This master thesis aims at developing the methodology to form dynamic PTDF families, for the implementation into the market model. The main procedures to create a dynamic PTDF family are developed and presented with a study case (See Chapter 4). By fitting data from the market model into the data mining tools, typical scenarios of the time intervals are selected. Furthermore, by dividing subareas according to nodal PTDF values, the resulted estimation power flow is getting closer to the real flow, as depicted in section 4.4 for the study case. In addition, the outliers of the initial input data are located and more typical scenarios are selected from this outlier data group. The effectiveness of the dynamic PTDF family is demonstrated by comparing the estimated power flow using the dynamic PTDF matrices, a fixed PTDF matrix and the real power flow calculated in PSSE, given the same economic dispatch run by Powrsym4 simulation without PTDF logic. The results reveal that with the dynamic PTDF family adopted, the power flow estimation error is significantly reduced in comparison with using a fixed PTDF matrix. In the study case, the peak value of the estimation error is reduced by 67%. The average error of the estimated power flow reduced from 16.9% with one fixed PTDF matrix to 3.7% when we adopt the PTDF family. In the last part of the thesis, the implementation of dynamic PTDF matrix in the market model is simulated with a Matlab economic dispatch model for one hour. The influence of incorporating dynamic PTDF matrix into the UC-ED decision is investigated. The UC-ED results and PSSE power flows are compared for cases without PTDF logic, with simple PTDF logic (only one fixed PTDF matrix) and with dynamic PTDF matrix. At last, an implementation scheme of the dynamic PTDF family for the UC-ED decision making is proposed.