uptime
Chapter 11 uptime + downtime
Availability
Concepts and Definitions (11.1)
Exponential Model (11.2)
System Availability (11.3)
Inspect and repair Model (11.4)
C. Ebeling, Intro to Reliability & Maintainability
Engineering, 2nd ed. Waveland Press, Inc. Copyright ©
2010
11.1 Concepts and Definitions
Definition: Availability is the probability that a system or
component is performing its required function at a given
point in time or over a stated period of time when
operated and maintained in a prescribed manner.
Chapter 11 2
Availabilities
1. A(t) = the availability at time t - referred to as the
point availability
2. A( T ) =
1
T z
0
T
A( t ) dt is the average availability over [0,T]
generalize: At 2 - t 1 =
1
t2 - t1
z
t2
t1
A(t)dt
3. A = lim A( T ) is the steady-state availability
T →∞
Chapter 11 3
Steady-State Availability
To have steady-state, a renewal process is necessary
1. Inherent A = lim A( T ) = MTBF
inh
T →∞ MTBF + MTTR
2. Achieved
MTBM
Aa =
MTBM + M
td
MTBM =
td m(t d ) MTTR + MPMT
td Tp
m( t d ) + M=
Tp m( t d ) + t d / Tp
Chapter 11
For a renewal process: m(td ) = td / MTBF 4
Achieved Availability
Availability
1
0.9
0.8
Achieved Availability
0.7
0.6
0.5
inherent availability
0.4
0.3
0.2
0.1
0
100
150
200
300
400
500
10
15
20
25
30
50
5
MTBF = a + b/Tpm PM Interval
Chapter 11 5
Operational Availability
MTBM
Ao =
MTBM + M '
td
m( t d ) MTR + MPMT
Tp
M' = and MTR = MTTR+SDT+MDT
m( t d ) + t d / Tp
MTBM + ready time
AG =
MTBM + ready time + M
Chapter 11 6
Availability with Minimal Repair
Inherent Availability
t2 − t1
t2
At2 −t1 = where m(t1 , t2 ) = ∫ ρ (t ) dt
t2 − t1 + m(t1 , t2 ) ⋅ MTTR t1
Achieved Availability
t2 − t1
At2 −t1 =
⎛ t2 − t1 ⎞
t2 − t1 + m(t1 , t2 ) ⋅ MTTR + ⎜ ⎟ MPMT
⎜ T ⎟
⎝ pm ⎠
Chapter 11 7
Availability with Minimal Repair
A machine has minimal repair upon failure with an
intensity function of ρ(t) = .00017t.7, t measured in
days. Repair times are lognormal with a median repair
of 20.3 hours and s=1.2. Estimate its availability over
its first 5 operating years assuming 7/24.
m(t) = .0001t1.7 and m(5 x 365) = 35.0
MTTR = 20.3 exp[1.22 /2] = 41.7 hrs = 41.7/24 = 1.74 days
A(5 oper yrs) = 5 x 365 / [5 x 365 + 35 x 1.74] = .9677
Chapter 11 8
11.2 Exponential Model
λ
1 2
r
d P1 (t)
= - λ P1 (t) + r P 2 (t)
dt
P 1 (t)+ P 2 (t) = 1
Chapter 11 9
Exponential Model
Point Availability
1.000
0.990
0.980
0.970
0.960
0.950
0.940
0.930
time
r λ -( λ +r)t
A( t ) = P1 (t) = + e
λ+r λ+r
Chapter 11 10
Exponential Model - Interval Availability
At 2 - t 1 = 1
t 2 -t 1 z t2 FHr
t 1 r+ λ
+ λ
r+ λ
IK
e-( λ+r)t dt
r λ
= + e
-( λ +r)t 1
- e
-( λ +r)t 2
r + λ (r + λ ) ( t 2 - t 1 )
2
steady-state Ainh = lim At − o
t →∞
r MTBF
= =
r+λ MTBF + MTTR
Chapter 11 11
11.3 System Availability
Series versus Parallel
n
As (t) = Π A (t)
i=1
i
n
As (t) = 1 - Π (1 -
i=1
Ai (t))
Chapter 11 12
Standby Systems
dP1 ( t )
= − λ 1 P1 ( t ) + rP2 ( t )
dt
dP2 ( t )
= λ 1 P1 ( t ) + rP4 ( t ) − ( λ 2 + r ) P2 ( t )
dt
P1 ( t ) + P2 ( t ) + P4 ( t ) = 1
lim d Pi (t)
for steady-state: = 0 and set Pi(t) = Pi
t → ∞ dt
Chapter 11 13
Steady-State Standby Systems
-λ 1 P1 + r P2 = 0
λ 1 P1 + r P4 - (λ 2 + r) P2 = 0
P 1 + P2 + P4 = 1
L
P = M1 +
λ λλ O
−1
N r r PQ
+ 1 1 2
1 2
λ1
P2 = P1
r
λ 1λ 2 A = P 1 + P2
P4 = P1
r2
Chapter 11 14
Generalize - steady-state availability
Σ j Rate into state i from state j x Pj
= Rate out of state i x Pi
Chapter 11 15
Example #1
- λ 1 P1 - λ 3 P1 + rP 3 = 0
λ 1 P1 - λ 2 P 2 = 0
λ1 1
λ3 P1 + P 2 + P 3 = 1
λ1 λ 1 + λ 3 -1
r P1 = (1 + + )
λ2 r
2 3
=
λ 1
P2 P1
λ2 λ2
λ1+ λ3
A = P 1 + P2 P3 = P1
r
Chapter 11 16
Example #2
(λ 1 + λ 2 ) P1 = r P4
λ1 1
λ2
λ 2 P2 = λ 1 P1
r
λ 1 P3 = λ 2 P1
2 3
r P4 = λ 2 P 2 + λ 1 P 3
λ2 4 λ1
P1 + P2 + P3 + P4 = 1
Chapter 11 17
11.4 Inspect and Repair Model
Let R(t) be the dormant failure distribution,
t1 = inspection time,
t2 = repair time if necessary
T = time between inspections
Then T + t1 + t2 [ 1 - R(T) ] is the expected cycle time
Time to inspect
and repair?
Chapter 11 22
Inspect and Repair Model
z z
T T
Expected uptime = R (t )dt = R ( T ) T + tf ( t ) dt
0
z z
o
T T
t f ( t ) dt = − tR ( t )| + R ( t ) dt
T
0
z z
0 0
T T
R (t )dt R (T )T + tf (t )dt
A( T ) = 0
= o
T + t1 + t 2 [1 − R ( T )] T + t1 + t 2 [1 − R (T )]
Chapter 11 23
Exponential Failure Time
T T − λt T
e
∫0 R(t )dt = ∫0 e dt = −λ
− λt
=
−λ
(
1 − λT
e − 1) =
1
λ
(1 − e − λT
)
0
Chapter 11 24
Inspect and Repair Model
Exponential Case
− λT
1− e
A( T ) = − λT
λ[ T + t1 + t 2 (1 − e )]
If t1 and t2 are negligible
− λT
1− e
A( T ) ≈
λT
Chapter 11 25
Example
1 − e − .0002 T
A( T ) =
.0002[ T + 16 + 48(1 − e − .0002 T )]
T 100 200 300 400 500 600 700 800 900 1000
Avail .847 .900 .913 .916 .914 .910 .904 .898 .891 .884
CFR=.0002
resupply time = 48 hr inspect time = 16 hr.
Chapter 11 26