RBM-Fuzzy Risk Based Maintenance Planning with Fuzzy Logic for Utilities

Asset Management
The life of the public utilities is basically defined by the changes of the last decade, the privatization and the unbundling, and direct changes in regulation and consumer requirements. Beside the external factors significant internal aspects play a part: for the sake of the long term sustainability halting of the network aging is a critical task.
The changes of the environment fundamentally modified the business strategy of the companies and the relationship to the network assets: the economically effective management and operation of the network became the core of these companies’ business. To achieve this an asset management system has to be operated that ensures that only really sufficient expenditures are defined and in which the rate of effectiveness can be demonstrated.
Risk Management
One of the main principles is the risk management of the ISO 55000:2014, the international standard of the asset management. The risk can be defined at various levels and processes of asset management (strategy, planning, implementation of the plans, education, IT support, etc.), and at the level of the assets themselves.
Risk based maintenance planning is strictly aligned with the risk management of the company, and basically evaluates the risk represented by the assets themselves and manages it through different asset intervention possibilities. Beside other procedures (like PM, RCM, CBM, etc.) RBM is mentioned as a separate method by the literature.

Beyond the technical condition of a network element the risk also represents the importance of the element’s role in the network. From a practical point of view the more data is available, the more exact evaluation of risk can be reached, but to start using risk evaluation is a conceptual and methodological issue.
Risk based planning can usually be introduced based on the utilization of the same data that is used in the current planning procedures of the companies. By introducing risk evaluation the currently used qualitative and heterogeneous planning may be replaced with a unified, transparent and quantitative procedure.

Maintenance Planning
One main element of an Asset Management System is Asset Management Planning. The purpose of planning is to achieve the strategic goals of the organization, also considering the gap between actual and required performance levels.
The basic units of the planning are the significant maintenance unit, which is different in each network type, and the potential operational or capital expenditure type interventions on these units. These interventions, can mitigate the risk with a given rate, or enhance the performance of the network with a given probability. The planning is a complex optimization task covering both technical and economic aspects and resulting in the optimal combination of basic units and the interventions on them.

Fuzzy Based Risk Evaluation Model
Geometria developed a risk evaluation model (available for medium voltage cable network) to support maintenance planning based on fuzzy logic, which is a scientifically developed procedure. The particular advantage of the fuzzy logic is that it can be effectively used in an incomplete data environment.
The scheme of the risk evaluation model is shown in the figure below.

Beside this analytically not, hardly or only implicitly expressible, nonlinear relationships can be mapped with verbal rule sets in an easy and transparent way.
When creating a specific model the knowledge and the experience of the experts have to be mapped with verbal rules in a structured way. The modification and the tuning of the mapped rules by introducing new rules and changing the weight of the rules is a simple task for the users, in contrast with the modification of difficult mathematical functions.

Pilot, validation, economic evaluation
In 2010, ELMŰ-ÉMÁSZ decided to test the evaluation model for medium voltage cables developed by Geometria. Therefore, ELMŰ-ÉMÁSZ and Geometria together updated the evaluation model and performed a pilot evaluation on a sample network.
The model has been tested on an approximately 100 km long, real medium voltage cable network. The value representing the network condition was calculated with the model, taking into account the 10 years earlier state of the network. Values from the fuzzy model and from other models (a pointing method, an age-based evaluation and a randomized selection method) were compared with the failures that later occurred. The Fuzzy model was significantly better in predicting failures than the other evaluations. By applying the model, 15-20% less cable reconstruction is needed to achieve the same rate of risk mitigation, compared to the other models.

RBM planning at ELMŰ-ÉMÁSZ
Based on the successful evaluation of the sample network, ELMŰ-ÉMÁSZ decided to implement a central risk based maintenance planning for the elements of the distribution system, applying the fuzzy based evaluation model. The application of this method enables the optimal, bottom-up planning of the budgets for each network intervention program.
IT Support – Decision Support System (DSS)
Based on the risk evaluation model a decision support system has been developed which supports maintenance planning. The system determines suggested interventions based on the risk evaluation of the network elements.

The system has four modules:

  1. Data Warehouse Generating Module: it is an ETL (Extraxt, Transform, Load) function to establish a data warehouse from the transactional databases (information systems of the network, events, failures, network condition, and customers).
  2. Fuzzy Based Risk Evaluation Module: the system predicts the value of the failure probability of the element based on technical, condition, environmental (e.g. vegetation) and failure data, and the consequence of the failure based on importance, function, supplied consumers and contingency state of the network.
  3. Planning Module: on the one hand it selects the optimal combination of significant maintenance units (e.g. per evaluation of parallelism of cables), on the other hand it supports the selection of the interventions using various indicators.
  4. Map Based Designer Module: it enables the evaluation of aspects that cannot be automated in the Evaluation and the Planning module by displaying information in the map based network registry system, enabling the exact bordering of the selected elements.

Actual state of development and the future
The above mentioned modules got in operation for all kind of LV and MV networks and transformer stations for 2014.
Later on, it can be expected that the risk based planning will be harmonized to different types of preventive maintenance methods (e.g. cycle time based), to risk based condition survey, and it will be also enhanced with a financial performance evaluation module (costs, TCO). As a functional development of the Designer Module the given interventions might be defined in different ways which then can be evaluated and compared with a net present value method.

Screenshots from the planning and the map based designer module: