OEE Calculator

$T_{pp}$ is the planned production time, the time in which the machine is meant to produce: the shift length $T_s$ less the planned stops $T_{pl}$, all in minutes. $T_r$ is the run time that is left after $T_d$, the time lost to unplanned stops and changeovers. Availability $A$ is the share of the planned production time in which the machine runs, from 0 to 100%.

Run time and availability follow Muchiri and Pintelon (2008), with their operating time as the run time here. Their formulas do not define loading time; it is read here as the planned production time, as in their TEEP diagram (Fig. 3), which takes the planned maintenance time out of the total time first, and consistent with their conclusion that scheduled downtime should not be used in OEE calculation.

• Changeovers count as stops, an availability loss: in Nakajima’s six big losses, as described by Muchiri and Pintelon (2008), set-up and adjustment is a downtime loss, and their framework keeps changeovers inside OEE. Plants that plan changeovers outside production time can enter them under planned stops instead. That is this calculator’s own allowance, not part of their framework; availability and OEE then come out higher, so OEE figures from different plants compare only when they follow the same rule.
• Planned stops are taken out of the time before availability is worked out. Muchiri and Pintelon (2008) conclude that only the production losses related to the equipment should be used in OEE, and that scheduled downtime should not be. Time with no orders is not scheduled downtime in their classification (Fig. 7 lists “No Demand or Low Demand Time” as a commercial loss), but it falls outside OEE by the first rule.
• Waiting for material, an operator or an inspection counts here as a stop, an availability loss. That is this calculator’s own rule: Muchiri and Pintelon (2008) do not classify waiting as such. Their framework puts a delay in raw material arrival with the logistic problems, outside OEE (Fig. 7), and keeps idling inside it as a speed loss, so where a wait belongs in their terms depends on its cause.
• A stop counts as an availability loss only when it is entered under unplanned stops. Minor stops left out of it lower performance instead; OEE is the same either way.
• The stops are entered as one total for the period, not stop by stop.

$R_i$ is the ideal rate in parts per hour: $n_c$ parts every ideal cycle time $t_c$, in seconds. $T_n$ is the net run time in minutes: how long the $N$ parts made, good and rejected, would take at the ideal cycle time. Performance $P$ is the share of the run time $T_r$, from Availability, that the parts would need at the ideal speed. Performance follows Muchiri and Pintelon (2008): their theoretical cycle time is the ideal cycle time per part here, $t_c / n_c$, and their actual output the parts made. The ideal rate and the net run time are this calculator’s restatements.

• The ideal cycle time is the cycle at the machine’s design speed. A target or an average cycle time instead hides part of the loss and can give performance above 100%.
• Every machine cycle makes the same number of parts.
• Short stops that are not entered under unplanned stops fall into the performance loss; the calculator does not separate them from slow cycles.
• All the parts made are counted, rejects included, so that the time spent on rejects is counted once, under quality.

Quality $Q$ is the share of good parts: $N_g$ is the number of good parts and $N$ the parts made, so $N - N_g$ are the rejects. $T_{fp}$ is the fully productive time in minutes: the time the good parts would take at the ideal cycle time $t_c$ for $n_c$ parts. Quality follows Muchiri and Pintelon (2008), with the good parts as their total production less the defect amount. The fully productive time is this calculator’s restatement.

• The quality loss is the machine time spent on rejects, counted at the ideal cycle time. The cost of scrapped material is not included.
• Only the rejects made by this machine in this period are counted.

OEE is the product of $A$, $P$ and $Q$, availability, performance and quality from the steps above. $T_{fp}$ is the fully productive time and $T_{pp}$ the planned production time, in minutes. $L_A$ is the time lost to stops, equal to $T_d$; $L_P$ the time lost to running below the ideal speed, the run time $T_r$ less the net run time $T_n$; and $L_Q$ the time lost to rejects, the $N - N_g$ rejected parts at the ideal cycle time $t_c$ for $n_c$ parts, all in minutes. With the fully productive time, they add up to the planned production time.

• The three losses and the fully productive time add up to the planned production time exactly; the calculator uses no approximation.
• OEE covers one machine, or a line run as one unit without buffers between stages, in the period entered (see the questions below). Planned downtime, which TEEP (total equipment effectiveness performance) also counts, is not part of it (Muchiri and Pintelon, 2008).

$K_A$ is the cost of the availability loss in one period: $c_h$, the cost of one hour of the machine or line in dollars, times $L_A$, the availability loss in minutes, over 60. $K_P$ is the same for the performance loss $L_P$, and $K_Q$ for the quality loss $L_Q$. $K$ is their sum, equal to the cost of the planned production time $T_{pp}$ less the fully productive time $T_{fp}$; $K_y$ is the cost in a year of $n_s$ periods a week and $n_w$ weeks.

$\Delta N$ is the number of good parts the planned production time would give with no rejects, $n_c$ parts every ideal cycle time $t_c$, less the good parts made $N_g$; $\Delta N_y$ is the same for a year.

• This step is the calculator’s own method: Muchiri and Pintelon (2008) note that cost is not measured in OEE.
• One hourly rate is used for every lost minute. Scrapped material, energy and lost sales are not included.
• The good parts lost count as lost sales only when every extra part could have been sold.
• The year repeats the period entered: all the shifts are assumed to be alike.