Prepare an Industrial Chiller Load Survey Before Sizing Equipment

A maintenance manager hands you a nameplate rating and asks for a matching replacement. The old chiller struggles on hot afternoons, production is losing time, and the plant wants a number, not a conversation. But a nameplate describes what someone selected years ago, not what the plant asks of the system today. Size a new unit to that old number and you inherit every load that got added since the original installation. The conversation that prevents this mistake is a load survey, and it has to happen before anyone proposes a size.

Why This Conversation Goes Wrong

Most chiller replacement requests start backward. Someone in maintenance or purchasing already has a number in mind, usually the tonnage stamped on the failing unit, and they want a quote that matches it. The pressure is real. Production is down or running hot, and every day of delay costs money. So the easiest path looks like agreeing to the number and moving fast.

The trouble is the number rarely reflects the current load. Plants add equipment over time. A laser cutting cell shows up, two molding presses join a line, a paint booth gets a cooling loop it never had before. None of that appears on the old nameplate. Size off the label instead of the current process and you either undersize the new unit and repeat the failure, or oversize it and the plant pays for capacity it does not use, often with worse part load efficiency as a result.

The requester is not being difficult here, they are being efficient with what they already know. They see a chiller failing on hot days and a label with a number on it. Asking them to slow down and gather flow data, temperature readings, and utility information can sound like stalling instead of helping. Your job is to reframe the delay as the thing that prevents a second failure, not as friction added on top of an urgent problem.

Prepare The Ground Before You Meet

Before you ask a single question in the room, decide what a complete survey needs to contain. Write the plan down. It should name the measurement points, the instruments needed to capture them, who has access to take each reading, and who confirms the value once it is recorded.

At a minimum, get supply and return fluid temperatures during the affected run itself, not just at startup or on a mild day. Find out which equipment operates during that run, because a chiller that struggles on a hot afternoon may be struggling because several machines draw load at once, not because ambient temperature rose a few degrees. Get water quality and flow data if the plant uses cooling tower water or a shared loop, since fouling and flow restriction reduce effective capacity even when the equipment itself is sound. And confirm the electrical capacity available, because a larger chiller sometimes runs into a service limit before it runs into a space limit.

Bring this plan into the meeting instead of building it live. When you already know what you need, you sound organized rather than skeptical, and the requester experiences the conversation as planning rather than interrogation.

What To Say When The Request Comes In

Picture a plant manager named Priya and a technical sales engineer named Daniel meeting about a failing chiller.

Priya says, "Our chiller cannot hold temperature on hot afternoons. I have the nameplate right here. Can you just quote a replacement at the same tonnage?"

Daniel says, "The nameplate tells me what this unit was rated for when it went in. It does not tell me what the plant asks of it now. Can you walk me through what changed on the floor since then, and what supply and return temperatures you see during a bad afternoon?"

Priya says a laser cutting cell and two molding presses were added to the line after the original chiller was installed, and nobody revisited the cooling plan when that happened. She admits she has no current temperature logs, only complaints from the floor.

Daniel says, "That is useful, and it points right at the gap we need to close. Let us set up a short survey. Facilities can pull water logs and confirm electrical capacity, production can flag a representative hot run, and I would like your maintenance technician there too, since they usually know where someone added a hose or a bypass that never made it onto a drawing. During that window we record supply and return temperatures, flow, ambient conditions, and which machines are running. Once we have that, I can bring your engineering team a sizing recommendation they can check against real data instead of a label."

Daniel never argues that the nameplate is worthless. He treats it as one data point and asks for the rest. He also names specific people and specific records instead of asking Priya to gather everything herself.

Mistakes That Undermine The Survey

The most common mistake is accepting the nameplate number and quoting from it anyway, because the requester sounds confident and pushing back feels like it risks the relationship. That confidence usually comes from urgency, not from data, and a wrong quote costs far more trust than a short delay for a proper survey.

A second mistake is asking for data without saying who should provide it. Tell Priya to go get water logs, electrical drawings, and temperature readings herself, and the survey stalls, because none of that sits with one person. Name the roles instead. Facilities usually owns water and electrical data. Production knows the schedule and can flag a representative run. Maintenance often knows about undocumented changes, like a temporary hose or bypass line that never reached a drawing.

A third mistake is treating the survey as a formality on the way to a number you already suspect is right. Walk in assuming the old tonnage is close enough and you will read the data to match that assumption instead of letting it tell you something different. Go in willing to be surprised by what the loads actually show.

A fourth mistake is naming a size out loud in the same meeting where you agree on the survey plan. Write the plan, run the survey, and let engineering review the results before anyone commits to a number. Say a tonnage too early and you anchor the requester's expectations, and walking that number back later is harder than never saying it.

Practice The Conversation Before You Run It

This conversation is easy to describe and hard to run well live, because the pull toward agreeing to a fast number is strong and the requester's frustration is real. Practice it the way you would practice any conversation with a predictable failure mode, by rehearsing the version where the other person pushes back.

Take a real equipment request from your files, one that arrived with only a capacity number attached, and rehearse asking what process created that number, what conditions surrounded it, and who can supply the underlying record. Say the questions out loud. Repeat them until they sound curious instead of corrective, because a defensive tone pushes the requester to dig in on the number instead of opening up about what actually changed on the floor.

A manager can score this by checking whether you asked for site facts before offering any conclusion, whether you named specific people for specific pieces of data, and whether you avoided naming a size before the survey was complete. DealSpeak runs this kind of rehearsal against an AI counterpart that plays the impatient requester, so you can hear how the conversation drifts when you rush it and fix that before you stand in front of the real plant manager. The Manufacturing practice conversations hub has more scenarios like this one if you want to work through other sizing and service conversations from the same industry, including how to Assess Injection Molding Screw Service With Process Evidence and how to Set a Kiln Refractory Service Scope From Inspection Facts.

A load survey takes longer than reading a label off an old unit, and that is exactly why it works. Run the survey first, let engineering size the equipment against real data, and the plant gets a chiller that matches what the floor actually demands instead of a repeat of the same failure at a higher price.