A logical data model is used to abstract a 'conceptual data model' by incorporating normalization rules to manage data integrity and relationships. It serves as an abstraction layer that defines the relationships between different data elements, entities, and attributes, focusing on the business concepts and rules that govern the data. This model helps in organizing data to reduce redundancy and improve data integrity, ensuring that the data structure is optimized for both business understanding and database design.