The demand response(DR)

which is adopted to decrease peak load and operational risk

is generally divided into electricity price based DR and incentive based DR. The focus of this research is to construct a time-of-use load model that contains load control and analyies the impact of load changes on power system reliability by using two kinds of DR measures

that is incentive mode of load control and time-of-use electricity price

power grid reliability

which analyzes the two operation modes of the distribution networks from different perspectives. Furthermore

a new method of calculating the reliability of distribution network under dynamic load by load curves clustering is proposed. By introducing the improved k-means clustering algorithm

the load probability model on different DR load curve is obtained. The reliability indices of distribution network under different load levels are calculated by Monte Carlo method

and the influence of DR on the reliability of distribution network is analyzed. The results of the example show that the demand response measures can improve the system reliability

and the time-of-use electricity price including the load control is more reliable than the single measure.