Lean Manufacturing and Optimizing Operations

SVEZA Group, the world leader in birch plywood production, has been actively introducing new methods to optimize company's operations. The highlight among those is the new system for production scheduling OptiPlan developed by Evolved Analytics Europe and currently being implemented on the first plant in Kostroma. Introduction of the SVEZA Optiplan will allow the company to align its strategic goals to be a 100% customer-oriented business with real-time business operations, optimized logistics, and improved customer account management. Highly specialized mathematical algorithms hidden under user-friendly interfaces of SVEZA OptiPlan are designed to streamline production and balance logistics while optimally using raw materials and minimizing inventory costs. SVEZA anticipates substantial (up to 3-4 times) reduction in the time of order fulfillment as well as noticeable reduction in WIP.

Plywood puzzles

Published with permission of SVEZA

Motivation

The vast majority of consulting businesses (in Business Intelligence) promote lean manufacturing and business optimization strategies, which involve client-oriented production, and reduction of all unnecessary inventory and operational costs. However, it is rarely advised on how exactly to do it.

How to align with client orientation while minimizing costs of operations? How to produce only what clients want, not manufacture anything which has not been ordered, while optimally consuming all available raw materials and minimizing operational changes? How to produce only what we can sell immediately without robust demand forecasts? How to orient on client wishes while satisfying business plan goals and minimizing logistics?

We feel there is a significant gap between knowing what to do – produce more, spend less, plan far ahead and see customer as king, and how to do it (all of the above are competing objectives – I can improve on one while get worse in the other).

We took up a mission to bridge this gap using advanced predictive analytics technology, field expertise and interactive insight-generating workflows and interfaces.

Sveza OptiPlan

Because the stakes are very high and production cycles differ quite a bit among manufacturers and sectors, custom data systems are the optimal solution for easy adoption, seamless integration, and straightforward propagation of the lean manufacturing strategy through an enterprise.

SVEZA OptiPlan is an example of a proprietary OptiPlan framework of Evolved Analytics customized for the largest birch plywood manufacturer SVEZA Group. (Our proprietary predictive analytics and business process optimization system OptiPlan has been heavily customized for our manufacturing customer – SVEZA Group).

Optimally matching demand with intermediate raw material production and consumption

SVEZA OptiPlan is designed to optimally schedule client orders of different types of birch plywood into production in order to dramatically reduce the time between order acceptance and order delivery while optimizing raw material consumption, minimizing idling time of expensive equipment, increasing the overall plywood output, maximizing the number of finished orders for each week in the planning horizon while balancing daily shipment volumes.

Fine-grain production optimization with equipment use optimization and shipping load balancing

Example A:

The smallest plant of SVEZA produces almost 500 different types of birch plywood in format 5x5 feet with the output of over 10000 m3 per month. Every product is characterized by the plywood grade (4 grades), thickness of a sheet and the finishing type, and packaging. Each sheet of plywood is composed of several sheets of birch veneers glued together with a special epoxy cured for a certain time in a hot press. Each SKU (stock keeping unit) requires a particular formulation in terms of intermediate raw materials – birch veneers in our case. I.e. for each SKU there exist a unique recipe of making it. E.g. to make a top plywood of grade B/B in thickness 4 mm we need two sheets of veneers of grade B with thickness 1.15mm to put on either side of the plywood sheet, and 3 sheets of veneers in grade “internal layer”' in thickness 1.50mm to put in-between and glue with a special epoxy.

The intermediate raw material is produced on the same plant from wooden logs. Logs get conditioned for 24 hours, cut to size, then peeled into very thin “wooden carpets”' (veneer), then cut and dried to get veneers to compose plywood. Veneer sheets have a fixed length and width, but can have several thicknesses. Dry veneers get sorted into 5 grades (B, BB, CP, C, internal layer - the whitest most uniform sheets are top grade B, with grades going down with increasing number of knots or color changes). How veneers get distributed into different grades depends on their thickness%mdash; thin veneers have a higher proportion of top grade, than thick ones.

In summary – logs are transformed into intermediate raw material – veneers in 10 grades, and distributions depend on the time of the year (and the quality of wood logs).

The limitations are that unless specially packed and stored, veneers have a limited life time and must be turned into plywood rather quickly. This requires that all veneers produced in a week (despite unequal accumulation over the grades) are consumed evenly and as quickly as possible. Optimally matching production with intermediate veneer levels is a major challenge on the road to an even, uninterruptible production process.

Example B:

For example, to make 3 packages of plywood B/B in thickness 4 we need to make 195 sheets (2.7m3) of this plywood. This requires twice as many sheets of veneers in grade B1.15 mm (1.15m3 in veneers) and three times as many sheets in grade internal Layer 1.50mm (2.25m3 in veneers). Because grade B only accounts for 2% of the peeled birch, to get the required veneers we need to peel a minimum of 66 m3 of wood. This means that after making the 2.7m3 of plywood we will be left with more than 63 cubic meters of intermediate veneers, which need to be turned into products as soon as possible! The question is

What to produce from these raw materials to minimize inventory and maximize speed to market?

SVEZA Optiplan takes into account capacities of all production stages, the season and priorities of client orders to compose an optimal production schedule for the entire planning horizon (2 to 6 weeks) with exact forecasts of production evolution of each SKU in each client order, shipment dates, pressing sequences, forecast of veneer inventory. Using state-of-the-art mathematical algorithms under the hood, OptiPlan explores hundreds of thousands of scheduling scenarios (for thousands of variable options and production constraints) just under 5 minutes, and provides the user with a best schedule that meets output volume plan requirements, maximizes the number of happy clients for each day of the planning horizon, and points out the bottle-necks in the production line. Knowing which part of the production line to improve or upgrade as well as how to quantify the investment provides unparalleled opportunities for the manufacturer to further optimize the production process.

For manufacturers with a variable demand we developed a unique system to balance the production load while minimizing inventory costs for products that are not being sold. Evolved Analytics' OptiStock is an intelligent data system which suggests optimal lists of orders to produce, such that the available raw material is maximally used and only highly popular and well sold products are made, profitability is maximized, and the risks for unnecessary inventory costs are cut down.

Example C:

When there are not enough client orders for the planning horizon the SVEZA OptiPlan will analyze the sales statistics, use state-ot-the-art prediction techniques to forecast the demand for the next X days and will synthesize additional orders, which only consist of popular production to be sold in the coming X days (even if no orders are yet taken from the clients).

If the pool of client orders is too large for the planning horizon and a limited volumes of raw materials are available, OptiStock will select subsets of client orders which optimally consume raw materials, while maximizing the number of fully finished orders and hence, happy clients.