ModelEnterprise case: Optimal scheduling of unit operations

Optimization-based online control of a vacuum filtration plant

"The application has resulted in a 25% decrease in the consumption of expensive precoat material."

Supply chain integration schematic – click to enlarge

Fig. 1 – Executive system

This project was undertaken for one of the largest privately owned companies in the world, which operates primarily in the food and food ingredients industry.

The project involved development of an optimal scheduling system for a vacuum filtration plant in which 3 semi-continuous filters produce filtrate for processing by a semi-continuous downstream unit.

The process

The filters remove waste solid material to produce a clear filtrate. Filters are precoated in order to ensure suitable filtering characteristics from the start.

During filtration a scraper blade progressively removes the precoat material along with deposited material; the filter cycle completes when there is no remaining precoat material.

The operation of each filter is characterised by:

cycle times
filtrate flow rate
total filtrate/ cycle.

The downstream filtrate requirements over time are imposed by the cycle of the downstream unit; any excess of filtrate produced at any point in time is recirculated back to the feed stream of the vacuum filters.

An important issue in the plant operation is the high cost of the precoat material, so minimising the total number of cycles, while keeping the filtrate production in the desired levels, becomes the major target.

Objectives

The project had two stages, each with its own objectives:

reduce precoat consumption by reducing the number of cycles, thus reducing recirculation flow and optimising the plant scheduling. This had to be achived while ensuring that operation remained responsive to downstream filtrate requirements.
provide an automated way of delivering and implementing the optimal schedule of the vacuum filtration plant during operation.

PSE approach

An automation system was built to collect data from the plant, optimise and implement the derived schedule (Fig. 1). This comprised:

a pre-processing module. This collects the raw operation data from the plant via an OPC client connection and translates it into input data to the optimisation model.
a scheduler module, a short-term planning model in ModelEnterprise that optimises the plant operation given current demands and constraints.
a post-processing module, which converts the optimisation results into actual orders to the operators and set-point profile to the operation variables.

Benefits

Implementation of the module within the plant automation system has resulted in:

reduction of the precoat consumption by 25%
elimination of missed filtrate orders due to poor scheduling of the filtration plant
more automated and overall more reliable operation of the plant.