PLM System and Workday Planning Development for a Food Holding

Client

The largest agricultural holding in CIS, specializing in turkey, duck, lamb, and other livestock-based food products. It holds 50% of the CIS turkey products market.

The holding includes enterprises covering the full production cycle: from feed production and livestock reproduction to finished goods manufacturing, including semi-finished products and meat delicacies.

The client produces goods both under its own brands and under retailers' private labels.

At the time KT.Team worked on the project, the semi-finished product line included about 900 products.

The Problem

Despite the production volume, many of the client's key processes were not digitized or automated. For example, product specifications were stored as simple text or spreadsheet documents in a decentralized way. The matrix of technological transitions between products, meaning the lists of actions and the conveyor switch-over time for producing a new product, existed in spreadsheet form, and workday planning became an analytical task each time, consuming a lot of time and resources.

These and other aspects of the work had to be automated as much as possible, reducing the cost of routine operations.

Result 1: specifications are stored centrally and can be changed easily when needed

The first step in the project was automating the process of working with specifications and process cards.

The process works roughly like this: when an idea for a new product appears, methodologists handle its development. They prepare the product specification, the list of source ingredients, and the production methodology. These documents are then sent to technologists for approval, and they can confirm or adjust the processes.

Previously, all documents, such as specifications, process cards, and so on, were stored separately, access control was difficult, and changes, such as an adjusted spice mix in the filling or changes in packaging, were not transparent to all process participants.

After moving the process to Pimcore, specifications and process cards are linked to a specific product. We implemented role-based access inside Pimcore so that only methodologists and technologists can edit documents, while approved specifications and process cards are passed to production. Production users cannot accidentally change key documents.

Result 2: routine production processes run faster thanks to parallelized sub-processes

Previously, all actions within the same process according to the process cards were largely sequential.

Together with the client, we reviewed each existing process and configured tools for analyzing roles and actions. Using these tools, the client identified tasks and sub-processes that could be performed in parallel without losing quality or efficiency. For example, when switching the smart conveyor between products, the knives had to be washed and the container for finished goods had to be cleaned. Since several people handle the switch, they can perform these tasks at the same time without violating safety rules. This significantly reduced time losses during production operations.

Result 3: preparation of the production workday plan is automated and takes fixed and variable factors into account

Workday planning in the client's production depends on several factors:

Almost all of the client's products are manufactured on three smart conveyor lines that can be configured either for basic poultry cutting or for preparing breaded cutlets.

The client had a so-called transition matrix, a table with 900 rows and 900 columns that recorded the switch-over time between products and the set of actions required for it. Each switch can take from several dozen seconds to ten minutes, depending on what workers need to do: replace the knives, add curing mix or seasonings to the relevant containers, wash the containers for finished products, and so on. The switch-over time is not obvious: for example, switching only between primary cutting and drumstick production takes 6 minutes, while the full sequence of switches from cutting to large fillet production to small fillet production to drumstick production takes 9 minutes.

Each time, creating the workday plan was an analytical task that took several person-hours or even person-days.

We digitized both the transition matrix and the vehicle arrival schedule from the client's partners using directories, a dedicated Pimcore feature. Now, to plan the workday, the engineer only needs to enter the list of products scheduled for today. The system automatically calculates valid options based on the existing constraints, which reduced the time needed to prepare the workday plan by several times.

• the production plan for today: this plan is adjusted daily based on sales trends;
• the shipping plan for retailers' private-label products, which also depends on the sales trends of those products;
• the vehicle arrival schedule from partners: each partner has strictly agreed shipping windows. For example, Magnit can pick up its order from 6:45 to 7:15, and by then the full assortment agreed with Magnit must already be in the loading area. If the order is not ready in time, the client faces fines, and private-label products that have already been packed would have to be unpacked and discarded.