What do you do when faced with too much information? Perhaps surprisingly, you gather more… The key is to ensure that data becomes meaningful information and that information then becomes useable intelligence.

When dealing with vast quantities of data, the system for collecting, interpreting, sorting and prioritising it is central in enabling data to become useful. Of course, how one goes on to use that intelligence is what really makes the difference.

Information overload?

For the team at London Underground Asset Performance, Jubilee, Northern & Piccadilly line (APJNP), a very real case of information overload was developing in its system to record and monitor alerts on systems used for running the railway – and becoming a physical as well as an intellectual problem.

“Between the dawn of civilisation through 2003 about 5 exabytes of information was created. Now, that much information is created every 2 days” Eric Schmidt, former Google CEO.

With such an explosion in the creation of data in recent years, it is no surprise that data handling systems are having to evolve quickly in order to remain effective.

Within the underground network there are many systems critical to running a service and ensuring stations, as well as the line, can operate effectively: sump pumps and groundwater pumping systems, ventilation systems, escalators, lifts and more. Failure, or threat of failure, in a critical system can result in station or line closure, accompanied by the full portfolio of associated troubles – safety issues, delays, service disruption and misery for travellers, not to mention potential expense for operators.

APJNP monitors a number of these critical systems at its Fault Reporting Centre (FRC) in order to respond immediately if a system fails.

When a failure or malfunction occurs, an alert is triggered and that information is fed through to the FRC for problem management.

The systems connect to more than 100 information management systems, collecting alerts from approximately 15,000 assets. There are over 225,000 individual alerts triggering more than 200,000 alarms per day – 1,800 per minute.

The monitoring systems in use have evolved over the last 16 years. Alert data arrives in the Central SCADA (supervisory control and data acquisition) system and other Station Management Systems, via an iMAC computer. Operators then perform a manual check, requiring specific instructions or processes from different asset owners to interpret what alerts mean and which of those alerts are relevant.

The ever-growing list of events requires continual scrutiny by operators, to seek out the important information. Until recently, this had to be done by proactively searching on a list of unique search strings. With this ongoing search being carried out manually, there comes a risk of something important being lost. In this case, as the system evolved and the quantity of data being received grew, the monitoring process became almost unmanageable and so the risks continued to grow.

The problem was how to spot effectively the relatively rare critical event ‘needle’ in the growing haystack of less important or non- critical events. Data handling specialist Telent was asked to simplify the system.

New system requirements

By creating a new Alert Gateway System, Telent aimed to provide a tool for organising and tracking the most important alerts within the APJNP infrastructure. The new system needed to:

Alert, inform and guide the operators;

» Prevent unnecessary service activities from affecting operation delays;
» Only present the operator with useful and relevant alerts;
» Use prioritisation to highlight critical alerts;
» Have a detailed response to each alert;
» Allow enough time for the operator to respond; » Exchange information with enterprise applications.

The potential risk of human error would need to be eliminated and the various data sources amalgamated into one, easy-to-use system. Operators really needed to see, at a glance, the priorities for action, with only the relevant alerts being brought to their attention.

Open architecture enabling further configuration was also a necessity if the system was to have a long life and cope with additional feeds in the future.

A simplified user interface that only required the operator to learn one system would significantly reduce the drain on resources and remove the need for significant training when other assets were added to the system.

Working in partnership

The Alert Gateway would be built on open standards, for either Windows or Linux, and be compatible with a wide variety of protocols, thereby offering interconnectivity in line with current industry standards.

The first stage was Alert Rationalisation: every possible alert was evaluated in order to ensure that, even though the data from all alerts would be recorded, only those alerts requiring action would be presented to operators. An investigation was also carried out to confirm that alerts were only raised by the most appropriate indicator of the root cause of an abnormal situation.

The development process that followed was divided into three phases. Firstly, the central SCADA and iMAC would be integrated into a common Graphic User Interface (GUI). Then the existing connected third party assets could be integrated, followed finally by the integration of existing disconnected third party assets.

Working closely with Telent at all stages, APJNP operators tested the interface and provided feedback for further development, ensuring the system met their needs.

Now, having been installed a few months ago, the new system is operating in tandem with the old system, in order to prove accuracy and reliability before decommissioning of the old system takes place.

Simon Pateman, APJNP stations manager, stated: “The new Alert Gateway has been positively received by APJNP staff in the FRC. The driving force behind this project was to simplify the operator’s task of undertaking remote checks.

“The future aspiration of the Alert Gateway is to be a central point for all remote monitoring for JNP assets. The Alert Gateway has the capability to expand to meet anticipated future demand. Such expansion could include a PC alert system for operators to ensure that alarms are dealt with in a timely manner and asset specific screens for the asset managers to oversee their asset performance.”

Next steps

As Alastair Norman, Telent’s head of asset condition monitoring, explained: “Applying a relatively simple process to a highly complex scenario means the FRC is now in a very different situation; laborious monitoring of vast quantities of data is a thing of the past. An extremely straightforward user interface displaying smaller numbers of colour-coded events ensures that, when critical alerts do occur, they cannot be missed.

“The system now in place is flexible, offering room to grow with future needs. It is also a more cost-effective system, placing much lower demands on manpower and desk space. The system can be applied to a wide variety of scenarios, relatively quickly.”

Not only have the issues of the old system been resolved, but possibilities for future development have been opened up. The role of the Alert Gateway is set to grow as more monitoring systems are added in, including the receipt of alerts by text as well as email.

The new system can also be configured to raise an alarm when specific combinations of alert occur. This would be exceptionally difficult with a manual searching process.

Beyond the basic requirement of highlighting clearly prioritised issues to be resolved, data collection on this scale can, with appropriate analysis in place, provide highly useful information. Trends, cycles, performance usage details and asset behaviour prior to failure can all be monitored, in turn creating a bank of intelligence. This intelligence can inform future Asset Condition Monitoring systems, enabling asset behaviour to be predicted and failures to be prevented.

Intelligent maintenance plans can then minimise and often prevent disruption as well as reducing maintenance costs.