3.06 Reliability Engineering

In our increasingly digitized and networked environment, the expectations for excellence in asset management are ever growing. While an abundance of maintenance and sensor data have become available, companies must develop the proper application of the data in their maintenance strategies. In this presentation, we’ll discuss the potential of your operational and maintenance data in the context of asset management, and explore different machine learning (ML) algorithms and how they may be leveraged to unleash hidden patterns in your asset management strategies. We’ll introduce some foundational topics required for ML, such as the taxonomy and data preparation steps critical to all ML approaches, the probability and statistics supporting ML, and how the evaluation of the quality of our models. C-MORE has actively applied machine ML methods to interesting real-world problems, such as the categorization of power generation units according to reliability characteristics, and anomaly detection in linear assets to optimize required maintenance actions. We’ll share a few of our case studies so participants can experience how ML methods can be used in maintenance, reliability and operations.

With many people at home either due to the closure of their operations, self-isolation protocols, or as a proactive measure to reduce non-essential staff on site, some might question how these individuals can be productive, particularly when assets are not operating. However, if employees have access to their CMMS/EAM/ERP data systems from home, here are some value-added activities that employees and employers should consider undertaking given the time they now have. Note these are in no particular order as priorities would be context-specific, and specific procedures are omitted for this same reason. 

This article is to address the difference in conditional probability of failure patterns, and the impact on how best to maintain assets based upon those differences.

This Project was established to review all facets of Maintenance within the St Lawrence Seaway Management Corporation (SLSMC) with a goal to improve productivity, maintaining a positive impact on maintenance staff moral and provide the same or increased equipment reliability. Maintenance Programs were reviewed for all major assets and analyzed using subject matter experts leveraging the FMECA (Failure Mode, Effects & Criticality Analysis) tool to determine areas of vulnerability within the assets ability to perform at the designed operational level Maintenance Processes were analyzed using some of the Lean Six Sigma and Work Measurement tools with focus on the six (6) steps of Work Management Cycle (Identify, Plan, Schedule, Assign, Execute and Learn) to get a better understanding of the problem areas and generate solutions to this issue backed by actual results. Work Organization main focus was to improve Supervisory awareness and availability in providing support to trades employees and conducting regular field audits to ensure accuracy and quality of task execution. Investigations and work process flow analysis are also planned for individual Trade Shops and Warehouse Facility Layouts to improve work space planning and component/part inventories. Change Management focus was on Vision Mapping, Stakeholder Analysis, Communication Planning and transition coordination of all improvements and changes that will affect the entire organization during the progression of each stage of the project. The findings of the project to date showed that there were a lot of excess maintenance tasks being performed on managed assets. The estimated labour times for task completion, travel and delay inefficiencies of work tasks being performed were excessive and daily performed tasks contained value and non-value activities over all process steps of the Work Management Cycle. All findings discovered and work that continuous to be performed at each stage of this project confirms that there is a lot of variability, inefficiencies and opportunities for improvements within all facets of the Maintenance within the Organization.

As the influence of the asset management approach continues to expand within Nova Scotia Power, we need a structured approach to ensure we continue to seek opportunities to optimize maintenance strategies. In a new installation, techniques such as failure modes and effects analysis (FMEA) and reliability centred maintenance (RCM) can be used to develop an optimized maintenance strategy from the start, in a top-down approach. However, the vast majority of Nova Scotia Power’s equipment was in place long before the asset management office—and, therefore, the asset management approach—existed. The result of that is a collection of value-added, but developed after-the-fact maintenance strategies. Each maintenance strategy has components of operator surveillance (rounds), testing, predictive pattern recognition (also known as advanced pattern recognition, APR), predictive maintenance (condition-based monitoring and risk-based inspections), online monitoring, and preventative maintenance. While efforts had been made to “baseline” the equipment processes when maintenance strategies were developed (i.e., “clean out” existing activities), the organic growth of the approach and the distributed nature of assets and personnel have made this difficult to maintain. Therefore, we needed an approach to optimize existing maintenance strategies, without recreating them. Nova Scotia Power has therefore undertaken an effort known as maintenance strategy optimization, and has made this activity a core accountability for the asset management team, which recognizes the need to seek continuous improvement (vs. a one-time exercise). With a focus on digitization wherever appropriate, Nova Scotia Power has asked a number of questions to streamline, standardize, and optimize its maintenance strategies. Is there opportunity to reduce PM frequency? Is there opportunity to collect more information such that we can strengthen our APR models? Can our in-house standards be revalidated to sustainably reduce operating and maintenance costs? Nova Scotia Power is answering yes to these questions, and more, and pursuing opportunities to optimize its maintenance strategies—from the bottom up! 

Enbridge Liquids Pipelines is at the midpoint of a multi-year journey to advance the M&R program for its electrical and mechanical assets. This presentation will share successes and challenges while aiming to apply M&R best practices to a midstream pipeline operator. Our initiative will improve M&R processes and culture, with a focus on improving multiple work streams. These include the planning, scheduling, and execution of preventative and predictive maintenance, information management, spare parts, and CMMS use.

Companies are continually faced with competing priorities and limited resources, and, as a result, it's critical to drive value from the business decisions they make. With more focus on asset management principles, these decisions must be justified through demonstrated and quantified value to the business. With complex systems of equipment, such as a pipeline, determining the value of a project or proposed improvement can often be difficult to estimate—especially when considering an asset's lifespan. In such cases, a more systematic, data-driven approach may be required to predict value. RAM models are one such tool that can be used to achieve this. In this case study, we'll look at how RAM models have been developed and used at Enbridge Pipelines to identify and quantify the risks to throughput volume. The models include key equipment and operational events, along with specific throughput impacts and simulated long-range demand forecasts in order to prioritize the risks and focus resources toward capturing the best opportunities. When specific improvement opportunities (e.g., additional equipment sparing) are identified, alternate cases of the model can be created and the results compared to the base case to support decisions and justify projects. We'll explore some of the challenges we encountered during development and highlight examples of how this tool is being used to quantify value, as well as the approach taken to disseminate the information within the organization.

When entering a maintenance record into a CMMS, there's often a place where the operator can enter free-form comments. These comments may contain valuable information about the health of the equipment, any maintenance activities that were undertaken, and plans or recommendations for the future. The flexibility of the comments is attractive to operators, as a precise description of the observations can be recorded. However, using the comments information in data analysis usually requires some codifying of the comments, which is time-consuming and results in a loss of nuanced information. A machine learning approach to using comments data has been applied to predict the health of hydroelectric generating units. By embedding comments into a matrix to generate a "bag of words," and applying neural networks on the vocabulary, comments can be used to assess the current state of the asset and predict its next state. In this presentation, we'll discuss three machine learning algorithms in a way that's accessible and relevant to M&R practitioners: a classification method, a clustering method, and a neural network method. Each method will be partnered with direct applications to real-life maintenance problems, including lessons learned and potential uses in other contexts.

Metrics, best practices, more than 40 key elements to implement, challenges, and opportunities all combine to make a successful implementation difficult. Where do you start, and how do you know how to work on what matters? Once you understand how it’s all related, you can focus on the vital few to leverage the maximum ROI. This presentation will clarify the importance of culture and employee engagement, along with other key plant floor performance indicators that will be clarified with data. We'll look at the current state of R&M; what’s working and what's not; survival skills for the next decade; impacts of connected technologies (edge computing, big data, machine learning, AI, 3D printing, augmented reality); the importance of getting your data ready for what's coming next; and relationships between R&M and safety, people engagement, quality, throughput/uptime, and cost.

Reliability Centered Maintenance – Reengineered, provides an optimized approach to a well established and highly successful method used for determining failure management policies for physical assets. It makes the original method that was developed to enhance flight safety, far more useful in a broad range of industries where asset criticality ranges from high to low. RCM-R® is focused on the science of failures and what must be done to enable long term sustainably reliable operations. If used correctly, RCM-R® is the first step in delivering fewer breakdowns, more productive capacity, lower costs, safer operations and improved environmental performance. Maintenance has a huge impact on most businesses whether its presence is felt or not. RCM-R® ensures that the right work is done to guarantee there are as few nasty surprises as possible that can harm the business in any way. RCM-R® addresses the shortfalls of RCM that have inhibited its broad acceptance in industry. Little new work has been done in the field of RCM since the 1990’s, yet demand for such a method, better adapted to industrial applications is higher than ever and growing. Demographics and ever more complex systems are driving a need to be more efficient in our use of skilled maintenance resources while ensuring first time success – greater effectiveness is needed. RCM-R® was developed to leverage on RCM’s original success at delivering that effectiveness while addressing the concerns of the industrial market. RCM-R® addresses the RCM method and shortfalls in its application. It modifies the method to consider asset and even failure mode criticality so that rigor is applied only where it is truly needed. It removes (within reason) the sources of concern about RCM being overly rigorous and too labor intensive without compromising on its ability to deliver a tailored failure management program for physical assets sensitive to their operational context and application. RCM-R® also provides its practitioners with standard based guidance for determining meaningful failure modes and causes facilitating their analysis for optimum outcome. It places RCM into the Asset Management spectrum strengthening the original method by introducing International Standard based risk management methods for assessing failure risks formally. RCM-R® employs quantitative reliability methods tailoring evidence based decision making whenever historical failure data is available.