Танай системийн SPOF шинжилгээг үнэгүй хийлгэе гэвэл: 7733-4478 эсвэл энэ сайтын AI туслахтай ярилцаарай.
In a continuous plant, even a few minutes of control-system downtime is expensive. Yet every device eventually fails — the question is not "will it fail?" but "what happens when it does?". Redundancy means architecting the system with duplicated elements so that operation survives any single failure. The guiding principle: no Single Point of Failure (SPOF).
Why redundancy is needed
Production loss. In continuous processes (processing plants, heating plants, water utilities) a control-system stop halts the whole line.
Safety. Some processes cannot simply be stopped: cyanide leaching, high-pressure boilers, chemical reactions — losing control creates immediate risk to equipment and people.
Equipment damage. Sudden stops cause their own failures: mills stopping under load, freezing lines, water hammer.
Restart cost. Restarting a large process takes hours to days, and every start-up adds risk.
How the risk is assessed and planned
Redundancy is an engineering decision based on numbers, not a nice-to-have:
Price the downtime. What does one hour cost? (lost production + restart cost + contractual penalties + equipment risk)
SPOF analysis. Walk the system diagram asking "if this one thing fails, what stops?" — power, CPU, switch, cable, server, disk...
Probability × consequence. Assess each element's MTBF and MTTR, and rank the risks.
Compare costs. Weigh the redundancy investment against the downtime price — a Hot Standby CPU usually costs less than one shift of lost production.
Redundancy levels
Redundancy is applied level by level, not "duplicate everything": power (dual feeds, UPS) → I/O and field network (dual-ring Ethernet, distributed I/O) → controller (Hot Standby CPU pair) → upper network (dual switches and paths) → SCADA servers (primary/standby with synchronization) → clients (multiple stations, automatic failover) → data (redundant historian, backups).
Key principle: the weakest link sets the system level. Duplicating CPUs but connecting them through one switch just moves the SPOF.
Which case needs which solution
Cold, warm and hot standby compared: recovery time, data, cost
Cold standby — spare on the shelf, manual replacement; recovery in hours to days. For tolerant auxiliary systems.
Warm standby — backup powered and waiting, manual/semi-automatic switchover; recovery in minutes. For important systems tolerating a short break.
Hot standby — both units run in parallel, state continuously synchronized, automatic bumpless switchover. Mandatory for continuous and hazardous processes.
Voting (2oo3) — three channels vote by majority; the domain of safety (SIS/SIL) systems.
Modicon M580 Hot Standby (HSBY)
Primary + Standby CPU pair — fully synchronized every scan over a dedicated HSBY link: program state, memory, timers, counters
Automatic bumpless switchover — the standby takes control instantly; outputs don't bump, the process doesn't notice
X80 Ethernet RIO — CRA-adapter drops on a dual-ring network: one broken cable loses no I/O
Hot-swap I/O modules during operation
CCOTF — change configuration on the fly: add drops and modules without stopping
Web diagnostics for the CPUs and every drop
Fully redundant architecture: SCADA server pair, dual network, M580 HSBY pair, dual ring, RIO drops
We deployed exactly this architecture at Erdene Mongol's Bayan Khundii gold plant: Hot Standby M580 pair, X80 RIO drops in every plant area, dual-ring Ethernet — no single point of failure.
AVEVA Plant SCADA redundancy
Server pairs (Primary/Standby) for the I/O, Alarm, Trend and Report server roles with automatic takeover
Automatic client failover — operator stations switch servers by themselves, with no operator action
Data synchronization — alarm states, acknowledgements and trends kept identical on both servers; nothing is lost in a switchover
Redundant network paths between servers, clients and PLCs
Distributed architecture and clustering — each cluster carries its own redundancy inside one unified system
Online changes deployed without interrupting operation
PLC-level HSBY and SCADA-level server redundancy complement each other into an unbroken chain: a single failure at any level is invisible to the operator and the process.
Conclusion
Redundancy is not a cost — it is insurance. Proper design starts from the downtime price and risk analysis, then balances every level. We deliver the full cycle: audit, Hot Standby architecture design, implementation and zero-downtime migration.
Get a free SPOF analysis of your system: 7733-4478 or talk to the AI assistant on this site.