Cyber-physical systems are self-adaptive, changing theirbehavior at run time to adapt to their context. Hence,their simulations must also handle variability at run time.The lack of support for variability in industrialequation-based modeling languages, such as Modelica, causesproblems when simulating self-adaptive systems, e.g., theyonly limitedly support structural variability, and statetransitions are based on if-then-else conditions that cancause conflicts, especially for complex control mechanisms.We present a modeling technique for equation-basedmodels containing variability by implementing concise anddedicated language constructs to express state space andtransitions via contextual modeling. Contextual modelingabstracts the modeled world and allows the definition ofconstraints that reduce the risk of reaching conflictingstates. We demonstrate the feasibility of our approach on acase study, presenting the advantages of our modelingtechnique regarding the definition of state control and thereduction of risk for reaching conflicting states.