Decision support system for predictive maintenance planning

The Project Background

In 2012, on the recommendation of our customers, we were contacted by a liquefied natural gas (LNG) producing company from Middle East region. Its operations consisted of the extraction, processing, liquefaction, storage and export of LNG and all associated derivatives.

The customer needed to service more than 500,000 pieces of equipment, which required about 1 million man-hours of maintenance per year. In addition, when planning, it was necessary to take into account many factors (constrains), such as the season, the availability of related equipment and specialists, the urgency and duration of work, the dependence of various types of work and appliances, etc.

When working with such a quantity of equipment, manual planning is extremely time-consuming. And the slightest mistake when entering data may cause huge financial losses. One of these errors resulted in equipment downtime that cost the company $ 15 Million a day.

The Challenge

At that time, the Customer had a whole planning department of 25 people who manually entered data into Excel. It was ineffective and time consuming process that ended up with a serious failure that led to multi million downtime. To solve the existing problems in maintenance planning and service workflow automation RNDpoint team was invited.

Like many other companies of this scale, the client used SAP to automate their work. Therefore, they turned to their SAP supplier to develop solution tailored to their needs. After two years of time-consuming work on a project based on SAP PM, it was decided to stop it and find another solution.

The Solution

So the client turned to us. Together with the client, RND Point’s experts examined the possibilities of using existing platforms (Primavera, Microsoft Project, etc.) and found that they do not comply with industry-specific business requirements and have low productivity. Moreover, existing systems were expensive to implement and support. As a result, we proposed to develop a custom decision support system, which was called AllMonitor.

In 2010-2012, we developed and implemented a solution that met the needs of the customer. The system included two main predictive planning modules: a long-term planning module and a short-term planning module.

The long-term planning module allows you to calculate the optimal amount of resources (machinery, people and equipment) for a long period of time (5-10 years) and offers the best option out of the many possible for scheduled maintenance. This takes into account a number of constrains:
• cyclic schedule (repetition of work with a given interval);
• seasonal restrictions;
• different performance at different times of the year;
• interdependent work;
• availability for service.

The short-term planning module draws up a work plan for the near future (from 2 weeks to 1 month) for unscheduled maintenance. The decision support system calculates the maximum amount of work, taking into account the following factors:
• importance, urgency and interconnection of work;
• availability of equipment and machinery;
• the presence of workers of a certain specialization;
• performance of work in different shifts.

The Impact on the Client’s operations

63%

Reduced cost of maintenance

3.1

Million USD saved annually

6 to 2 weeks

reduces unscheduled work

  • The developed system for predictive maintenance allowed to significantly reduce the cost of equipment maintenance and personnel (including subcontractors, downtime, processing, equipment).
  • With long-term planning, equipment downtime was reduced and the amount of human resources used was optimized. This allowed the customer to save more than $ 3 Million per year.
  • With short-term planning, it was possible to reduce the time for unscheduled work, to reduce the likelihood of downtime. After the implementation of the system, the queue of unscheduled work reduced from 6 to 2 weeks. Customer saves more than $ 1.5 Million a year.

Have questions? Get in touch!