THE SCENARIO
A homeowner flushes the toilet, then some time later finds hot water gushing out from the bottom of the tank in the back of the toilet.
She calls for help, which fortunately arrives quickly.
The water leak is too hot for either person to safely reach the toilet water shutoff valve at the wall. So he shuts off the water going into the hot water tank, then can shut off the toilet valve as only cold water is coming out at that time.
What happened? Why would hot water be gushing out of a toilet that is only plumbed to cold water?

FACTS
- Weather had been unusually cold for this area, for a few days, thus there was risk of loss of water pressure into the house from freezing of lines that are not buried as deep as they might be in colder climates, or of failure of pressure regulators in the city distribution system.
- The only potential connections between hot and cold water systems in the house are:
* Cold water input to water heating tank.
* Faucets in kitchen and bathroom sink (single handle type, others were dual handle thus have separate valves for cold and hot water, though a common spout thus some crossflow potential if both are open).
* Possibly clothes washing machine input solenoid valving. (The responder turned both cold and hot off at the wall.)
- The water heater felt very hot at exposed plumbing.
- Opening the pressure relief valve on top of the water heater produced very hot water vapour. (Colloquially called "steam".)
- Cold water flow out of kitchen and bathroom faucets seemed normal.
- Water was dripping on the furnace, having entered the heat outlet grill in the bathroom and ran along the air distribution duct). (Another complication to restoring normal function of the house - has the furnace been damaged? The responder turned it off at the circuit breaker panel.)
- The night before the owner had noticed air spurts out of the kitchen faucet, and some sand in the bottom of the sink. (Which would be consistent with problems with city water supply.)
- Dirt on the floor of the bathroom below the toilet looked and felt consistent with sediment in the hot water tank.
- The commonly used flexible input line from wall shutoff valve to toilet tank, a gray plastic material, was melted at the valve end, which resulted in the leak.
- The rubbery flapper valve inside the toilet tank was warped, so once the water input was restored the tank did not fill and water ran into the toilet bowl.
- Foam insulation on hot water lines was partly melted for some distance from the tank.
- The over-pressure valve vented hot vapour when manually actuated.
- Checking the heating element indicated a short of one side to ground, removal showed it was split open, exposing the electrical circuit to the water.

Additional help concluded that the heating element housing had split, exposing the element which was shorted to the tank through water, thus to the protective ground of the electrical system. The enabled current was boiling water, regardless of thermostat action, with enough pressure to overcome the cold water pressure into the house. Flushing the toilet allowed that hot water to go through plumbing into the toilet tank, melting the standard plastic tube that connects the toilet water input valve to the toilet tank. (City water was not a factor, but an example of a "red herring" (potentially misleading factor, one to be considered in looking at the complete set of facts, in this case that look rejected it.)

The mishap raises questions of what could be done to prevent it, including:
- periodic checks or replacement of water heater thermostat and pressure relief valve.
- a backflow prevention valve on the cold input to the water heater tank.
- can the furnace be protected against water from above? (The configuration of the ducts and furnace is typical, open grills in the floor.)
- ensuring plastic materials in residential cold water plumbing can withstand boiling water. (There are of course different types of plastic. In this case the gray plastic input line was replaced with a better line having stainless braided flex covering. An option of a flow-break valve in that line is being reviewed (the maker of the better line offers a simple valve in the large end that is claimed to close when water flow rate is excessive). Does use of plastic depend on certain features like backflow preventer? (The traditional toilet inlet line is a bendable metal material, braided stainless is an option but has plastic lining so would not fully contain water if overheated. (It is also easier to install than semi-ridig lines.)
- other measures (for example, the house already has a CO detector near the main bedroom, and smoke alarms including one near the furnace).

Each requires evaluation of risk and feasibility.
(For example:
- the warped flush flapper valve does not result in damage just waste of water, because there is an overflow tube in the tank.
- the flow-break valve won't prevent a leak from backflow if the plastic lining of the better line melts, as it is at the top of the line. The stainless braid covering does increase resistance of the line to high temperature by preventing bulging that would lead to burst, and should reduce rate of leaking. (The flow break valve will only help in some cases of failure downstream such as failure of the fitting it connects to at the toilet tank or of plumbing inside the toilet tank that results in high flow. Failure of the standard toilet tank's water level control to shut off would not result in high flow, and would be mitigated by the overflow tube in the tank. I don't know about the new accumulator flush systems.)
- whereas water getting into the furnace is potentially damaging and risky.)

I am impressed by the features in the replacement water tank and installation, any one of three of them would have prevented the mishap.

(At the same time I am impressed by the longevity of the old tank itself, albeit made of bronze material that would cost more than an entire new tank today. It could have been kept except for lack of safety features desired today - and would have continued to be intact beyond the life of the new tank. Fortunately scrap value of the old tank is high due to the material.)

Analysis of failure modes is standard practice today in many fields. I expect that specific failure modes are considered in building codes, which for example today require a vacuum relief valve, certainly there is no warning in stores about use of the plastic line with old water heaters. This is not a thorough engineering analysis, only an example of considerations, presented to educate on considerations in designing and maintaining.