Many spare units on site arrive this way: the budget approves an identical machine, it is put to one side without any pipework connected, and the drawings mark it as standby. When the main unit fails and the standby is connected, it turns out that the pipework connections do not match, the vacuum source is shared, the standby has not been run for a long time, and the condition of the pump oil and the seals is unknown. One more machine in the count does not necessarily mean one more share of protection.
Parallel running and redundancy are two different accounts
Running in parallel divides the flow and the task between two machines: both work at the same time, each contributes a share, and the combined processing capacity is higher than that of one machine, but as soon as either one stops, the capacity halves at once. Redundancy holds the capacity back: one machine works and the other stands by, and when the working machine fails the standby takes over, so the processing capacity does not lose anything. The two have completely different goals, so their configuration logic differs naturally.
Parallel running pursues total output, with the focus on the load distribution and the balance of the pipework between the two machines; redundancy pursues availability, with the focus on how quickly the switch-over happens and whether the standby machine can establish the same vacuum level and flow rate by itself. Mixing these two together makes the configuration work against itself.
| Point of comparison | Parallel running | Standby redundancy |
|---|---|---|
| During normal running | Both machines contribute | One runs, one stands by |
| When one machine fails | Processing capacity halves | Capacity is maintained after the switch-over |
| Requirement on the pipework | Both connected to the same header, and it has to be balanced | Two independent paths, with a reliable switching point |
| Equipment utilisation | High | The standby machine stands idle for long periods |
For standby switching to hold, three conditions have to be met
First, the standby machine has to be able to establish the same working conditions as the main machine by itself. If the two machines share one vacuum source, the vacuum source stops when the main machine stops, and after the standby is connected it still has to build the vacuum from scratch, which lengthens the switch-over time considerably. Second, the switching process must not break the oil path for long. A connection method with a shut-off valve and a non-return structure allows the switch-over to be completed while the pipework stays connected, so there is no need to stop the pumps and decant the oil on site. Third, the standby machine has to be run regularly.
The third point is the one most easily overlooked. A standby pump that has not turned for six months has a mechanical seal that has settled in one position, the pump oil may have already stratified, and the motor winding insulation falls in a long-term damp environment. When it is really needed and started for the first time, the success rate is not one hundred per cent. Regular trial switch-overs and regular no-load running are the step most easily left out of a redundancy configuration, and the one that should be left out least.
- Write the trial switch-over into the maintenance plan, with the interval, the responsible person and the record sheet all fixed.
- Record the trial switch-over and the trial run separately: the former looks at the switching action, the latter at the processing result.
- The filter elements and the pump oil of the standby machine have to be changed on the same cycle; do not exempt them because it does not run.
- There has to be a fallback action when the switch-over fails; it is not permitted to force the connection with the pressure not released and the pipework not open.
The redundancy options a single machine offers
Part of the redundancy does not have to be achieved by adding whole machines; it can be carried by the configuration of a single unit. The dual vacuum pump set of the ZYD-300 can run as duty and standby, so one machine on its own covers a larger daily throughput, and the standby vacuum pump set redundancy can be selected directly in the configuration; in the safety design of the ZYD-200, the overpressure alarm shutdown, the phase loss and phase sequence protection and the dry burn protection together form redundancy in the electrical safety. All of these are ways of narrowing the area affected by a fault without adding another whole machine.
The criterion for the selection can be plain: how much does the site lose if this oil treatment task stands still for one day. That loss determines how far the redundancy has to go: whether to keep one more whole machine, to give a single machine a duty and standby pump set, or simply to make sure the key consumables have spare parts available.
Buying one more machine is an outlay; holding on to one more share of processing capacity is the protection, and the two cannot replace each other on the account.
