Mining Equipment Maintenance
Fundamental Concepts
Old Underground Coal Mines
Modern Longwall Mine
The effect of Mechanisation on Productivity
25000
20000
Tonnes/Man-Year Produced in Queensland Coal Mines
15000
10000
5000
19 16 19 20 19 25 19 30 19 35 19 40 19 45 19 50 19 52 19 55 19 60 19 69 19 72 19 75 19 80 19 85 19 88 19 95 20 03
Maintenance Costs
30-50% of the operating costs Annual bill is about $10 billion Equally significant is the cost of lost production when the machine is down Every 1% improvement in equipment availability or productivity improves the company profits by up to 3.5%
Objective Function for Maximising Company profits
The Annual Production for a given investment is a suitable objective function
Total Annual Production
Is this total annual production ? Tonnes/Hour x Total Hours More realistic prediction is Tonnes/Hour x (TotalHours - Downtime) Probably the following is more illuminating Tonnes/Hour x (TotalHours -Scheduled Downtime - Breakdown maintenance)
Total Hours of Possible Production
Subtract Planned Maintenance
Subtract Breakdowns
This is the time available for production. Obviously, you want to maximise this time. That is why the ratio of this time to the total time is a very important KPI.
Availability
TH - PM - BM A= TH
TH PM BM Total Hours Planned Maintenance Breakdown Maintenance
Availability
TH - PM - BM Operating Time ! A= TH Total Time
Is also expressed in terms of MTBF and MTTR as
MTBF A= MTBF+MTTR
Availability
TH - PM - BM Operating Time ! A= TH Total Time
Is also expressed in terms of MTBF and MTTR as
MTBF A= MTBF+MTTR
This is the textbook definition of AVAILABILITY
Availability
Operating Time = N x MTBF Downtime = N x MTTR Total Time = N x (MTBF+MTTR)
Operating Time N x MTBF A= ! Total Time N x (MTBF+MTTR)
Maximise the Return to the Company
The aim for the company is to maximise the return from its productive assets. Let us keep it simple and express this as maximising the production through the year. To achieve this, (a) The machine availability must be high (b) The machine must be producing at a high rate when it is operating The solution is a compromise between the two. We will see this in an example.
Optimum Bucket Size
The recommended suspended load for your dragline is 150 tonnes. A competitor put a larger bucket on a similar unit and they are running it at a suspended load of 170 tonnes. Your Manager wants you to install an even bigger bucket, taking your suspended load up to 210 tonnes. You know that the safe static load for this dragline is 250 tonnes. -Does a larger bucket necessarily mean higher production? -Why not increase it to 250 tonnes if this is the safe load? -How would you determine the optimum bucket size?
Question
Present Production Po = 9000 t/h.
A=
TH - PM - BM TH
Total Mine Operating Time in a Year = 8640h 12-hour planned maintenance shift every month. Breakdown maintenance downtime is expected to vary with the production rate as
BM ! 336 " R
P is the production rate with the bigger bucket. Find R that maximises the annual production.
P where R ! Po
Solution
Lo Br st p ea ro k d du ow ct n m i on ain du te e t na o nc e
Optimum Point
Dependence on MTBF
Maintenance
Maintenance must be considered in the context of asset utilisation Mining is an asset-rich industry Optimum utilisation of these assets is the only way a company will stay competitive This is a task for both production and maintenance engineers.
Some Basic Concepts
Failure
Loss of ability of an item to perform its required function
Failure
Broken teeth on shearer downdrive gear
Failure
Cost associated with this failure: Lost production when the machine was down Replacement gear Maintenance labour
Failure Rate, MTBF, MTTR
MTTR MTBF MTTR
Re pa ir
MTBF
Re pa ir
Re pa ir
Failure
100
200
Hours
Failure Rate, MTBF, MTTR
MTTR MTBF MTTR
Re pa ir
MTBF
Re pa ir
Re pa ir
100
200
Hours
MTBF = Mean Time Between Failure (100 h) Failure Rate = Number of failures per unit time (0.01 h-1) MTTR = Mean Time To Repair (20 h)
Historical Records
Failures occur randomly. The repair time is also not constant. How do we find MTBF and MTTR?
If we treat failure as a random event, then we can use the well-established tools of probability and statistics to model the uptime, downtime and availability for our equipment.
Random Failure?
All nature is but art, unknown to thee All chance, direction thou canst not see All discord, harmony not understood
Alexander Pope, Essay on Man
Poisson Process
Poisson distribution is commonly used in forecasting to represent the number of occurrences of a specific event in a given continuous interval. Ships arriving at a dock on a given day Traffic accidents on the SE freeway in a month Mad cow disease breakouts in the world in one year Typos per page in a long report typed by Hal Gurgenci Cable shovel failures in one day of operation
Poisson Distribution
p ( x) ! e
& %t
# %t $
x!
This is the probability distribution of the Poisson random variable X representing the number of outcomes occurring in a given time interval t. The parameter % is the average number of outcomes per unit time.
Reliability
Assume failure events follow a Poisson distribution. What is the probability of having NO FAILURES in a given time interval t? This can be found by substituting x=0 in the Poisson distribution function:
p (0) !
& %t
# %t $
0!
!e
& %t
This is referred to as the survival probability or the reliability
Reliability
R (t ) ! e
& %t
Reliability is the probability that a product will operate throughout a specified period without failure
when maintained in accordance with the manufacturer's instructions; and when not subjected to the environmental or operational stresses beyond limits stipulated by the manufacturer
The value of e
Two centuries ago, a Polish Statistician, Ladislaus Bortkiewicz, investigated the Prussian army fatalities caused by horse kicks. According to army reports, the rate was about one fatality every 1.64 years. Ladislaus collected the reports for one year. These were 200 reports and 109 recorded no deaths at all. Can you estimate the value of e using the above data?
Example
A major piece of equipment fails twice a day on average. Consider its reliability over a period of month. What is the probability of failure at any time during that month?
Reliability
R ! e &2t
Zoom In
R ! e &2t
Failure Probability
If the reliability, R(t), is the probability to survive through time t, then the probability of failing through that period is 1 R(t) or
F (t ) ! 1 & R (t ) ! 1 & e
Let us plot this
& %t
Failure Probability
This chart gives the probability of the failure from 0 to time t. In other words, it is the Cumulative Distribution Function for failures. How would we find the probability density function or p.d.f. This is sometimes useful. We differentiate the cumulative distribution function.
Failure Distribution Function
Cumulative Distribution Function
F (t ) ! 1 & R (t ) ! 1 & e
& %t
The time-derivative of the c.d.f. gives the Probability Distribution Function or p.d.f.
dF & %t ! %e f (t ) ! dt
This form of p.d.f. is called the Exponential Distribution. It represents the case when the hazard rate or failure rate, %, is constant over time.
Failure p.d.f.
The probability of failure through a unit time interval is given by
p!(
t '1 t
f (t )dt
Let us zoom into the p.d.f.
The probability of failure through a unit time interval is given by the area under the curve
p!(
t '1
1 1 f (t ) dt ) f (t ' ) "1 ) f (t ' ) ) f (t ) 2 2
Hazard Rate
The hazard rate is the conditional probability of failure in a small time interval (t, t+dt). It is conditional on there being no failure until t:
h(t ) ! f (t ) R (t )
For exponential failure distribution, the hazard rate is constant:
f (t ) % e & %t ! & %t ! % h(t ) ! R(t ) e
Constant Hazard Rate
The exponential distribution corresponds to a constant failure rate a.k.a. constant hazard rate
Is % always constant ?
f (t ) ! % e
Failure p.d.f.
& %t
The constant % is the failure rate (%= 1/MTBF) So far, we treated % as constant
Reliability function
R (t ) ! e
& %t
This is called the exponential distribution Is this always true? Let us first give an example
Uniformly Increasing Rate
Assume the failure rate is increasing by the formula 0.1t where t is measured in days It starts from zero and at the end of the month, the hazard rate is 3 failures per day. How do we generate the reliability function for this component?
Reliability with varying m=0.1 hazard rate
In a sample of N, dN will fail in a time interval dt
dN ! & Nmtdt
t dN ! & mtdt ln N ! & m ' ln C 2 N
N ! Ce
& mt 2 / 2
! Noe
& mt 2 / 2
Where N is the value of N at t=0
Then the reliability function is
N & mt 2 / 2 !e No
This corresponds to a Weibull distribution
Weibull Reliability
R!e
R * + Shape factor Scale factor
,t&. / 0+ 1
The probability of surviving through time t
Weibull Distribution Curves
c.d.f.
F (t ) ! 1 & R(t ) ! 1 & e
dF * f (t ) ! ! *t dt + e
,t&. / 0+ 1
*
p.d.f.
,t&. / * &1 0 + 1
Hazard Rate
f (t ) * * &1 h(t ) ! ! *t R (t ) +
Weibull curves with different hazard functions
R!e
,t&. / 0+ 1
*
* * &1 h(t ) ! * t +
Bathtub Curve
Failure Rate, %
Infant Mortality
Useful Working Life
Falling Apart
Optimum Point to Replace
Life in operation, hours
Actual Failure Patterns
4% 2% 5% 7% 82%
(68% of this with infant mortality)
These curves are the failure patterns observed on aircraft components in a study completed in 1978 by Nowlan and Heap. This shows that only 4% of the components go through a bathtub curve.
RELIABILITY OF SYSTEMS
Series Systems Parallel Systems (Redundancy)
Series Systems
A series system is a chain of components. When one of these parts fails, the entire system fails.
Series Systems
R ! RA RB RC
Parallel Systems
The failure for a parallel system means the failure of each individual component. The system failure probability is then the product of individual failure probabilities (1 R). A Most mining machinery systems are series systems. In other words, the failure of one component fails the entire system. The redundancy in mining can be provided by having multiple systems, eg spare trucks or shovels.
R ! 1 & (1 & RA )(1 & RB )(1 & RC )
Managing Reliability
Optimum utilisation of its capital investment in equipment is essential for company profits Equipment reliability plays a major role in this Therefore, managing reliability is a core business for a mining company This is a task for both production and maintenance engineers. In the rest of this presentation, we will talk about the maintenance function.
Maintenance Function
Preventive Maintenance
Prevent failures by performing a set of maintenance tasks at periodic intervals
Service Inspection Replacement
Corrective Maintenance
Repair after a failure to bring the machine back to an operating state
Which one delivers higher overall system availability?
Corrective Maintenance
We assume that corrective maintenance brings the system to as new state. Then it has no effect on system reliability Its impact on system availability is measure by the Mean Time To Repair (MTTR)
Mean Time To Repair
Fault Identification
What caused the failure? What needs to be repaired?
Set-up time
Find and bring the right person to the job Actual repair
Logistic delays
Waiting for the spare part
Restart time
Time spent to bring the system back to normal operation after the fault is repaired
How to minimise MTTR
Identify the failed components quickly. This is achieved by experienced operators, on-line fault detection tools For frequent failures have the repair crew with the right skills on standby Ditto for the frequently failing spare parts Design the equipment and the operating procedure to minimise re-starting time
PM Trade-Offs
The cost of failure
MTTR The cost of the repair and the replacement
The cash cost of the planned maintenance action (salaries, consumables, etc) The opportunity cost (lost production)
Preventive Maintenance Issues
Service
Effect on System Reliability
Inspection
P-F time (between potential and actual system failure)
Replacement
Failure distribution curves Effect on System Reliability
Would inspections help?
The time when we can recognise Potential Failure
P-F Interval
Failure Time
Scheduled inspections help when
Potential failure condition is clearly defined The P-F interval is consistent It is practical to inspect at intervals less than the P-F interval The P-F interval is long enough to implement corrective maintenance action
Scheduled replacements help when
The component breakdown has costly consequences (eg chain of failures, distance from the workshop, etc) The dominant failure mode is age-related with the hazard rate consistently increasing above an acceptable value at around the set replacement period
Weibull curves with different hazard functions
R!e
,t&. / 0+ 1
*
* * &1 h(t ) ! * t +
Periodic replacements - 1
* ! 0.5
Decreasing Hazard Rate Scheduled replacement increases failure probability
Periodic replacements - 2
* !1
Constant Hazard Rate Scheduled replacement has no effect on failure probability
Periodic replacements - 3
* !2
Increasing Hazard Rate Scheduled replacement decreases failure probability
Reliability Research
A significant part of the academic and research community has been continuing to develop increasingly complex mathematical models of the engineering systems and the expected modes of failure under various loading assumptions. While the intellectual rigour in these studies and the amount of effort that go into them cannot be ignored, the applications to real manufacturing and mining processes have been limited primarily for the lack of data needed to support these models
Industry Data
Analysis and Representation Tools
job_no
Site Maintenance Records
DaJob ojob (s job_title faci date_raised date_act_start date_act_end elapsed_ cosjob_reco trigger_reading s jo instructions R 432 R 432 R 432 R 432 1 2 3 4 1 2 8 1 1 2 3 4 1 5 6 1 7 8 1 1 1 2 3 4 3 1 1 1 3 1 Hoist Motors - George Gilbert & Co. -ED Petersen Report (Shutdown Hoist Motors - George Gilbert & Co. -ED Petersen Report (Shutdown Hoist Motors - George Gilbert & Co. -ED Petersen Report (Shutdown Hoist Motors - George Gilbert & Co. -ED Petersen Report (Shutdown E LIMIT SWITCH UPGRADE FOR DRE23 E LIMIT SWITCH UPGRADE FOR DRE23 E LIMIT SWITCH UPGRADE FOR DRE23 E LIMIT SWITCH UPGRADE FOR DRE23 REPLACE / OVERHAUL #1 HOIST BRAKE CYLINDERS REPLACE / OVERHAUL #1 HOIST BRAKE CYLINDERS AM MAINTENANCE SERVICE NO. 5 CHANGE HOIST BRAKE SOLINOID AM HOIST BRAKE SHOES CHANGEOUT AM HOIST BRAKE SHOES CHANGEOUT AM HOIST BRAKE SHOES CHANGEOUT AM HOIST BRAKE SHOES CHANGEOUT ADJ # 2 HOIST BRAKE AM HOIST BRAKE SHOES CHANGEOUT AM HOIST BRAKE SHOES CHANGEOUT CHANGE QUICK RELEASE VALVE ON # 3 HOIST BRAKE AM HOIST BRAKE SHOES CHANGEOUT AM HOIST BRAKE SHOES CHANGEOUT REPAIR HOIST BRAKE CHANGE CYLINDER ON # 2 HOIST BRAKE CYLINDER CHECK/ADJUST HOIST BRAKES. CHECK/ADJUST HOIST BRAKES. CHECK/ADJUST HOIST BRAKES. CHECK/ADJUST HOIST BRAKES. FIT TWO NEW RAIL CHOCKS CHECK ADJUSTMENT OF No 3 HOIST BRAKE REFITTED No 4 HOIST SHOE PINS FIT 2 NUTS TO #4 HOIST BRAKE SPRINGS AM HOIST BRAKE SHOES CHANGEOUT #4 CHECK #4 HOIST BRAKE DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR 11/04/96 11/04/96 11/04/96 11/04/96 6/05/96 19:50 6/05/96 19:50 6/05/96 19:50 6/05/96 19:50 30/06/96 1:15 30/06/96 1:15 29/07/96 14:03 8/08/96 16:56 21/10/96 16:11 21/10/96 16:11 21/10/96 16:11 21/10/96 16:11 21/10/96 19:28 21/10/96 16:11 21/10/96 16:11 21/10/96 21:59 21/10/96 16:11 21/10/96 16:11 5/11/96 5:29 13/11/96 16:14 4/12/96 21:37 4/12/96 21:37 4/12/96 21:37 4/12/96 21:37 30/12/96 16:23 27/12/96 6:35 27/01/97 20:09 5/02/97 20:41 19/03/97 19:41 17/03/97 4 55 18/03/96 18/03/96 18/03/96 18/03/96 18/03/96 18/03/96 18/03/96 18/03/96 28/06/96 13:00 28/06/96 13:00 30/07/96 12:30 8/08/96 15:20 21/10/96 21/10/96 21/10/96 21/10/96 21/10/96 15:30 21/10/96 21/10/96 21/10/96 21:00 21/10/96 21/10/96 5/11/96 4:00 13/11/96 4/12/96 17:10 4/12/96 17:10 4/12/96 17:10 4/12/96 17:10 27/12/96 26/12/96 23:15 27/01/97 10:30 4/02/97 19/03/97 17/03/97 1 45 22/04/96 22/04/96 22/04/96 22/04/96 22/04/96 22/04/96 22/04/96 22/04/96 29/06/96 1:00 29/06/96 1:00 30/07/96 22:00 21/10/96 21/10/96 21/10/96 21/10/96 21/10/96 16:30 21/10/96 21/10/96 21/10/96 21/10/96 5/11/96 5:15 13/11/96 840 ## MD 840 ## MD 840 ## MD 840 ## MD 12 ## RP 12 ## RP 9.5 ## SM DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR DR 3 ## RC 3 ## RC 3 ## RC 3 ## RC 1 ## RP 3 ## RC 3 ## RC 3 ## RC 3 ## RC 1.25 ## RP 2 ## RC DR DR DR DR DR DR 86,387.00 86,387.00 86,387.00 86,387.00 86,387.00 86,387.00 86,387.00 86,387.00 87,006.00 87,006.00 87,601.00 87,860.00 89,086.00 89,086.00 89,086.00 89,086.00 89,086.00 89,086.00 89,086.00 89,086.00 89,086.00 89,086.00 89,350.00 89,490.00 89,888.00 89,888.00 89,888.00 89,888.00 90,123.00 90,123.00 90,867.00 90,988.00 91,693.00 91 693 00 # CHANGE OUT THE "CATS WHISKER" LI # CHANGE OUT THE "CATS WHISKER" LI # CHANGE OUT THE "CATS WHISKER" LI # CHANGE OUT THE "CATS WHISKER" LI # # # # # # # 21/10/96 16:11:36 CHANGE #2 HOIST B # 21/10/96 16:11:36 CHANGE #2 HOIST B # # # # # # # # # # # # 3 4 4 4 NEED 2 EA 1" UNC NUTS 4 4 8/08/96 16:40 1.33333 ## RC
HOIST BRAKES
R29 R29 R29 R29 S0000055 S0000055 S0000055 S0000055 Z0025077 Z0025077 Z0026785 Z0027463 Z0031974 Z0031974 Z0031974 Z0031974 Z0032221 Z0031974 Z0031974 Z0032254 Z0031974 Z0031974 Z0033251 Z0032240 Z0035454 Z0035454 Z0035454 Z0035454 Z0033882 Z0037064 Z0039305 Z0039805 Z0043323 Z0043264
21/10/96 21:40 0.66667 ## RP
4/12/96 18:50 1.66667 ## AD 4/12/96 18:50 1.66667 ## AD 4/12/96 18:50 1.66667 ## AD 4/12/96 18:50 1.66667 ## AD 27/12/96 27/12/96 0:00 27/01/97 11:30 4/02/97 19/03/97 17/03/97 2 45 3 ## RP 0.75 ## AD 1 ## RP 0.5 ## RP 8 ## RC 1 ## AD
It is usually necessary to process 100000+ lines from multiple sources of sometimes questionable accuracy
Pareto Analysis
Pareto Principle : Significant few and Insignificant many In any application, a large part of the failures are due to a small number of causes A Pareto plot helps to identify the most significant causes The benefit is incurred only by attending the significant issues
Pareto Chart
Pareto Analysis for Longwall Face Equipment Failures
Shearer Drive Shaft
Scatter Plot Diagrams
A scatter plot is a logarithmic plot of MTTR against the number of failures N. Since the total downtime associated with each failure is NxMTTR, constant downtime curves appear as lines on logarithmic axes. Downtime MTTR ! 2 log MTTR ! C & log N N
Longwall Scatter Plot
Lines of Constant Downtime
Reliability Analysis
Pareto Analysis and Scatter Plots are good tools to identify the reliability sinks in the equipment The next step is to calculate the failure probability distribution curves for all critical components. The MTTR statistics may also be required if MTTR is not reasonably constant for each item This step requires high quality data
Features of High Quality Data
Large enough set to have at least 4-5 failure events for each target failure mode Cover a sufficiently long time period to eliminate local effects Uniform operating conditions over this period Accuracy free of collectors bias
Failure History Example
Suppose that we have the failure log for a component as 180, 216, 930, 990, 1300 and 1850 hours. Estimate the MTBF assuming an exponential probability distributions.
Failure Data (example)
Failure times 180h 216h 930h 990h 1300h 1850h 714 310 60 310 550 550 36 36 714 60
36 ' 714 ' 60 ' 310 ' 550 MTBF ! ! 334 5
Curve Fitting
The failure log was 180, 216, 930, 990, 1300 and 1850 hours. The TBF array , TBF = {36, 714, 60, 310, 550}.
i TBF Fi 1 36 20% 2 60 40% 3 310 60% 4 550 80% 5 714 100%
This is cumulative probability distribution data. We can then compare it against exponential or Weibull distributions. For example, use the following form to try an exponential distribution:
3 1 4 & %t ! 5 & ! % t F t e ln 6 1 ( ) 7 & F t 1 ( ) 8 9
Estimate the Failure Distribution
TBF, h 36 60 310 550 714 Rough Median Estimate Estimate 20% 10% 40% 30% 60% 50% 80% 70% 100% 90%
Exponential Fit t
F (t ) ! 1 & e
& %t
! 1& e
&
334
Longwall Equipment
Failure Probability Distribution Functions for some Critical Items
AFC Blockage/Overload (2002 only)
3 1 4 ln 6 7 81 & F (t ) 9
Exponential
AFC Blockage/Overload (2002 only)
Weibull
AFC Chain Failures
3 1 4 ln 6 7 81 & F (t ) 9
Exponential
AFC Chain Failures
Weibull
BSL Chain Failures
3 1 4 ln 6 7 81 & F (t ) 9
Exponential
BSL Chain Failures
Weibull
Shearer Cable Failures
3 1 4 ln 6 7 81 & F (t ) 9
Exponential
Shearer Cable Failures
Weibull
What is to be done?
Increase overall availability
Minimise the time spent on PM Decrease number of breakdowns
More effective PM Condition monitoring with long enough P-F time Engineering changes
Design changes Changes to the operating procedure
Decrease MTTR
Do all this while maximising the profit Learn lessons for next equipment purchase
The Reliability Function?
Failed =0 Failed =2 Failed=4 Failed =6
t=0 Failed=8
t=1 Failed=10
t=2 Failed =12
t=3 Failed=14
t=4
t=5
t=6
t=7
Is this best represented by a Weibull or an exponential distrubution?