A plastic injection molding machine ran its hydraulic oil at 62°C in summer. The OEM spec said 45°C maximum. The oil was thin, the valve spools were sticking, and the machine tripped on low pressure at the accumulator fill. The plant engineer ordered a bigger air-cooled cooler, the biggest that would fit the existing mounting frame. It dropped the oil to 58°C. Not enough. The problem was not the cooler size. It was the heat transfer medium — air versus water — and the plant had picked the wrong one for the duty.
Air-cooled coolers are simple, need no water supply, and fail gracefully. Water-cooled coolers (shell-and-tube or plate) transfer heat far better per unit size. The difference is roughly a factor of 10 in heat transfer coefficient: air at 20-60 W/m²·K against water at 500-2000 W/m²·K. A water cooler the same size as an air cooler moves ten times the heat. The molding machine needed 18 kW of heat rejection. The air cooler that would do that at 35°C ambient is huge. The water cooler that does the same is the size of a shoebox. The plant had a cooling tower on the roof feeding the mold chillers. The water was available. Nobody connected the two.
Doing the heat balance
The first number: how much heat does the hydraulic system generate? For a pump running at 150 bar and 60 L/min, the heat is the difference between input power and useful work. At 150 bar and 60 L/min, the hydraulic power is P = p × Q = 150 × 10⁵ × 0.001 = 15 kW. If the system is 30% efficient overall (typical for a machine with throttling), the heat is 70% of the input — about 10.5 kW, plus the pump losses. A realistic figure for this machine: 12-15 kW of continuous heat at full cycle.
The cooler must reject that heat plus keep the oil at 45°C when ambient is 35°C. That means a 10°C approach. An air cooler with a 10°C approach at 15 kW needs a large core and a big fan. The water cooler with the same approach is small. The plant’s mistake was sizing the air cooler by “biggest that fits the frame” instead of by the heat balance. The oversized air cooler still could not reach 45°C because the approach temperature (oil-to-air) was limited by the 35°C ambient. The oil could never get below about 50°C with a 35°C ambient, no matter how big the cooler. That is the ceiling nobody recognized.
The approach temperature ceiling
Every cooler has an approach limit. For air-cooled, the oil cannot go below about 10-15°C above the ambient air. For water-cooled, the oil cannot go below about 5-8°C above the water inlet temperature. In the molding shop, the cooling tower water came in at 28°C. The water-cooled cooler could hold the oil at 36°C. The air-cooled cooler, even perfect, could hold it at 47°C at best. The machine needed 45°C. The air cooler was physically incapable of the job. No size, no fan speed, no second cooler in series would change that. The limit is thermodynamics, not hardware.
That is the single most useful check in cooler selection: write down the ambient air temperature and the water temperature, add the approach margin, and see if the result is below your oil temperature target. If the ambient is 35°C and you need 45°C oil, air cooling works but barely. If the ambient is 40°C and you need 40°C oil, air cooling is impossible. Water cooling is the only option. This one check would have saved the plant the wasted air cooler purchase.
The water cooler installation
The plant installed a plate-type water cooler rated at 18 kW, fed from the cooling tower return line. The oil dropped to 41°C. The valve spools stopped sticking. The low-pressure trip disappeared. The machine ran 3°C below spec even in the hottest week.
The installation rules that made it work:
- The water flow was set to 1.5x the oil flow. The plate cooler’s catalog curve assumed that ratio.
- A differential pressure valve on the water side regulated flow. The cooling tower water pressure varied with the chillers, and the valve held the water flow constant.
- The oil side was piped so the hot oil enters at the top and exits at the bottom — the natural convection direction, which prevents air pockets in the oil circuit.
- Both sides got isolation valves and drain plugs, so the cooler can be cleaned without draining the whole tank.
One detail: the water supply must be filtered. The cooling tower water carries scale and algae. The plate cooler channels are narrow. Without a strainer, the cooler plugs in three months. The plant added a 300 μm strainer on the water inlet. Quarterly cleaning of the plates became a 20-minute job instead of a day.
When air cooling is still the right answer
Air cooling is not wrong. It is right when the heat load is modest, the ambient is cool, and there is no water supply. A mobile machine, a hydraulic power unit outdoors in a temperate climate, or a low-duty system under 10 kW — air cooling is simpler and cheaper to maintain. The failure mode of the molding machine was using air cooling for a 15 kW continuous load in a 35°C shop. The decision table:
| Condition | Choice |
|---|---|
| Heat load under 10 kW, ambient under 30°C | Air-cooled |
| Heat load over 15 kW, continuous duty | Water-cooled if water exists |
| Oil target within 15°C of ambient | Air-cooled, check approach |
| Oil target within 10°C of ambient | Water-cooled |
| No water supply on site | Air-cooled with a bigger core, accept the temperature |
The plant’s total cost to switch: the plate cooler ($1,200), a strainer ($80), two isolation valves ($140), and a day of piping labor. The wasted air cooler was $2,800 sitting in the stores. The machine has run at 41°C through two summers.
Check the approach temperature before picking the cooler type. If the oil target is within 10°C of the ambient air, air cooling will not get there, no matter the size. Water cooling drops the approach to 5°C and handles ten times the heat per unit area. The molding machine was not under-cooled; it was cooled with the wrong medium.
