Planning Electrical Infrastructure for Industrial Facilities in South Africa

Most switchgear procurement problems trace back to a decision made months earlier, at the infrastructure planning stage, long before anyone specified a panel. A site master plan that under-forecasts load growth forces a disruptive retrofit three years in. A single point of connection with no redundancy turns a routine transformer outage into a full production stoppage. A capital plan that treats the grid connection application as a formality, rather than the long lead-time item it actually is, delays a project by a year regardless of how fast the switchgear itself can be delivered.

This article is written for facility managers, project engineers, and finance teams responsible for planning new industrial sites or major capacity expansions in South Africa. It sits upstream of the equipment-level decisions covered in our switchgear sizing guide and buyer's guide — this is about the site-wide plan those decisions get made inside.

---

Start With a Load Forecast, Not a Current Requirement

A common planning mistake is sizing infrastructure to today's connected load. Electrical infrastructure — transformers, main switchboards, cable routes, servitudes — has a service life measured in decades, while the equipment it feeds gets replaced and expanded every few years. Plan against a load forecast that covers the site's realistic growth horizon (typically 10–15 years for primary infrastructure), not the load schedule for the equipment being commissioned this year.

Build the forecast from three components: committed load (equipment already ordered or firmly planned), probable load (expansions under active consideration, weighted by likelihood), and strategic headroom (capacity reserved for growth that hasn't been scoped yet, typically 20–30% on top of committed plus probable). Document the assumptions behind each figure — a forecast without assumptions can't be revisited when circumstances change, and it will change.

---

Map the Site Before You Size Anything

A site electrical master plan starts with a single-line diagram at the facility level: point of supply, main intake, primary distribution, and the major load centres it feeds — not yet the individual panel schedules that come later. This master single-line should show:

• Point of connection to the Eskom or municipal network, including the agreed or applied-for notified maximum demand (NMD)
• Primary distribution voltage and the reasoning for it (LV reticulation is simpler and cheaper up to a point; MV reticulation reduces cable losses and sizes on larger or more spread-out sites)
• Major load centres and their approximate location, so cable route lengths and voltage drop can be estimated early
• Reserved space and servitudes for future substations, cable routes, and switchrooms — space that is cheap to reserve now and often impossible to find later once the site is built out

Getting this map right before detailed design starts avoids the common failure mode where a site is built out load centre by load centre, each individually sensible, that collectively leaves no viable route for the next expansion.

---

Plan the Grid Connection Early — It Is the Longest Lead-Time Item

An Eskom or municipal supply upgrade — increasing notified maximum demand, moving from LV to MV supply, or securing a new point of connection — routinely takes 6–18 months from application to energisation, driven by network studies, servitude registration, and construction queues that are entirely outside the project's control. This is almost always the longest lead time in the entire project, longer than switchgear procurement, longer than civil works, and it is the item most often started too late because it doesn't feel urgent until the rest of the project is already committed to a date.

Submit the connection or upgrade application as soon as the load forecast is firm enough to state an NMD, even before detailed switchgear design is complete. If the final NMD shifts by the time detailed design finishes, it is far easier to amend an application already in the queue than to join the queue for the first time when the rest of the project is ready to energise.

---

Design In Redundancy Where an Outage Is Not Acceptable

Redundancy is a planning-stage decision, not an equipment-stage one — retrofitting a second intake or a bus-tie into a site designed with a single point of failure is disruptive and expensive compared with designing it in from the start. Decide, load centre by load centre, which parts of the site can tolerate a supply interruption and which cannot:

• N (no redundancy): acceptable for non-critical loads where an outage causes inconvenience, not loss — general lighting, non-process HVAC, offices.
• N+1 (single redundant path): the common standard for production loads, achieved with dual transformers and a bus-tie switchboard, or dual incomers from independent points of the network where the network topology supports it.
• N+2 or independent supplies: reserved for loads where an outage causes safety risk or unrecoverable loss — cold chain, continuous process lines, life-safety systems — typically combined with on-site generation or battery backup (see our load-shedding resilience guide for how BESS fits into this tier).

Over-specifying redundancy across an entire site is expensive; under-specifying it on the wrong load centre is the failure that actually gets remembered. The classification exercise, done honestly at the planning stage, is what prevents both mistakes.

---

Sequence the Capital Plan Around Lead Times, Not Just Budget Cycles

Capital planning for electrical infrastructure often gets sequenced around when budget is approved rather than around what actually has a long lead time. A more resilient sequence works backwards from the site's required energisation date:

1. Grid connection / NMD application — submit first; it has the longest and least controllable lead time.
2. Civil works and servitudes — substation foundations, cable trenching, switchroom construction; these depend on the site master plan being finalised, not on switchgear being ordered.
3. MV/LV transformers and switchgear — order once the connection voltage and NMD are confirmed, since a late change to either can invalidate equipment already specified.
4. Protection coordination study — commission this in parallel with equipment ordering, not after installation; a coordination study that only starts once switchgear is on site routinely delays commissioning by weeks.
5. Commissioning and energisation — the final step, and the one every earlier step in this sequence exists to protect.

A capital plan built in this order surfaces the true critical path early, when there is still time to act on it, instead of discovering during commissioning that the constraint was a grid application nobody submitted until civils were already finished.

---

Account for Demand Charges When Sizing the Connection

The point-of-connection size isn't only a technical decision — it sets the tariff structure the site pays for the life of the connection. Eskom and municipal industrial tariffs bill a demand charge based on notified maximum demand (or actual peak demand, depending on the tariff), independent of energy consumed. Applying for more NMD than the site will realistically use "just in case" locks in a higher fixed monthly charge for capacity that sits idle; applying for too little forces a costly upgrade application the first time the site trips a demand penalty or a network capacity limit.

This is where the load forecast from the first section earns its keep: NMD should be set against the probable-load figure plus a modest margin, not against the strategic-headroom figure reserved for growth that hasn't been scoped yet. When that unscoped growth becomes real, it triggers its own NMD upgrade application — submitted early, per the lead-time guidance above, rather than assumed away at the outset.

---

Common Planning Mistakes We See

• Sizing infrastructure to the first phase of equipment only, with no reserved servitude or transformer capacity for the expansion everyone already expects. The retrofit that follows costs several times what reserving the space and ducting upfront would have.
• Submitting the grid connection application after switchgear has already been ordered. By the time the NMD approval comes back, equipment specified against an assumed connection voltage or fault level may no longer match what was actually granted.
• Treating redundancy as a single site-wide setting rather than classifying load centres individually — either paying for N+1 everywhere, or discovering during an outage that the one load centre that needed it didn't have it.
• No documented assumptions behind the load forecast. When the forecast is inevitably revisited two years later, there's no way to tell which figures were conservative estimates and which were guesses, so the whole forecast gets redone from scratch instead of updated.
• Protection coordination studies commissioned after equipment installation, turning what should be a design input into a commissioning-phase scramble that delays energisation.

---

Working With GNTech

GNTech supports facility and project teams at the site master-planning stage — load forecasting, single-line development, redundancy classification, and specification of Telergon LV and Pronutec MV equipment once the site plan is confirmed. Bringing us in during planning, rather than only at the procurement stage, gives more room to get the infrastructure decisions right the first time. Contact the GNTech team to discuss planning support for your next site or expansion.