Measuring phosphorus in wastewater using a self-organizing RBF neural network

A system for monitoring a value of effluent total phosphorus (TP) of a wastewater facility, the system comprising:

• a molybdenum-blue colorimetric analyzer placed in an influent tank of the wastewater facility and configured to measure the value of influent TP, the wastewater facility comprising the influent tank, an anaerobic tank, an aerobic tank, a second setting tank, and an effluent tank;
• an ORP electrode sensor placed in the anaerobic tank as well as the effluent tank and configured to measure a value of redox potential (ORP) in the anaerobic tank and the effluent tank;
• a temperature sensor placed in the effluent tank and configured to measure a temperature in the effluent tank;
• a pH sensor placed in the in the effluent tank and configured to measure pH in the effluent tank;
• a Teflon-coated-film based sensor placed in the aerobic tank and configured to measure a value of dissolved oxygen (DO) in the aerobic tank;
• a COD sensor placed in the effluent tank and configured to measure a value of chemical oxygen demand (COD) in the effluent tank;
• a TN sensor placed in the effluent tank and configured to a value of total nutrient (TN) in the effluent tank;
• a TSS sensor placed in the aerobic ad tank and configured to measure a value of total suspended solids (TSS) in the aerobic ad tank;
• and
• a computing device comprising:
• one or more processors, and
• memory to maintain a plurality of components executable by the one or more processors, the plurality of components comprising:
• a parameter receiver configured to receive real-time information of variable parameters consisting essentially of:
• the value of ORP in the anaerobic tank and the effluent tank,
• the temperature in the effluent tank,
• the pH of the effluent tank,
• the value of DO in the aerobic tank,
• the value of COD in the effluent tank,
• the value of TN in the effluent tank, and
• the value of TSS in the aerobic ad tank,
• a neural network handler configured to:
• initialize a self-organizing radial basis function (SORBF) neural network,
• train the SORBF neural network based on pre-measured training data comprising the variable parameters, and
• predict, in a real-time manner, a value of current effluent TP based on the received real-time information of variable parameters.