The Raspberry Pi 4 installed in the outside enclosure is mounted in a box on one of the internal DIN Rails. On extreme weather days, the air temperature in the area can reach nearly 50°C (122°F). Coupled with the afternoon direct sunlight, this has the potential to serious increase the temperature in the enclosure and hence the temperature of the air in which the Raspberry Pi 4 sits.
Although the Raspberry Pi is equipped with heat sinks on the significant heat-generating chips, this is not expected to be sufficient during heat waves. The Pi has one more trick up its sleeve in that it has an internal temperature sensor and will throttle back the CPU as it starts to overheat. While this will help protect the PI, it does mean that the software running on the Pi must be designed to cope with a throttled back CPU.
The top of the enclosure is also vented to help hot air escape from inside the enclosure, but again, it is unlikely that this will be sufficient to keep the electronics from overheating. A small available fan that could be attached to the case was tried, but the air flow was essentially insignificant for this purpose and the fan had a draw of 150mA from the 5V supply. One of the reasons that this fan had such poor air flow is that the area of the central hub holding the motor was much greater than the area available to the blades. The blades on a larger fan are larger proportionally since the central doesn't get larger at the same rate.
The image below shows that initial installation of the fan shown in the image above. There was not temperature control at this stage, so the fans ran simultaneously and constantly. The horizontal axis is the time in UTC, meaning that 12am UTC corresponds to 10:30am Local Time. The left of the graph shows the CPU temperature overnight. When the sun rises and begins to strike the (uninsulated) rear wall of the cabinet, the CPU temperature begins to rise.
At the point of the discontinuity, the system is powered down and the Raspberry Pi case is removed and the base swapped for the base holding the fans. After the fans are wired to the +5V supply, the system is restarted. The reason that the temperature has dropped so much is that the system has been powered off and the cabinet open to the cool morning air while the modifications were made.
After the enclosure stabilised, the CPU temperature had a relatively small riuse until about local noon when it began to slowly cool. The front side of the enclosure (the door side) has a thick layer of polystyrene foam on the inside wall and does get some protection from the radiant heat.
Around 5pm local time, the power was removed from the fans and the rise in CPU temperature is immediately obvious. After it peaks just over 60°C, the temperature continues the cooling trend that is expected as the sun goes towards the horizon.