Hinge Production (sustainability enriched)

This is a simulation model and simulated event data of a hinge production process and its environmental impacts.

Download:

Short Description #

The simulated process describes the production of hinges. It begins with a steel coil which is split into steel sheets which are heated, formed and coated before being split into the male and female parts. Subsequently, both components are assembled with a steel pin and packed. The process is describes using events and objects

This is an artificial event log according to the OCEL 2.0 Standard defined in Berti et al. simulated using an extended version of CPN-Tools with a connector to climatiq API. The extension to CPN-Tools and the event log simulation are results of the Bachelor Thesis project of Marco Heinisch at the Chair for Process and Data Science at RWTH Aachen University.

Overview #

In this example scenario, hinges are produced in a German facility. The production process is divided into three workstations, all supervised by a single worker.

The production line is started by this employee workstations for workstations every morning at 7:30. The employee pauses work for a lunch break at 12:00 pm and then continues working from 1:00 pm until 3:00 pm. The production process starts with two primary inputs: steel coils and steel pins. The steel pins, also known as steel rods, are each 3.0 cm long and are used to hold the leaves together. The steel coils hold a rolled steel strip that is approximately 0.3 centimeters thick and 3.0 centimeters wide. This steel strip is then processed into hinge leaves through the following steps.

In the first operation of our model, one steel coil at a time is processed at the first workstation. Each coil is continuously cut into steel sheets that are 3 cm in length. As these strips are cut, we assign each resulting steel sheet an identification number. This process continues until there is only an unusable steel remainder of the steel coil, which we do not track further in this process. This remainder needs to be classified as waste in preprocessing. The amount of waste created during the production of rolled steel strips varies depending on the initial length and could be optimized. The cut steel sheets are handled in a line and heat-treated in a gas oven. A different oven technology could reduce impacts. Then, each 3.0 cm steel sheet is rolled on one side to form a barrel for the insertion of hinge pins. This operation is performed using a hydraulic press powered by electricity. In our model, we refer to these produced parts as FormedParts. Finally, in a very energy intense step, plasma coating is conducted using a specialized machine. The FormedParts are automatically collected in a transport carriage. If the carriage reaches full capacity, an alert signals the operator to manually move the batch of FormedParts to the second workstation.

At the second workstation, FormedParts are shaped into alternating male and female hinge leaves (MalePart and FemalePart) using a laser cutting machine. This is achieved by cutting out specific sections from each part. Notably, the method generates more waste compared to an alternative process where the male and female parts are directly cut out from the coil and then shaped further. The metal waste containing the coat is hypothetically not recyclable, which is reflected in a separate impact indicator. The waste created during this operation is not recorded, however, the weight of the workpiece object before and after the material removal is recorded.

In the next step, the worker manually checks both FemalePart and MalePart for quality criteria, which is mass, representing various possible measurable attributes, including geometrics and material properties. If everything is within tolerance, the parts are placed on a moving band to the third workstation. If not, the part represents waste, which again is to be quantified in a preprocessing step. At the next station, an assembling machine uses one MalePart, one FemalePart, and one SteelPin to create a Hinge. The material origin of a hinge or its parts could potentially be discovered. A set of 10 hinges is automatically buffered and placed into a cardboard box by a packaging machine.

Sustainability Data #

The event log is enhanced with sustainability-related attributes for objects and events, as can be seen in the extended meta model below.

The specific-sustainability attributes can be identified by their naming structure as they begin with either i, p or s and include a unit in square brackets at the end of the attribute name structure, i.e., “i_electricity[kWh]”. These additional attribute are classifies into three different categories of data:

• Impact Indicators (“i_”): contain information on all different impacts caused by an event, object, or process. Impact Indicators should be based on some sort of sustainability framework, such as ReCiPe2016 standard. In this event log, all relevant impact indicators considered in reviewed literature on manufacturing assessments are included according to the Climatiq-API specification to show the potential of automated assessment. Indicators further represent LCI-Items according to the LCA standard. However, to both indicate their relevance and a possible inability to get this data from the process directly some of the values read “??” as they have to be estimated using given impact paqrameters.
• Impact Parameters (“p_”): As data availability issues may lead to an inavailability to determine impact indicators directly, impact parameters can be included into the event log to support the estimation of impact indicators. These process-specific parameters include, e.g., material, mass, volume, geometric properties, and operation duration.
• Impact Scores (“s_”): Impact scores describe the overall effect of a system or event on the environment in a standardised manner. Multiple impact scores for different impact categories, such as climate change or toxicity, can be calculated and represented as impact scores. This process involves aggregating and weighting individual impact indicators according to a standardized methodology, such as ReCiPe2016. Climatic API (used for the simulation of the event log) uses values for impact indicators to return impact scores.

Find an overview of the included indicators, parameters, and scores in the Figure below.

Artifacts #

A simulated event log as well as a simulation model and the corresponding thesis project can be found under the resource linked above and in this Git Repository.

Event Log #

An overview of log properties is given below.

Simulation Process #

The result is a simplified hinge production line, modeled as a Colored Petri Net (CPN) in CPN Tools, a tool for editing and creating colored Petri nets. This model is an idealized representation of an exemplary manufacturing process. It is designed to examine and present the use case of SD in OCPM and does not realistically represent a specific manufacturing process in detail.

The model’s design process was structured into four steps. First, the workpiece flow was outlined, identifying the sub-products used in hinge production. Next, control flow elements were incorporated, adding batching and queuing behaviors as well as quality control procedures. Sustainability data relevant to this example were integrated into events and objects following the sOCEL specification. Finally, the simulation of values such as waiting times and weights now includes randomness and deviations.

Model Simulation #

The repository with the CPN Tools model as well as the CPN-Tools extension including the connector to Climatiq API can be found in the following Repositiry: https://github.com/rwth-pads/sOCEL

Further information on the simulated process, the simulation model and the CPN-Tools extension can be found in Marco’s thesis entitled “Integrating Sustainability Data into Event Logs for Object-Centric Process Mining”.

Authors #

Marco Heinisch

Nina Graves (corresponding author)

Contributing #

To contribute, drop us an email! We are happy to receive your feedback.