In 2001, a publicly traded energy company headquartered in Newark, New Jersey, Public Service Electric and Gas (PSE&G), which provides electric services to over 2.2 million customers, launched a system to enhance the efficiency of its equipment maintenance operations.
It was through PSE&G’s Electric Delivery Asset Strategy department that a Computerised Maintenance Management System (CMMS); was envisioned and launched, providing a proactive, condition-based maintenance (CBM) approach to help predict equipment failures before they occurred. Thus, for the past 15 years, the utility has seen tremendous value in the CBM programme, which allows the company to focus efforts on performing the right maintenance at the right time. This tool predicts equipment failures and calculates transformers’ true age, which is a critical asset for the planning and operations.
In every utility, planning is important to ensure that the system is properly designed and maintained to handle peak demand periods (whether summer or winter), while meeting the design criteria, i.e. n-1 or n-1-1. Here n-1 (normal -1 asset) refers to losing the most critical asset, such as a transformer or a transmission line, during peak demand without resulting in any overload or voltage concerns.
However, it is of utmost importance to monitor the health of key electrical assets in order to ensure that the system is functioning as designed and to implement the needed repairs/ diagnosis before the peak time. Taking a troubled asset out of service during the peak period to provide the needed maintenance outweighs a component failure.
The main objectives of the CMMS are as follows:
- Enable users to perform the appropriate maintenance to ensure a safe, reliable and cost effective approach;
- Centralise and correlate operational, diagnostic, maintenance, real-time sensor and nameplate/characteristics data, down to the asset level; and
- Develop a CBM and life cycle algorithms transforming data into business plans for asset replacement.
The CMMS platform is comprised of numerous modules; this system allowed PSE&G to save more than $50 million in maintenance and capital costs, where the majority of these savings are attributed to substation transformers and LTCs failure avoidance.
Setting the programme to complete the task
The strategy behind CMMS is to collect and centralise data from multiple systems, to transform that data into actionable information via CBM algorithms, to save the results to a data warehouse, to analyse the data and for personnel to act on the information. The CMMS data architecture collects and centralises real-time operational, real-time nonoperational, diagnostic, maintenance and equipment characteristic data. The CBM platform includes factors such as cooling performance, load performance, dissolved gas analysis (DGA) tests, taps movements and physical tests from comparative equipment; and ranks the equipment according to the sum of the factors. CBM logic data extracted during operations and/or maintenance intervals is utilised to forecast the need for additional or future maintenance.
The CBM’s goal is to allow maintenance personnel to locate and resolve asset functionality issues before they escalate. The designed algorithms are set up to provide a transformer ranking system; the algorithm is based on the data available for each particular asset, and the factor weightings vary between voltage classes. The CBM transformer algorithms look at detectable acetylene, moisture, dielectric strength, cooling performance and combustible gas rate of change.
Based on the CBM algorithm, the replacement algorithm is based on the condition score, chronological age, IEEE loss of life aging factor, asset’s physical condition, bushing condition, spare availability and critical customers supplied. The calculated score is then reviewed with a focus on assets with the highest score.
The ultimate goal is to have a high degree of reliability and relatively low failure rates. At PSE&G, it is known when the transformer was built and when it was installed; the issue or lack of data resides in the fact that sophisticated monitoring equipment were not installed on the 60+ year old transformers when they was first installed. As a result, there is no track record of detailed transformer loading and how many faults this specific transformer has witnessed since it was first installed. Thus, the difficulty of knowing the true transformer age, not to mention the lack of knowledge regarding all the maintenance that was provided or never provided prior.
Industry information presented during an IEEE Transformers Committee Tutorial in 2007 indicated a tendency for low failure for transformers younger than 25 years, moderately increasing failure for transformers around 35 years of age and rapidly increasing failure rates beyond 35 years of age.
During normal transformer operation, dielectric integrity deteriorates over time due to thermal aging of insulation. Moisture introduction, through leaks, and fault activity outside the transformer can accelerate this deterioration, thus increasing the probability of failure. Deterioration can be slowed and transformer life extended through CBM. As degradation progresses, the cost of maintenance will increase and the effectiveness of repeated maintenance will diminish, resulting in a detrimental impact on transformer and system reliability; thus replacement will ultimately become necessary.
Adding a monitoring programme with real-time sensor data
The transformer monitoring programme (TMP) is another PSE&G tool that uses real-time sensor data to determine the true age of a transformer by utilising IEEE C57.912011 (guide for Loading MineralOil-Immersed Transformers and Step-Voltage Regulators) equations in order to determine the transformer (insulation) aging acceleration factor and loss of life for specified periods of time. The data collected includes transformer top oil temperature, winding/hot spot temperature, ambient temperature, transformer load, voltage, cooling hours and LTC tank temperature. The CMMS utilises the algorithms to create a useful satellite view of the substation and a heat map representation of transformer loading as shown in Figure 1.
By clicking on any station in the online application, the transformer load curves are drawn up compared to the normal or emergency ratings. This satellite map also allows the user to forecast the next day’s load by plotting the expected load versus the ratings. The IEEE Loss-of-Life calculation provides the needed computation to determine the number of days a transformer has aged based on the heat run properties of the transformer, continuous load, top oil temperature, ambient temperature and winding/hot spot temperature for a selected timeframe.
The calculation allows planning and asset management to determine whether to keep any transformer in service that has surpassed its expected life cycle; it also calculates aging per year and helps drive transformer replacement algorithms. The CMMS provides engineers and planners with easy access to transformer load and loss of life profiles; it also provides the ability to forecast data and integrate with other data sources.
Initial expense well worth the outlay
In today’s world, planners are challenged to find creative and out of the box ways to defer capital expenditures and still reliably meet the forecasted system peak demand. Prior to implementing the CMMS, the planner’s goal was to ensure that the available capacity could meet the forecasted load under design contingency. The main assets monitored throughout the peak days are transformers, cables and transmission lines – basically the most critical, expensive assets, not to mention those assets that take the longest to restore.
Thus having an analytical CMMS tool that can accurately predict failures and inform planners and others about the overall health of the fleet is priceless. With the current shift in reducing operation and maintenance (O&M) costs and capital expenditure, a platform based on the likes of the CMMS will continue to be very useful.