedure normal with invalidation signal of 2 kinds of operating mode. Use all sorts of wearing away the broach with different level and change cutting dosage, achieve the signal of disparate test condition and the influence rate to signal thereby. Data is collected implement use 4 channels, collect respectively force, torque, perpendicular the oscillatory signal with horizontal way. Get the metabolic law of signal of condition of the cutting tool when cutting the data below condition to collect cutting tool to initial stage wears away and wear away normally normally; The change of change rotate speed to signal of cutting tool condition the rule; The change feed change to signal of cutting tool condition the rule. The data below imitate state of fault collects cutting tool to wear away quickly the metabolic rule with signal of condition of the cutting tool when damaged; Cut the metabolic pattern of signal of condition of the cutting tool when bits jams; Direct the metabolic rule of a signal of condition of the cutting tool when burst. Because signal analysis measured signal to contain the useful information that reflects object job condition not only, also included a large number of useless background noise at the same time, the useful information that reflects working condition often floods the background noise in no use in, discover very hard directly commonly and extract. Be opposite 4 kinds of signal undertake vibration of force, torque, perpendicular direction and horizontal way vibration time domain weaveform analyses discovery, primitive signal is distinguished inside time domain very not apparent, what cannot be used at broach is online monitor, for this, should undertake handling to primitive signal. Make discover when analysis of frequency domain signal to 4 kinds of signal, force and torque signal do not have apparent chart peak in frequency domain, also do not have clear metabolic rule, only the ability when cutting tool damaged has diagnostic frequency, use time alignment analysis to force and torque signal so. Use AR model to undertake to its parameter estimates, the incomplete of the sequential model that gains strength differs equation. The oscillatory signal change in frequency domain signal is apparent, analyse via the spectrum of the oscillatory signal to horizontal way and perpendicular direction, see a picture 2. Graph Duan Zhen of 2 low frequency moves analysis of signal power chart to be able to reach by afore-mentioned analysises the following conclusion: The characteristic that oscillatory signal power spectral density changes in chart of low frequency Duan Pin is as cutting tool tatty increases, amplitude of the highest peak in a mountain range increases faster, next incline to is gently, and position of frequency of the highest peak in a mountain range criterion by high frequency move to low frequency direction. Wear away quickly when cutting tool when close damaged, frequency of horizontal way the highest peak in a mountain range is centered in 400Hz to control, and appear a lot of new chart peaks; Frequency of perpendicular direction the highest peak in a mountain range is centered in 210Hz left and right sides, also appear a lot of new music peaks. This kind changes, reflected generation of cutting tool tatty and progress, arouse a cutting tool through cutting force - workpiece - the harmonic composition of system of machine tool treatment, bring about parameter of systematic vibration mode to produce change. The occurrence of peak of much a fair amount of confidence, make signal energy dispersive, accordingly, chart peak amplitude increases slow, and the cutting damp that cutting tool and aggravation of condition of workpiece contact area cause increases, make position of frequency of perpendicular and oscillatory the highest peak in a mountain range is reduced. The metabolic rule that chart of rate of high frequency Duan Gong pursues and low frequency paragraph agree basically, but the rule does not have low frequency paragraph outstanding, the frequency that amplitude of the highest peak in a mountain range appears in high frequency area is very small, chart peak is little, because cutting is machined,this is of the system force Zhen Yuan to be in commonly 1, 000Hz less than. Accordingly, high frequency chart is kept apart effectively or weaken the harmonic composition that processes a system. Pattern recognition condition is classified whether successful, depend on greatly diagnostic analysis is chosen with what the feature measures. Deep Kong Jia is versed in dynamic system is random process, be analysed very hard with function of some determinism time. The purpose that the feature analyses alternates primitive signal namely measure for the feature, find out the relation of it and operating mode. Capacity of this kind of feature is very much, but should reflect operating mode condition the rule, sensitivity and get together mediumly in mode space type, separability is not identical, the diagnostic volume with good, strong sensitivity regards regularity of the choice on the foundation that need analyses in the feature as mode vector, make its have better separability. This experimental research was analysing force of oscillatory, axial and torque signal to be in respectively the characteristic of time domain, frequency domain, on the base that using chart of sequential model, power and auto-correlation chart to wait, the incomplete of the sequential model of the diagnostic chart peak that chooses horizontal vibration and force differs variance to regard mode as vector, use perception implement algorithm obtains W=(-4 of the vector that increase advantageous position, 540, 131, 090, 300)T. Accordingly, get classified function: Through afore-mentioned classified function we determined the treatment position of cutting tool, if express 1 examine example place is shown. Via experimenting for many times, correct rate amounts to 90% above. Epilogue experiment proves, designs deep Kong Zuan cuts much sensor to measure force and measure brace up device, installation is simple, move reliable, assorted signal has complementary sex, the position that cutting tool cuts to getting in raising monitored accuracy to apply to deep Kong Zuan to cut a course undertakes monitoring. Oscillatory signal has reflective cutting tool to wear away in frequency domain the striking feature of condition. Get in what differ cut condition and auger cut dosage to fall. Change of its power chart is different. But metabolic trend is consistent. The incomplete that chooses amplitude of horizontal vibration the highest peak in a mountain range and force differs variance to regard a feature as vector. Form mode vector. Can reflect the condition that deep Kong Jia is versed in the system moves in the round. CNC Milling