<p>With the rapid expansion of renewable energy resources, the demand for effective demand flexibility programs (DFPs) has grown to address challenges posed by surplus energy and peak load spikes. To assess the potential for demand response (DR) flexibility, we propose an empirical data-based estimation methodology that utilizes only high-resolution historic energy data. Pattern clustering is performed to identify similar energy usage behaviors due to no prior knowledge of processes associated with industrials. Since data points in clustering have high dimension due to high resolution, we employ a Self-Organization Map (SOM) technique to reduce feature dimension and then K-means algorithm to group similar patterns of energy usage. For each identified cluster, we assess its demand flexibility potential and calculate the total demand flexibility potential for industrials. We conduct a case study on 10 industrial factories, such as two cements, two forges, one paper, three metal, and two steel. As a result of clustering, the proposed method showed an average performance improvement of 9.43%. The case study demonstrates that cement, forge, paper, metal, and steel exhibit a flexibility potential of 2.93%, 16.36%, 2.42%, 8.81%, and 6.47%, respectively.</p>