Diesel engine trial is an important step to determine the performance and carry out acceptance of diesel engine. The main testing platform is the dynamometer system

in which the mechanical energy generated during the trial process is converted into electrical energy through the dynamometer system

and then to heat energy through a load box

resulting a huge waste of energy. Currently

the event with low probability but high impact due to the extreme disasters has seriously affected safe operation of power distribution network

while the access of the flexible distributed resources is an effective measure to enhance the resilience of the distribution network and mitigate the impact of extreme disasters. Therefore

it is considered in this paper to feed back the electrical energy from the dynamometer system into the distribution network as a flexible resource to support its operation. Furthermore

the impact of the access of the diesel engine dynamometer system on the resilience of the distribution network is studied so to search the optimal dispatching strategy of the dynamometer system under the disaster scenario. Firstly

the output characteristics of the dynamometer system is modeled. Then

the key resilience data is selected to quantify the impact of the grid connection of the dynamometer system on the resilience of the distribution network. The optimal scheduling model of the diesel engine dynamometer system is then set up with the objective of improving resilience. This issue contains multiple decision agents

which are solved by multi-agent reinforcement learning method. Finally

the example analysis is conducted to simulate the impact of the access of the dynamometer system as a distributed energy resource into the distribution network under different capacities and grid connection points of the network and its optimal scheduling strategy is solved. The analysis of example shows that the access of the diesel engine dynamometer system contributes to enhancing the resilience of the distribution network

thereby reducing the load loss rate of the distribution network under disaster scenario.