Semi-Markov processes with semi-regenerative states for the availability analysis of chemical process plants with storage units

Olga Fink, Enrico ZioProceedings of the Institution of Mechanical Engineers Part O Journal of Risk and Reliability 2013 · DOI

TL;DR

This paper introduces a novel analytical method for evaluating the availability of complex chemical production chains that include storage units. By extending Semi-Markov processes to handle "semi-regenerative states" representing storage units, the authors provide a more accurate way to balance design-to-capacity principles with high production availability.

What Problem Does This Paper Address?

The chemical industry increasingly operates large, interconnected plants, where a failure in one unit can cascade through the entire network. While technical equipment is a primary concern for availability, storage units play a crucial role in bridging production interruptions. However, traditional availability analysis methods often struggle to accurately incorporate the buffering effect of these storage units, especially when dealing with complex interdependencies and the desire to minimize over-capacity (the "design to capacity" principle). This paper tackles the challenge of developing an analytical framework that can effectively model and quantify the impact of storage units on overall production chain availability, allowing for better design decisions.

Methods

The core of the paper's contribution lies in extending Semi-Markov processes to incorporate storage units. Standard Semi-Markov processes are regenerative at each state transition, meaning they "forget" their past. Storage units, however, introduce "semi-regenerative states" where the process's future depends not just on the current state but also on the remaining time within that state (e.g., how long a tank has been filling or emptying).

To handle these semi-regenerative states, the authors propose several techniques:

  • Assigning distributions for remaining holding times: Instead of assuming immediate regeneration, they assign probability distributions to the time left in a semi-regenerative state.
  • Determining equivalent transition rates: They adapt methods to calculate constant transition rates that approximate the behavior of the Semi-Markov process, making it more manageable.
  • State space aggregation: To reduce computational complexity, similar states are grouped together.

This approach allows for the integration of storage unit characteristics, such as their capacity and filling/emptying dynamics, into the availability analysis.

Datasets

Not discussed in the paper. The paper uses hypothetical data for its case studies, with real failure and repair data being alienated for information protection.

Key Insights and Ideas

  • Storage units are critical for availability: The paper strongly emphasizes that storage units are not just passive buffers but active components that significantly influence the overall availability of a production chain.
  • Semi-regenerative states are essential for accuracy: Ignoring the "memory" of storage units (by treating them as simple regenerative states) can lead to significant inaccuracies in availability calculations.
  • Balancing design-to-capacity and availability: The proposed method provides a quantitative tool to find the optimal balance between minimizing over-capacity and ensuring high production availability by evaluating the impact of storage unit size and other design parameters.
  • Decoupling effect is quantifiable: The paper demonstrates how the buffering capacity of storage units can be quantified in terms of the mean time between production unavailabilities, showing a clear relationship between tank volume and system resilience.

Key Results

The paper presents results from two case studies:

  • Simple system (two units with one storage tank): This case study demonstrates that with a storage tank volume of approximately 5000 m³, the units become effectively decoupled, and the availability is primarily determined by the downstream unit's failures.
  • Cx production chain (five units and three storage units):
    • The analysis shows that the production chain is unavailable for only about 0.6% of the time, with an average restoration time of 23 hours.
  • Storage units effectively buffer repair times, with degraded states having holding times between 4 and 43 hours, while unit repair times range from 24 to 72 hours.
  • Planned maintenance activities, when coordinated with storage unit regulation, can reduce the time with reduced production output to 5% of the total time. Without such coordination, this figure could be threefold.

Limitations and Future Work

  • Manual state space management: The current approach requires manual management of the state space, which can become complex for larger systems. The authors suggest automating this process by creating predefined components with their own state spaces.
  • Load profiles not considered: The current application does not explicitly consider varying load profiles of individual plants, which would further increase the state space. However, the authors note that their approach can be extended to include these.
  • Phase-type distribution approach not used: The paper mentions that approximating Semi-Markov processes using phase-type distributions is an alternative that could be explored in future research.
  • Assumptions: The analysis relies on certain assumptions, such as exponentially distributed failure and repair times (though the method is flexible enough to handle other distributions) and a fixed filling level for storage tanks. These assumptions could be relaxed in future work.

Summary updated

Abstract

Design to capacity is an engineering principle that is increasingly applied in chemical industry, among others owing to increasing plant sizes and associated investments. This principle aims to reduce over-capacity, over-sized buffers and excessive redundancy. Concurrently, a high level of availability is targeted over the entire production chain. The consequences of unavailability of highly interconnected chemical process plants can be significant because a technical disruption in one plant is able to spread over the entire production network. In chemical process plants not only technical equipment determines the availability, but also storage units, which are able to bridge times of planned or unplanned interruptions of production. To find a balance between the principle of design to capacity and high production availability, the influence of different design parameters, such as capacity of production units, redundancy concept and the size of storage units, have to be evaluated and integrated in the design process. In this article, we present an analytical method for availability evaluation based on extending Semi-Markov processes integrating storage units and multiple production lines. Semi-regenerative states are used to capture the characteristics of storage units, and an approach is proposed in this work to assign distributions for the remaining holding times in these states. The proposed modelling and analysis are demonstrated on two case studies.
BibTeX
@article{fink2013semi,
  author = {Olga Fink and Enrico Zio},
  title = {{Semi-Markov processes with semi-regenerative states for the availability analysis of chemical process plants with storage units}},
  journal = {Proceedings of the Institution of Mechanical Engineers Part O Journal of Risk and Reliability},
  year = {2013},
  doi = {10.1177/1748006x13480765},
}
Figure