Neural net implementation of steam properties on FPGA

Real time applications like model predictive control, monitoring and data reconciliation of power plants industrial processes employ nonlinear mathematical models require thermodynamic properties their derivatives working fluids. Applications super heater temperature control based on energy balance real reconciliation, an efficient a compact method for simultaneous estimation properties, partial suitable implementation in field-programmable gate array (FPGA). However, the complex formulations these prohibit direct implementations FPGAs. Single artificial neural network (ANN) architecture is used to replace entire code higher level languages, running into few thousand lines. FPGA range presented. Large arguments sigmoid function are factored product integer fractional part which represented using series approximation with five terms only integers look up table (LUT). This ensures optimum storage computational burden above applications. The ANN implemented IEEE 754 floating point synthesis Xilinx ISE design suite Verilog HDL. results presented typical pressure versus saturation temperature.