Genset skid base design: flatness, welds and DFM

A skid base is judged in the first hour after the set is commissioned, not on the day it leaves the shop. If the four engine pads are 1.5 mm out of plane, the anti-vibration mounts load unevenly, the flexible coupling runs out of alignment, and the set vibrates at a level no mount selection can fix. The OEM then spends a site visit shimming a base that was signed off against a drawing with no flatness callout on it.

The numbers that decide it are small. Two thousandths of a metre across a pad, five tenths across the whole set, and a weld sequence that puts heat into the middle of the frame last. This guide covers genset skid base design as a fabrication problem: what the base carries, which tolerances belong on the drawing, how the weld sequence protects flatness, and where a fabricated base stops being the right answer.

Base frames for canopies and acoustic enclosures are covered across the genset and canopy hub, and the companion article on welded genset base frames deals with the welding route in more depth.

What loads does a genset skid base carry?

Four load cases, and only one of them is the static weight. A 1,000 kVA set with its radiator and fuel is typically 3,500–4,500 kg, spread over four to six anti-vibration mounts, so each mount carries 600–1,100 kg static. That is the easy number.

The other three are harder. Engine torque reaction during a load step puts a rocking moment into the base in the order of the engine's rated torque multiplied by the mount height above the frame neutral axis. Transport over Indian roads adds vertical accelerations of 2–3 g on a low-bed trailer, which is why a base that is stiff enough on the shop floor can still crack at a cross-member weld after 800 km. Lifting is the fourth: four-point lift with a spreader, and each lug designed for twice its share of the mass.

Stiffness matters more than strength. A base that deflects 2 mm between cross-members under the alternator mass will hold alignment cold and lose it hot, because the frame and the engine expand at different rates and the deflection path changes. Cross-members under the alternator feet and a closed box section along each side rail are the usual way to get the section modulus without adding plate thickness everywhere.

Genset skid base design tolerances worth putting on the drawing

The table below is the working set for a fabricated base in the 3–8 mm plate range. The middle column is common practice; the third column is what we hold at Ganesh.

The coplanarity row is the one that gets argued about. A drawing that says "mounting faces flat" without a value is not a specification, and a drawing that asks for 0.1 mm across a 2 m welded frame is asking for a machined base rather than a fabricated one. We will tell you which side of that line a requirement falls on before quoting, because a base that needs machining after welding is a different supply chain.

Engine pads, alternator pads and shaft alignment

Shaft alignment is the reason pad flatness is specified. The engine and alternator are aligned by shimming under the mounts, and every shim pack is a stack of flat plates that only works if the surface under it is flat. Typical practice allows shim stacks up to 3–5 mm with a maximum of three shims, because each interface adds compliance.

Coplanarity across the set is the stricter requirement, and it is measured from a single datum. Put a datum on the frame and dimension every pad from it in the same view. A base dimensioned from three different references will be inspected three ways and accepted three times.

Cooling and service access belong in the same review. Leave 100 mm of clear space under the base for airflow and for cleaning a fuel spill, and keep the drain plug at the lowest corner of the tray. A tray that drains uphill holds 20 litres of diesel against a seam weld for years.

Does the base need a drip tray, a fuel tank or both?

A skid base can be three different things and the drawing has to pick one. A plain base carries the set. A drip tray base has a welded shallow pan under the engine to catch spills and leaks. A tank base integrates the day tank into the frame, which saves floor space and adds a pressure vessel to the weld schedule.

A tank base needs internal baffles to stop fuel surge, a filler with a lockable cap, a breather routed away from hot surfaces, a level gauge accessible from the canopy door, and a leak test recorded before the set is mounted. It also needs the welded seams to be accessible for repair, which is why the external seam is usually continuous and the internal baffles are stitch welded.

The tray version is cheaper and most common. Two engineering details decide whether it works: a continuous external seam with a proper drain boss rather than a drilled hole with a nut welded over it, and a slope of at least 1:200 to the drain corner. Test at 0.3 bar with soap solution across every seam, or fill with water for 24 hours and inspect the underside. A base that passes either test on the first attempt usually indicates the welder had room to work.

Weld sequence: how a base stays flat

Flatness is made by the sequence, not corrected afterwards. Four rules cover most of it.

Weld from the centre outward. Start at the middle cross-member and work towards the ends in a back-step pattern, so the shrinkage from one run is opposed by the next rather than accumulating at a corner. Weld the two side rails alternately rather than completing one side first.

Do the pads last. Tack the frame, check diagonal, then fit and weld the pads in a fixture that holds the pad plane. The pads carry the tolerance, so they should not be welded while the frame is free to move. On a base where a cross-member runs under the engine, weld the cross-member to the rails first, let it cool, then attach the pads.

Keep the heat balanced from side to side. A 3 mm fillet on 3 mm plate puts enough heat into the section that a single continuous run along one rail will bow the frame 2–3 mm in the middle. Stitch welds of 40–50 mm with a 100 mm gap, welded alternately, hold the same strength with a fraction of the distortion. Where the drawing shows a continuous seam for fuel or water tightness, the tray is a separate sub-assembly welded to the frame after the frame is square, not part of the frame weld.

Check the diagonal twice. Once after the structural welds, and once after coating. The oven cycle releases residual stress and a base can move 1–2 mm on a 2 m diagonal as it cools from 180–200 °C. A frame that is square before coating and out of square after it will be blamed on the paint shop, and the real cause is the weld.

Where distortion still shows up

The pattern we see most often is a base that is flat on the pads but twisted overall. Twist comes from unequal weld volume on the two rails and from the base being lifted from one end before it has cooled. Both are process problems rather than design problems, and both are caught by the twist check in the table.

The second pattern is a tray seam that cracks at the corner where it meets a cross-member. A tray welded rigidly to a frame that flexes in transport will crack at the first stress raiser. Isolate the tray with a small gap at the cross-member or use a continuous weld with a generous corner radius, and test it after the base has been moved once, not only at the bench.

Transport, jacking and the details that crack

Two details account for most transit damage on a fabricated base. The first is bracing. A base designed for a set sitting on mounts has almost no resistance to the horizontal loads of a trailer braking, and the cross-members are loaded in a direction they were not designed for. A pair of diagonal braces bolted to the lifting lugs and removed at site costs little, and it prevents the parallelogram that shows up later as a cracked pad weld.

The second is the jacking point. Service engineers lift one end of a set to change a mount, and if the drawing does not name a jacking pad they will use the tray edge. Put four jacking pads on the drawing with a stated load each, and keep them out of the masking schedule so they stay bare steel and do not crush a paint film under the jack.

What Ganesh actually runs for base frames

Base frames at Ganesh go through one route: plate and channel blanks cut on the 12 kW fibre laser with a 40 ft × 16 ft bed, sections formed on the 200 T, 5-axis CNC press brake with a 4,450 mm bed, structural welding in fixtures built in our own tool room on one of three CLOOS robotic MIG cells or at one of the 30-plus manual stations, then shot blasting in the 16 ft × 16 ft chamber, seven-tank pretreatment, powder coating in the 18 ft × 14 ft booth and oven cure.

The long bed is what makes a 3–4 m base practical as one piece. Side rails are cut and formed in single lengths, so there is no splice in the member that carries the alternator. Mounting pads are laser cut with the mounting holes in the same setup as the pad outline, which holds hole-to-edge position without a second operation. Welded assemblies that have to arrive flat are checked on a bed with a dial indicator before they leave, and the inspection record carries the pad flatness and diagonal figures.

We build base frames and canopies as one coated kit for Mahindra Powerol, which means the frame that carries the set and the frame that carries the canopy are measured against the same datum. The genset parts programme is where that supply is described, and the welding side of it sits under welded assemblies. Distortion control in general is covered in controlling weld distortion in frames, and heavier plate work in welding for heavy gauge material handling parts.

What the canopy drawing has to settle

• Set mass and centre of gravity, in kg and mm from a named datum, so stiffness and lug loads can be checked.
• Mount positions and mount type, with the pad size and the hole pattern, dimensioned from one datum.
• Pad flatness and coplanarity, with the measurement plane named.
• Lifting arrangement: number of points, spreader requirement, and design load per lug.
• Tray or tank: capacity in litres, drain position, baffle layout and the leak test pressure.
• Weld symbols on the structural joints, and a note where the seam must be continuous for sealing.
• Finish schedule per face: blast, pretreatment, primer, powder, and which faces are masked.
• Handling and packing: how the base is lifted in transit, and whether it ships with the canopy.

The canopy RFQ checklist covers most of these fields as a form, and it flags when a base is being quoted without a mass or a mount drawing.

When a fabricated skid base is the wrong route

Fabrication loses to casting or machining above certain sizes and below certain volumes. A small set in high volume is usually cheaper on a pressed and welded base with dedicated tooling, and a large set above roughly 2,000 kVA often needs a rolled-beam base with machined pad faces, because the flatness callout stops being achievable by welding alone.

We do not machine pad faces, do not perform modal or finite element analysis on your behalf, and do not stress relieve fabricated bases in-house. We also do not design the anti-vibration mount selection; that belongs with the mount supplier and the OEM's dynamics engineer. What we can do is weld to a sequence that holds the flatness in the table, measure the result, and tell you when a drawing asks for something the process cannot deliver. That conversation is worth having before the first base is cut, because pad flatness is the cheapest tolerance to specify and the most expensive one to discover on site.