Case Study

IMPROVE 4.0 has the capability to directly acquire data from the field regarding downtime, scrap numbers, and production flow speed. These three Key Performance Indicators (KPIs), when combined, yield a percentage value known as OEE (Overall Equipment Effectiveness).

This data is gathered through Programmable Logic Controllers (PLCs), sensors, and direct input from operators using Human-Machine Interface (HMI) devices.

Real-time data on production efficiency is made available through the Andon system. Andon boards located on each production line provide the following information for the benefit of all stakeholders:

• Takt time
• Expected pieces, actual production, first-pass yield
• Remaining pieces for the batch
• Number of scraps
• Processing time, expected downtime, unexpected downtime
• Current OEE value

Aggregated real-time data for all production lines is also accessible through dedicated interfaces designed for managerial roles.

Aggregated data

The efficiency data is then aggregated to be accessible over daily, weekly, monthly, or annual timeframes.

Efficiency Monitoring (OEE)

In this interface, Pareto diagrams of downtime and scrap causes are also available. These causes have been acquired either from PLCs or directly from operators through HMI devices.

Statistical Process Control (SPC) is a fundamental part of the data acquisition process. This process provides an in-depth understanding of the production process. This knowledge can be leveraged to monitor and control variable values in real-time, allowing for immediate intervention in the event of an out-of-control situation (OOC).

Data acquisition

Data from the field is acquired and allocated in memory areas of the PLC (Programmable Logic Controller). IMPROVE then collects data from the PLC, creating a snapshot of the current state of variables. IMPROVE will then compare these values with the previously configured values, generating real-time feedback on the process or triggering other actions (such as sending emails or scheduling maintenance).

Before data acquisition, each PLC is configured with the respective datasets of interest and the memory areas where the acquired data will reside.

Dataset

For each variable, a description is provided, a domain is defined, measurement units are specified, and reference values or tolerance thresholds are established. During data acquisition, real values are compared with the configured values, determining whether an “out-of-control” condition exists.

Thanks to this system, it’s possible to monitor and control any variable related to various food or pharmaceutical products.

The concept of being “out of control” is highly relevant.

We know very well that just one variable not being within the expected values can have consequences on the taste of food or the effectiveness of a medicine. This underscores the importance of this phase in the process.

The configured variable values can take different forms. For example, it can be represented as a limit beyond which an “Out Of Control” (OOC) condition occurs, or it can consist of a threshold or range. In the latter case, different zones are identified and represented by various colors: the red color represents the OOC area, while green and yellow represent areas with acceptable values.

Configuration of variables

Other values can result from the connection with other datasets acquired through the specification of an aggregation criterion or through combinations with other variables using mathematical models. Additionally, it is possible to import deep learning models to perform functions that may be too complex for a simple mathematical model.

.

Data analysis

Once all the collectible data is gathered, the next phase is data analysis. For this reason, the data is organized in tabular forms as well as graphs that are derived directly from statistical process control.

Control chart

Through multivariable analysis, it is possible to further enhance our ability to monitor the trends of various product variables.

Multivariable analysis

The various operators or responsible individuals also have the opportunity to report anomalies and their respective causes through an HMI (Human-Machine Interface) device.

Monitor HMI

In the food industry, digitalizing the weighing phase of products is of crucial importance.

Many companies have not yet fully digitalized this phase, opting to save the weighing results as files in order to maintain a historical record of these values, as required by law.

What IMPROVE 4.0 does is read the weight value and save it, ensuring compliance with Law 690.

Law 690/78 guarantees the end user (consumer) that the declared weight in prepackaged products is in accordance with the law, using approved scales for use in transactions with third parties.

According to this law, the weight value can deviate from the expected value within certain thresholds:

• Normal Weight
• T1 and T2 as mandated by Law 690
• -T1 and -T2 as mandated by Law 690
• Plus any other limits specified by the customer

Once this data is collected, IMPROVE 4.0 generates real-time reports that can be filtered by period, product, production line, or department. The available data includes weight and various types of analyses.

Weighing Values

The Machine Ledger 4.0 originates from the World Class Manufacturing methodology, designed to manage all types of maintenance within a single environment, ranging from reactive to preventive, autonomous to professional.

Next has transformed the WCM tool, originally a complex Excel file, into a web-based tool suitable for all types of organizations, not limited to the WCM environment.

Additionally, a shared platform (machineledger.it) has been created, facilitating collaboration between machine suppliers and end-users. By using this platform, suppliers can upload the machine ledger and machine Bill of Materials (BOM) for their customers.

Machine Ledger

The Machine Ledger is a calendar that consolidates all types of maintenances, including both scheduled and unscheduled ones.

It interfaces with the calendars of operators and maintenance personnel responsible for individual interventions. The system tracks events that need to be performed, those that have been postponed, and those that have been completed.

Thanks to this solution, maintenance can be managed at higher levels, leading to benefits in terms of waste reduction. This transition shifts from a Time-Based Maintenance (TBM) approach, where components are replaced at regular intervals, to a Hit-Based Maintenance (HBM) approach. In HBM, time is no longer the sole factor in component replacement; instead, actual usage is considered.

A higher level of precision in determining the right time to replace a component is achieved with Conditional Based Maintenance (CBM). The end-of-life of a component is determined by the combination of one or more variables, whose thresholds are monitored through Statistical Process Control (SPC) techniques.

Conditional Based Maintenance

All the information collected through the Machine Ledger allows for the acquisition of important Key Performance Indicators (KPIs) such as MTTR (Mean Time To Repair) and MTBF (Mean Time Before Failure).

Finally, it is possible to identify the costs associated with both reactive and preventive maintenance and appreciate the trend of values over time.

Preventive Maintenance Cost

Reactive Maintenance Cost

The implementation of a Factory 4.0 project involves having production orders automatically transmitted from the management system (SAP) to the machines, without the need for human intervention.

This interconnection is facilitated by IMPROVE 4.0, which is responsible for transferring orders from SAP to the machines using protocols such as Siemens S7, Modbus, and OPC-UA.

The communication between IMPROVE 4.0 and SAP is bidirectional, as IMPROVE 4.0 also handles sending production data back to the management system.