availability by undertaking essential maintenance tasks. However, when availability drops below a specific threshold, the asset’s performance will not improve unless it is upgraded or replaced. There are several ways to calculate availability. The first is as follows:
An alternative means of determining availability is as follows:
Here is a simple example using the uptime/downtime method: A critical machine ran for 700 hours in a given month. During that month, the asset also had 12 hours of unplanned downtime because of a breakdown, and 8 hours of downtime for weekly PMs, which equals 20 hours of total downtime. Therefore 700 uptime hours + 20 downtime hours = 720 total hours. Using these numbers, we can determine the availability for the month is equal to 700/(700+20) = 97.22%.
Tracking availability can help identify opportunities for improvements by identifying problematic equipment. The typical availability benchmark is above 95% for most assets. However, it can differ depending on how necessary the equipment is to your operations.
Figure 2.7 Hypothetical machine history. Green (solid) arrows indicate machine is running and red (dashed) arrows indicate machine is down for maintenance.
Critical Machine Events
Without historical lifecycle data, we cannot make objective decisions about machines. To determine a critical machine’s availability, we need to know how long it ran between outages and how long it took to make repairs. Therefore, we need a database that can capture life cycle events for all your critical machines.
Consider the life cycle of the hypothetical machine seen in Figure 2.7. At t1 the machine is started for the first time. The machine runs reliably from time t1 to t2 and fails. The machine is repaired from time t2 to t3 and then restarted at time t3. The machine runs from t3 to t4 and fails. The machine is repaired from time t4 to t5 and then restarted. The machine runs from t5 to t6 and fails at time t6. The machine is repaired from time t6 to t7 and then restarted.
All of these start-up and shutdown times should be captured in machine records. In addition to these times, you also need to record:
Repair or PM work scope performed during downtimes.
Denote of the outage was planned or unplanned.
Repair costs.
Root causes of unplanned work.
Losses incurred as a result of an unplanned outage. Losses can be related to production losses, demurrage, fines, etc.
If all this information is available, then the machine’s availability (Ao), MTBF, MTTR can be determined and trended.
Some other tools used to track critical machines are:
Process Outage Trends
A production outage is when a processing unit is shut down due to a machinery, equipment, or processing issue. Unplanned production losses are especially costly. A great deal of effort is expended to ensure production losses are minimized. Production losses outages are typically reported in, 1) hours or days of downtime, 2) pound of production lost, 3) barrels of production lost, and 4) dollars of production lost, etc. If a production unit has no “make-up” capacity, then the production lost during an outage is lost forever.
Figure 2.8 A trend of machinery outage cost will tell you the magnitude of the production losses over a period of time and if the losses are increasing, decreasing, or staying constant.
Here are a few types of plots that can be used to track process outages:
1 Production losses related to machinery outages (see Figure 2.8)
2 Breakdown of unit outages (see Figure 2.9)
3 Outage downtimes causes (machinery, exchangers, controls, etc.) (see Figure 2.10)
4 Root causes of outages (see Figure 2.11)
Process Outage Related to Machinery Outages
Pareto of Production Losses Across a Site
Figure 2.9 A Pareto of production losses across a site. This Pareto indicates that most (89.4%) of production losses are associated with four process units. The next step would be to drill down into this data to determine the root causes of these outages and address them.
Pareto downtimes causes (machinery, exchangers, controls, etc.)
A Pareto made up of RCFA causes can assist your organization in determining:
1 The best types of training that might benefit your organization.
2 How to best use your predictive maintenance resources.
3 Where design upgrades could be justified.
Note that to generate this type of Pareto chart your organization will need to perform RCFAs to determine the root cause of failures. RCFAs are essential for organizations that wish to have operational reliability.
Figure 2.10 Pareto of causes of production losses over the last 12 months. This Pareto summarizes the reasons for unit outages. At a glance, management can see what is causing production losses. In this hypothetical example, we see that 68.7% of the process losses were caused by operational factors and heat exchanger issues. One conclusion you could draw from these results is that more operator training may be warranted to reduce process losses. Another recommendation that could stem from this Pareto is to review the present heat exchanger inspection program to see if improvements could be
