Simio Simulation - Serial Manufacturing Line

Summary

This video guides viewers through building a simulation model for a serial manufacturing line in Simio. The process involves understanding parts' arrival rates, service times for machining, grinding, and packaging, and handling defective parts. It demonstrates how to set up servers, sinks, and connectors in Simio, along with defining the necessary parameters for inter-arrival times and processing durations. The video also covers running experiments with 100 replications to measure key metrics like utilization and system performance, which are analyzed using SMORE plots to assess efficiency and identify areas for improvement in the manufacturing process.

Takeaways

• 😀 Parts arrive at a rate of 60 per hour, with three phases in the manufacturing process: machining, grinding, and packaging.
• 😀 Machining takes approximately 45 seconds per part, grinding takes 55 seconds, and packaging takes 63 seconds.
• 😀 10% of parts will require re-machining after machining, and 10% of ground parts will be deemed bad and disposed of.
• 😀 The inter-arrival time of parts is modeled using an exponential distribution, with a mean of one minute between arrivals.
• 😀 The simulation model for the manufacturing process includes a source, three servers (machining, grinding, and packaging), and two sinks (good and bad parts).
• 😀 The service times for machining, grinding, and packaging are modeled using exponential distributions with means of 45, 55, and 63 seconds, respectively.
• 😀 Selection weights are set for the transitions between processes, with 90% of parts progressing from grinding to packaging and 10% being sent to the bad parts sink.
• 😀 90% of parts from machining are considered good and move on to grinding, while 10% are sent back for re-machining.
• 😀 The simulation experiment is run with 100 replications to measure performance and gather data on utilization, time in system, and number in system.
• 😀 SMORE plots (Simio Measure of Risk and Error plots) are used to visualize and analyze the results, showing mean, median, confidence intervals, and percentiles for key metrics.

Q & A

What is the arrival rate of parts in the manufacturing line?

The parts arrive at a rate of 60 parts per hour.

What are the service times for each phase of the manufacturing process?

The machining phase requires 45 seconds per part, the grinding phase requires 55 seconds per part, and the packaging phase requires 63 seconds per part.

What percentage of parts require re-machining after the machining phase?

10% of the machine parts need to be re-machined.

How are the bad parts managed in the grinding phase?

10% of the ground parts are deemed bad and are disposed of.

What distribution is used to model inter-arrival times and service times?

Exponential distributions are used to model both the inter-arrival times and the service times for machining, grinding, and packaging.

How is the inter-arrival time of parts calculated?

The inter-arrival time is calculated as 1 divided by the arrival rate (60 parts per hour), resulting in an average inter-arrival time of 1 minute.

What are the roles of the connectors in the Simio model?

Connectors in the Simio model link the source to the servers (machine, grind, and package) and link the servers to the sinks (Good for good parts, Bad for defective parts). There is also a connector that loops parts back to the machine for re-machining.

How is the selection weight used in the grinding phase?

The selection weight in the grinding phase is set to 90% for parts that proceed to packaging, and 10% for parts that are deemed defective and go to the Bad sink.

What kind of experiment is conducted in the simulation, and what responses are measured?

An experiment with 100 replications is conducted, measuring responses such as machine, grind, and package utilization (MUtil, GUtil, PUtil), as well as time in system (TIS) and number in system (NIS).

What do SMORE plots represent in the analysis?

SMORE plots show various statistical measures including the mean, median, confidence intervals (for both lower and upper percentiles), and the minimum and maximum values of the simulation responses.