Because the plant does not yet exist, there is no waste stream to sample directly, so the waste-generation rates have to be built up from process knowledge, analogous industry data and, once construction begins, from real measurements — in that order of increasing reliability.
Review the process flow diagram and mass/water balance. Obtain the plant's process design (unit operations, raw material and water inputs, product output) from the process engineers, and identify every point where water contacts a raw material, product or cleaning agent.
Consult published industry-specific data. Use standard references (Metcalf & Eddy industrial-waste tables, EPA/Environment Canada effluent guideline development documents, industry association data) for waste factors reported per unit of production for the same or a comparable SIC-coded industry.
Benchmark against a similar operating facility. If the parent company or a comparable widget manufacturer operates an existing plant, obtain its actual metered flow and pollutant concentration data as the best available surrogate.
Bench-scale/pilot treatability testing. Where the process chemistry is novel, generate representative wastewater in the laboratory from the specified raw materials and run jar tests/bench reactors to estimate both generation rate and treatability.
Build in a design safety factor. Because pre-construction estimates carry real uncertainty, size the treatment plant with contingency capacity and design for future monitoring-based re-calibration.
Commissioning-phase monitoring. Once the plant is built, install flow meters and a sampling program (composite over the production cycle, capturing diurnal and batch variability) to confirm or correct the design-basis estimate before final plant tuning.
