Click Image to Enlarge

Indeed, the economic crisis in 2008 already led to enormous spending cuts across the globe. In Europe, the post-war infrastructure, especially bridges, is ageing. Despite that, the maintenance backlog, i.e., the amount of maintenance and rehabilitation that should have been completed in order to maintain infrastructures in good condition but has been deferred, is growing considerably. This problem could be being amplified because of the COVID-19 Pandemic, that causes further service cancellations, delays and consequently spending cuts.

This article aims to demonstrate the importance to implement the reliability and maintenance program during infrastructure concept and design phase, as well as A.I integrated to Asset Management and Asset Integrity Management during operation phase.

2 – Reliability and Maintenance Program for Infrastructure

The maintenance activities applied to infrastructures physical assets need to be under the context of a Maintenance Program, that considers different reliability engineering methods to be implemented throughout the different life cycle phases from concept to the decommissioning phases.

The Reliability Centred Maintenance (RCM) is a method initially applied during the design phase, that aims to define the maintenance tasks based on the infrastructure failure modes, causes and effects as well as the associated risk. The information that is input in order to perform the RCM are the infrastructure’s failure mode and effect analysis (FMEA). In the Railway Industry, the infrastructure’s systems play an important role in terms of the safety and performance of the railway. One good example is the rail component that in case of failure, will trigger impact on a Railway System’s operational availability and may trigger a major accident such as derailment as shown in figure 1.

Click Image to Enlarge

Figure 1: Rail FMEA/RCM ECC Database. Source Eduardo Calixto 2021.

Figure 1 describes the rail RCM where the risks are assessed as intolerable based on the combination of the cause frequencies and the consequence severity. Based on such assessment, different maintenance task types and the frequencies at which they are carried out are defined to mitigate the risk, such as Visual inspection and Track Road Vehicle Inspection (Ultrasonic Test). Similar methods comparable to the RCM approach is the RBI.

The Risk Based Inspection (RBI) is applied initially in the design phase and later during the operation phase. The RBI Infrastructure scope focuses on a failure that can trigger a major accident. The RBI method is implemented based on specific procedures and standards such as: API 580, API 581 and EN 1691, which can be qualitative or semi-quantitative based on the RBI application levels.

In addition to qualitative methods, it is important to have quantitative analysis to predict the Infrastructure Physical Assets RAM performance such as Lifetime Data Analysis, also popular known as Weibull Analysis, RAM Analysis and Reliability Growth Analysis. These methods can be applied to assess, verify and validate the infrastructure system’s RAM performance.
However, the Probabilistic Degradation Analysis (PDA) is more appropriate to predict such an infrastructure’s reliability performance. The PDA aims to define the infrastructure’s physical asset reliability based on integrity degradation failure data related to the thickness of a crack, corrosion and erosion measured by non-destructive test methods. By applying these methods, it is possible to predict when the functional failure will be achieved based on the trend of degradation such as thickness or depth (Crack or Corrosion) and by considering the degradation limit as shown in figure 2.

Click Image to Enlarge

Figure 2: Probabilistic Degradation Prediction. Source: Calixto E, 2021 – Software Weibull++ HBK.

The pink line 1.5 mm in figure 2 is the limit of corrosion, where a functional failure is expected to occur. The other different lines are different measurements that predict the trend of the evolution of corrosion a different points in time. Therefore, if we project the interception of each of these lines with the straight line (1.5 in Y axis) in x axis, there will be different times of functional failures. These functional failure times are used to predict the reliability and the failure rate function. Based on such information it can be defined when the inspection needs to take place.

3 - Maintenance 4.0 applied for infrastructures.

Artificial Intelligence (A.I) aims to enable a machine to think and make its own decisions based on data collected and assessed automatically without any human intervention. Based on the EFNMS – European Committee Maintenance 4.0 (ECM4.0) 2021, Industry 4.0 is a new paradigm and the last industrial revolution, that has been implemented across the globe intensively in the past five years and is supported by the utilization of Enabling Digital Technologies named 4.0.

Concerning the application of A.I for Infrastructures Physical Assets, machine learning is applied for the equipment criticality and critical alert levels classification, failure regression predictions and the automatic application of the Prognostic Health Management (PHM). In the case of an Infrastructure system, the stress factors measured by sensors are vibration, voltage, temperature, humidity; Non-destructive test measurements are also taken such as crack thickness, corrosion depth and other physical parameter that lead equipment degrade to functional failure.

Click Image to Enlarge

Figure 3: Flow Chart of A.I Deep Learning applied for Bridge. Source: Mohammad Noori.0 2021.

The Deep Learning (DL) methods, a special type of Machine Learning, can also be applied to support the preventive maintenance of an Infrastructure System. The DL is a more sophisticated machine learning method, that applies a deep neural network that encompasses several hidden layers as shown in figure 3. The principles of Deep Neural network consider different layers such as Convolution Layer, Pooling Layer, ReLu, Fully Connected, Softmax and the output image classification. (https://www.eduardocalixto.com/paper-2021/)

Despite the advantage of applying A.I Deep Machine Learning and other A.I methods as well as reliability engineering methods, it is necessary to integrate such methods in an Enterprise Asset Management System. This will enable the management of the preventive maintenance tasks defined for all these methods and will ensure that the proper resources are allocated in the proper time to mitigate the Infrastructure Physical Asset risk of unavailability along time and the possibility of a major accident. The next item will discuss the Asset Integrity Management as part of the Asset Management.

4 - Infrastructure Asset Integrity Management (AIM)

An Infrastructure Physical Asset integrity failure may lead to unavailability, or a major accident with multiple fatalities. Therefore, the so-called safety critical elements (SCE) are the physical assets, which in case of failure, may lead to a major accident such as jet fire, toxic cloud release, explosion, fire, toxic product spill, aircraft crash, trains collision or derailment. In fact, a major accident can be triggered by Infrastructure integrity failure, software, hardware or human error or a combination of such factors.

In order to mitigate such risks of a major accident it is necessary to implement a Reliability & Maintenance (R&M) Program immediately at the first stage of the concept and design of a physical asset’s life cycle and implement all recommendations from such R&M methods. After that, it is necessary to implement the risk management and inspection & test program concerning the A.I technologies during the operation phase.

The Asset Integrity Program can apply the same elements of the AM defined in ISO 55000 such as context of the organization, leadership, planning, support, operation, and performance evaluation but needs to focus on the critical safety elements management.

Click Image to Enlarge

Figure 4: AIM flow into AM Process. Source: Stewart Paul. Integrity PRO, Enkelt,2018

Since 2010, the new era of Industry 4.0 has become a reality for many industries across the globe. In the last five years new IOT technology development has been integrated with EAM solutions concerning technologies such as Big Data, PHA and Machine Learning, Reliability 4.0 and the usual maintenance management routine.

Despite all development that enables integrated AM, too much focus has been given to availability performance and maintenance, with a lack of effort for safety concerning the safety critical element management.
However, it is very important to establish a process to enable an effective AIM flow integrated to the AM process. Figure 4 describes the AM and AIM flow, highlighted in green, as part of the AM flow. In the case of AIM, safety management takes place in the fourth step, which encompasses the safety routine management (safety meeting and incident reports) as well as the Barrier Management. The Safety Barrier model is part of the barrier management that defines the level of risk of each SCE automatically, based on real online data.

Since the SCE is defined based on previous risk analysis considering severity criticality, the Risk Management of such SCE performed by the Barrier Model is automatically updated, enabling Asset Integrity and helping Safety managers manage the risk of the SCE on a daily basis.

The Infastructure’s Physical Assets need to have in the end, all information integrated in an EAM that encompasses the best A.I technologies and reliability engineering methods to enable the leaders to make a fast and reliable decision.

Dr. Eduardo 

Calixto,

Eduardo Calixto Consulting (ECC), RAMS Engineer and Asset Management Expert, EFNMS member