It is normal and acceptable for an overload condition to exist on a power circuit. Examples are motor starting currents and initial charging of power supply capacitors. Some standards have tests specifically for this situation. IEC 950 requires that the steady state current shall not exceed the rated current by more than 10% under normal load. In North America, for some types of equipment such as electric discharge lighting (fluorescent, metal arc), computers, and most continuous loads, the rated current may not exceed 80% of the branch circuit (plug) rating.
You understood correctly that it is the overheating which is the critical point of concern. That is what overcurrent protection is designed to control. >From a practical design point, your biggest problem may be the avoidance of tripping overcurrent devices, but careful measurement of conductor and connector temperatures is important. You mention several minutes of overload per hour, but are not clear about the duty cycle. Several minutes at 30 amps is not going to work on a 20 amp circuit. If these overloads are short and well distributed, you may have no problems. It may be necessary to specify special overcurrent protection such as time delay fuses or HACR rated circuit breakers. Start by consulting the fuse and circuit breaker manufacturers for their performance curves and allow a sufficient margin to avoid nuisance tripping. Connector quality and tight connections may be a significant issue for you since it appears you will see significant temperature cycling in you connections. Loose screws or contacts will quickly lead to oxidized surfaces, more heating and burnout of the connection. Some startup or surge situations can be helped by a series resistor which is shorted out by a time delayed relay. If you need to go to a larger connector, the twist lock series is a less expensive solution in North America, but the IEC 309 is the only common solution for the rest of the world. Bob Johnson
