Abstract: To explore dimension reduction methods and patterns of complex safety systems, based on safety system data field, a research on dimension reduction model of complex safety system was carried out. Firstly from the perspective of data representing safety characteristics and states, a concept of data field was introduced, and connotations of data field in safety space were described from both macro and micro levels. Meanwhile, four basic elements were summarized. Then based on all above, the effects that data field brought in safety space were analyzed from three levels to explain how accidents occurred from small incidents. Finally, a dimension reduction model of complex safety system based on data field was built by taking the dimension reduction, capacity reduction and change reduction process as main line. The results show that data field can describe and explain the dimension reduction process and rules of complex safety system effectively, and self-organizing clustering, capacity reduction in the dimension, dimension reduction between dimensions, and that change reduction in or between dimensions are four key processes of safety system dimension reduction.

Key words: complex safety system, data field, effect, dimension reduction, theoretical model