Injection moulding sets its monthly peak before anything is moulded. Why a grid-tied plant cannot be relied on to cut the demand charge, which loads it genuinely matches, and what a 260 KWp plant sits against.
The highest demand a plastics factory records in a month is often set on a Saturday morning, before a single component has been moulded. Every barrel and every hot runner zone comes up to temperature at once after the Friday shutdown, the compressors catch up, the chillers start from a warm building, and the demand meter records a peak that the plant will then pay for across the whole billing period.
That single fact does more to shape the economics of a rooftop plant on a plastics or packaging site than anything on the module datasheet, because it is the part of the bill a solar plant is least able to touch.
An injection moulding machine draws in bursts. Clamp close, injection, hold, cooling, screw recovery, clamp open, eject, and around again. On a machine with a fixed-displacement pump the motor runs through the whole cycle and the peak is smoothed out. On an accumulator machine the injection peak is sharp. On an all-electric machine the peaks are sharper still, with regeneration on servo deceleration putting energy back.
With thirty machines cycling out of phase the aggregate at the incomer smooths into something close to a steady load. With four large machines it does not, and when their injection strokes happen to coincide the plant sets a new maximum demand for no productive reason at all. That is worth knowing before anyone blames the tariff.
Around the moulding machines sits a second layer of load that behaves quite differently, and it is the layer that determines whether solar is worth building.
| Load | When it runs | How well solar matches it |
|---|---|---|
| Moulding machine drives | Whenever the line runs, in short peaks each cycle | Partly. The energy matches; the peaks do not. |
| Barrel and hot runner heaters | Heavy at start-up, then cycling to hold temperature | Badly at start-up, well once running |
| Process chillers and cooling towers | All day, and hardest on hot bright afternoons | Very well. The best match on the site. |
| Air compressors | Continuously while the plant runs | Very well |
| Material dryers and dehumidifiers | Near constant while hoppers are loaded | Very well |
| Granulators | Intermittent and operator driven | Neutral |
| Lighting, packing and offices | Shift hours | Well during the day shift only |
Read that table as a targeting exercise. A rooftop plant on a plastics factory is not really offsetting the moulding machines. It is offsetting the chillers, the compressors and the dryers, and those three run in daylight, run harder when it is hot and bright, and never spike.
Industrial tariffs in Bangladesh bill on maximum demand as well as on units consumed. Factory managers reasonably ask whether a solar plant reduces the demand charge. The honest answer has two halves, and the second half is the one that gets left out of proposals.
A grid-tied plant reduces energy reliably. Every unit generated is a unit not bought, and over a year that is a predictable number. It reduces billed maximum demand only by coincidence. The demand meter records the highest average demand over a short integrating window, and it does that all month. For the solar plant to reduce that figure, it has to be generating strongly at the exact moment the peak occurs, on every occasion the peak might occur. A cloud passing over the array removes most of its output in seconds. Build the business case on demand-charge savings from a grid-tied plant and you are betting the return on the weather at one particular minute.
What does reduce demand charges on a moulding site, in the order most factories should try them:
That list is not an argument against solar. It is an argument for selling solar on the thing it actually does, which is displacing a large, steady, daytime energy consumption at the retail rate.
A moulding hall is hot, and the roof above it is hotter. Module output falls as cell temperature rises, so a coplanar array lying flat against an uninsulated sheet roof over a hot hall gives up yield every sunny hour of its life. Mount on rails that leave a real air gap under the modules so that the underside can shed heat, and resist the temptation to reduce that gap to save on rail.
The inverter room deserves the same attention. String inverters derate above their rated ambient temperature. An inverter room on the sunny wall of a moulding hall will reach its derating threshold at around midday, which is exactly when the array is producing most. Put the room on a shaded elevation, ventilate it properly or air condition it, and check the manufacturer's derating curve against a realistic room temperature rather than against an ambient figure taken from a weather table.
Roof structure on packaging and plastics sheds is usually modern light steel with wide-span portal frames, which is good news: large contiguous roof planes, few obstructions and clean access. Check purlin section and spacing anyway, and check what is already hanging from the structure, because moulding halls accumulate ducting, cable trays and evaporative coolers over the years.
Servo-driven machines, variable frequency drives on chillers and compressors, and inverter-driven auxiliaries all put harmonic current onto the busbar. Most plastics factories also have a capacitor bank for power factor correction, and a good number of those banks are untuned.
Connecting a fleet of solar inverters to a board that already carries significant harmonic distortion and an untuned capacitor bank can excite a resonance between the bank and the supply impedance. The symptoms are failed capacitors, tripping contactors and overheating in the bank, and they are almost never attributed to the solar plant because they appear weeks later.
Vvon engineered and commissioned 260 KWp at Innovation Plastic Cans. A plant of that size on a packaging manufacturer is not aiming to run the factory. It is sized to sit under the site's daytime consumption so that generation is absorbed on site rather than exported, which is where the value is under net metering.
The question to ask of any proposal at this scale is simple and it is rarely asked: on a bright day in April at one in the afternoon, is the factory's own consumption higher than the array's output? If the answer is yes with margin, the plant is sized sensibly. If the answer is only yes when every machine happens to be running, the plant will export more than the model assumed, and the return will land lower than the spreadsheet said.
Getting that answer takes a logger on the incomer for a week, and any EPC contractor worth signing with will ask for one before quoting a capacity.