This plant has several warehouses. The main warehouse’s inside security cameras have been experiencing reliability problems while the outside ones have not. The plant engineer remarked that it’s too bad this isn’t the other way around because the solution would likely involve simply sealing outside cameras against moisture.
The cameras send information to a cloud-based analysis system. The outside cameras are motion detection activated to record comings/goings and any unusual activity, such as a violent act in the parking lot. The footage is archived for later retrieval if needed by police or insurance investigators.
The inside security cameras play a more intense role. The footage from these is actively scanned by AI to identify possible security risks and send an alert to the appropriate party. An archive is also kept, but it’s the real-time nature of the inside camera system that makes these especially important for safeguarding people and property — yet not all of the reporting is real-time. There’s also an availability report that comes out weekly. And for some reason, there are apparent “blank outs” that occur. These cameras, either individually or collectively, are not sending as much information as they should be sending.
The plant does not have a maintenance contract on the camera system, and it was installed by the maintenance department. It seemed to work fine for a few months, but then these blank out reports started showing up.
The plant manager has instructed the plant engineer to get the blank spot problem solved, and the plant engineer has assigned you to troubleshoot and fix this. What’s an approach you might take?
Answer to quiz. Just because the cameras are inside doesn’t mean they aren’t subject to moisture. If you can identify which cameras are blanking out (and those are located near a moisture source, such as a loading dock, which is open to the outside, or process heat that creates thermal cycling), the solution would be to replace those with cameras suitable for an outdoor location.
How are the cameras sending their data to the control system that sends the data to the cloud? If wirelessly, that would most likely be via Wi-Fi. Two strategies here:
- Calculate or measure the data transfer through each router on the system and compare that to the router’s transfer rate. You may need to upgrade one or more routers or just add routers and redistribute the load.
- Carefully analyze the Wi-Fi coverage for weak spots and blind spots, especially those that may be transitory due to non-stationary elements. You can use a smart phone to test Wi-Fi signal strength. A warehouse is full of things that inhibit or block Wi-Fi coverage. If you cannot identify and fix the blockage between a given camera and the router it sends data to, then use a hard-wired method such as Ethernet to connect the camera to the router (this may require replacing the camera with an Ethernet-capable one).
If cabled, you may have any of several problems that attenuate the signal through the cable. If the cameras are Ethernet-connected, use a high-quality Ethernet cable tester to see where signal is being attenuated. It might be at a kink in a cable that was mishandled during installation and thus exceeded its bend radius. Or it might be across a poorly made connection. Or a cable connector that isn’t securely snapped into the port. Do this testing methodically, and ensure you test the entire system.
Another possibility is there are just too many cameras feeding too much data to the controller for the bandwidth you have available to send to the cloud service. It’s a simple matter to ask the cloud service how much data they received over a given period (e.g., 10 specific days) and how much the cameras sent to the controller. If these numbers are not the same, you have inadequate bandwidth and will need to reduce the number of cameras, put some cameras on motion detection, or add a second controller with its own separate transmission means.
Something the plant manager can’t rule out is that someone is periodically disabling cameras in a given area. If the previous troubleshooting didn’t identify any weaknesses, then someone will have to go down this path and it may be someone the plant manager has to bring in from the outside. Keep this in mind while doing your troubleshooting. Look for signs of tampering, such as finger oils or dirt smeared on the camera near the Ethernet port. Is there a ladder or step stool situated near a camera, for no apparent reason? Is there any notable shrinkage of items located within sight of security cameras (you can ask the warehouse manager or the Plant Controller for shrinkage numbers)?
