ACXION banner

How Many Temperature Points Does Your Switchgear Need

The question arrives before the product does. Somebody has decided that the switchgear needs wireless temperature monitoring. The first practical problem is not which device to buy. It is how many sensors to order, where to put them, and what the system should do when one of them reads high. Get those three wrong and the installation reports numbers that nobody acts on.

ACX6600 wireless temperature monitoring receiver with dual temperature display
ACX6600 wireless temperature monitoring receiver with dual temperature display

Scope note: this covers continuous temperature monitoring of bolted joints, busbar connections and cable terminations in low voltage and medium voltage switchgear, using permanently mounted wireless sensors and a local receiver. It is not about infrared inspection windows or periodic thermal imaging. Those remain useful and they answer a slightly different question. ACXION builds the ACX6600 and ACX6500 receivers and the WTC405 and GTC405 sensors described here, so this reflects how we specify them.

Why joints heat up

A bolted joint carries current through the contact area between two conductors. That area is never the full face of the joint. It is the sum of a number of small asperity contacts, and its effective size depends on bolt torque, surface condition, and the pressure that remains after years of thermal cycling.

Every load cycle heats the joint and lets it cool. Each cycle relaxes the stack a little. Aluminium and copper also creep. An oxide layer grows on the contact surface, and the oxide conducts worse than the metal underneath. Resistance rises, which produces more heat, which accelerates the same process. It is a slow failure with a long flat period and then a short steep one. That is exactly the shape periodic inspection is bad at catching.

The temperature rise limits for connections are set out in IEC 62271-1 and in its Chinese equivalent GB/T 11022. Those limits come from type testing under defined conditions. They are not a description of a cabinet in a hot substation in August, which is why an absolute alarm threshold copied straight from the standard is rarely the right answer.

WTC405 and GTC405 wireless temperature sensors strapped on busbar joints for switchgear thermal monitoring

Counting the points

Start from the circuit, not from the cubicle. Every point where current crosses a mechanical connection is a candidate. The candidates are not equally likely to fail.

Cable terminations are the most common failure location. The connection depends on a lug being crimped correctly and then bolted correctly, and both are done by hand. A three phase cable termination is therefore three points if you want all phases. It is worth wanting all phases, for the reason set out below.

Busbar joints between cubicles come next. On a long lineup there are more of them than people expect. Then the fixed and moving contacts inside the withdrawable unit. Those are the hardest to reach and the most valuable to monitor on a critical feeder. Then the disconnector or earthing switch contacts on the same circuit.

On a typical incoming cubicle, with cable terminations at the bottom and busbar connections at the top along with the breaker contacts, the honest count lands somewhere between six and twelve points. That is why the receivers come in 3, 6, 9 and 12 point versions rather than one fixed size. Most installations we supply end up at six or nine points per lineup. The twelve point version is usually reserved for incoming and bus coupler panels, where the consequences are worst.

The reason you want all three phases rather than one is that the useful signal is often differential. A joint running at 70 degrees on all three phases in a 40 degree room is carrying load. A joint running at 70 degrees on one phase while the other two sit at 48 is failing. No absolute threshold set tightly enough to catch the second case will ever be usable in summer.

Choosing the sensor

There are two sensor types in this range. The difference is how they are powered, and that decides where each one can go.

The WTC405 is battery powered. It transmits at 2.4 GHz with an open line range of about 80 metres, and it measures 0 to 110 degrees Celsius. It can be strapped onto a joint during a short outage, and in some cases without an outage at all. Being self powered means it works anywhere. That includes circuits that carry little or no current most of the time, and joints that are not near a conductor you can clamp onto.

The GTC405 is passive. It takes its energy by induction from the conductor it monitors, so it needs an operating current of at least 5 A to run. Where the circuit always carries more than that, the passive sensor is the better choice. There is no battery to replace and no battery life to account for. Where the circuit is lightly loaded or intermittent, it will not work. Specifying it there is a mistake we have had to unwind on more than one order.

Both sensors cover the same 0 to 110 degree range. The receiver itself covers 0 to 120 degrees, with an accuracy of plus or minus 1 degree Celsius.

The receiver, and what it does with the data

The ACX6600 is the panel mounted display. The ACX6500 is the DIN rail version, with the same 3, 6, 9 and 12 point options. Both have two normally open relay outputs for alarms and an RS485 Modbus-RTU port. Device MTBF is rated at no less than 50,000 hours.

Two relay outputs is fewer than it sounds. It forces a decision about what the alarms mean. Our recommendation is to use one for a warning and one for a trip or an urgent call. Base the warning on a differential between phases rather than an absolute temperature. A warning that fires at a fixed value will either be ignored, because it fires every summer, or set so high that it fires too late.

The Modbus port is what makes the data useful beyond the substation door. Trending a joint over months shows the slow climb that absolute thresholds miss. A joint that rises a few degrees each month is more alarming than one that sits at a stable value.

ACX6100 switchgear status indicator with mimic diagram and live indication

If the cubicle also needs switchgear status indication, live indication and local control, the ACX6100 covers that. Its E variant accepts up to twelve wireless temperature points on the same panel face. That avoids a second cut-out on a crowded door. On new build rather than retrofit, it is usually the tidier answer.

Where the sensors go, and what they are not

A sensor measures the temperature of what it is touching. Strapped to the outside of a busbar joint, it reads close to the conductor temperature but not exactly it. The offset depends on how well it is clamped, and on airflow inside the cubicle. For trending and for differential comparisons this does not matter. For arguing with a type test certificate it does.

Position the sensor on the joint body rather than on the bolt head. Keep it clear of a heater or a fan discharge. Record which physical joint each channel maps to. That last step is the one that gets skipped, and it is the reason an alarm ends up being investigated on the wrong cubicle.

Wireless monitoring is also not a substitute for torqueing joints correctly at installation, and it does not fix a joint. It tells you early enough to plan an outage instead of taking one.

What we need to specify a system

We are the manufacturer, so we should say plainly that we sell this equipment. To size a system we need the lineup drawing, the number and type of joints to be monitored, and the load current on each monitored circuit. We also need to know whether the panel is new build or retrofit, whether the receiver goes on a door or on a rail, and how the alarms need to be presented.

Systems in this range are configured to the project drawing and quoted per configuration. Sensor and receiver specifications, mounting formats and channel options are in the power monitoring and protection catalogue on our download centre. Where the panel face needs status indication as well, our notes on switchgear status indication panels cover that choice. Cabinet climate control is a separate problem, described in our article on condensation in switchgear.

Limits

Wireless sensors inside a metal cubicle do not behave the way they do in open air. The 80 metre range figure is an open line figure for a reason. A cubicle that is heavily compartmentalised, or one with a lot of steel between the sensor and the receiver, needs the antenna position thought about before installation rather than after. Where a survey is required for insurance or for a utility, a calibrated portable instrument over a survey period is still the instrument of record.

Ratings in this article come from the current catalogue. Confirm channel count, sensor type and receiver format on the order confirmation. Use the drawing supplied with the equipment for final installation detail.