Real-time power management is a plant manager’s secret weapon
A modern power management system requires new techniques and cutting edge technology to allow electrical power users and producers to be competitive. In light of rising electricity costs and disruptive power outages, it’s imperative for power management system software to put plant managers in control of operations, maintenance, and planning of the electrical system – resulting in optimum system utilization, lower costs, and financial stability.
This new breed of model-based power management application should have the capability to integrate an active blueprint of the system including system topology, engineering parameters and other pertinent information with time-synchronized-data acquired for the purpose of depicting the actual operation of the system.
Advanced applications and simulation engines should allow improved situational awareness, look-ahead proactive approach and improved decision making for operators under emergency conditions. The same power management system should therefore serve operators, engineers, planners and managers by providing pertinent information to various levels within the enterprise. The information should be available to engineers and planners via desktop clients while the operators and managers can also rely on thin clients.
One of the key advantages of utilizing a model based power management system is maintaining consistency of the network model across engineering, planning, protection and operations department. Traditionally, real-time systems use power system models that differ considerably in detail and structure from the models used for offline studies. Linkages between the different models are typically not maintained, and the different models often have incompatible data formats.
Planning and operating decisions are based on the results of power system simulations. Optimistic models can result in under-investment or unsafe operating conditions while pessimistic models can also lead to unnecessary capital investment, thereby increasing the cost of electric power. Realistic models are needed for ensuring reliable and economic power system operation.
Hence it is a very crucial phase to verify & validate (V&V) the network model with real-time and/or archived data and prepare a benchmarked model for:
- State estimation / Monitoring
- Predictive simulation “what-if” analysis
- Forensic “root cause & effects” analysis
- Proactive contingency analysis & Remedial Actions
This can be achieved by utilizing a power management system that offers traditional simulation analysis tools on the same platform as the real-time operations tools. Doing so avoids the necessity to rebuild and maintain separate network models across various departments.
System monitoring is the base function for any power management software. In addition, seamless integration with metering devices, data acquisition, and archiving systems are essential to monitoring software. Real-time or snapshot data are linked to an online model of the system for proper presentation of actual operating status.
All this information should be accessible to the system operator through advance man-machine interfaces such as an interactive one-line diagram that provides logical system-wide view.
The next step is to process the telemetry data and determine the missing or faulty meter values using advance techniques such as State and Load Estimator (SLE).
The system should also be able to compensate for absence of physical meters by providing virtual metering of devices. Standard power monitoring systems are inadequate since they can only monitor based on the “eyes” you provide in the form of digital measurement devices. These devices can cost $5,000/unit depending upon their complexity and it quickly becomes prohibitively costly to install such meters at every location.
Virtual meters not only improve situational awareness, but also provide a means to alarm equipment (especially low-voltage) that is not visible to a traditional power monitoring system. A model-based power management system uses existing metering devices and makes estimates for the portions of the system that is not monitored.
A chemical plant avoided installation of five such meters and relied on estimated data for non-critical areas and realized a savings of $20,000 in capital expenditure immediately and enjoyed the supplementary benefit of complete system visibility and information for every load in the system.
Dashboards and thin clients
Energy dashboards summarize and record alarm conditions in case of unusual activity and provide continuous visual monitoring of user-selected parameters in any mode of operation. This provision would allow early detection and display of problems before a critical failure takes place. A modern power management system should not only provide monitored data via thin client, but also offer the following key advantages:
- Utilize the same electrical model as the desktop client and the offline planning model without having to recreate or maintain copies of the model. This results in significant time and cost savings when building Human-Machine Interfaces (HMIs). Traditional power monitoring systems are inexpensive to purchase, but take up a significant amount of time, resources and engineering cost to setup the HMIs. Extensive engineering man-hours are also spent modifying the existing HMIs. While in a model based power management system, the offline study model can be simply transitioned and connected with real-time data.
- Ability for the operator to recall and run pre-defined scenarios and get a simple decision (go / no go) especially when he/she is facing emergency conditions. Information overload will not only slow down every decision, it may invariably lead to complete system shutdown.
System engineers and operators must have instant access to energy information and analysis tools that allow them to predict an outcome before actions are taken on the system.
In order to design, operate, and maintain a power system, one must first understand its behavior. The operator must have firsthand experience with the system under various operating conditions to effectively react to changes. This will avoid the inadvertent plant outage caused by human error and equipment overload. The cost of an unplanned outage can be staggering (See Figure 1 at top.)
For industrial and generation facilities that utilize power system analysis applications, the ability to perform system studies and simulate “What If” scenarios using real-time operating data on demand is of the essence. For example, using real-time data, the system operator could iteratively simulate the impact of starting a large motor without actually starting the motor.
Sequence of event playback
The ability to recover from a system disturbance depends on the time it takes to establish the cause of problem and take remedial action. This requires a fast and complete review and analysis of the sequence of events prior to the disturbance. Power management software should assist your operation and engineering staff to quickly identify the cause of operating problems and determine where energy costs can be reduced. The software should also be able to reconstruct exact system conditions to check for operator actions and probe for alternative actions after-the-fact. This important tool serves as an on-going learning process for the operator.
Besides reducing losses and improving data gathering capability, such an application should assist in increase plant reliability and control costs. The event playback feature is especially useful for root cause and effect investigations, improvement of system operations, exploration of alternative actions, and replay of “What If” scenarios.
Event playback capability translates into savings. For example, a conservative estimate of 10% reduction in downtime for an outage that lasts an hour yields about $33,000 in savings.
An advanced power management system should provide the options for full remote control to the system elements such as motors, generators, breakers, load tap changers, and other protection devices directly or through existing Supervisory Control and Data Acquisition (SCADA) system.
In addition, the software should provide user-definable actions that can be added or superimposed on the existing system for automating system control. This is similar to adding PC-based processors/controllers (kV, kW, kvar, PF, etc.) or simple breaker interlocks to any part of the system by means of the software.
Supervisory and advisory controls
State-of-the-art supervisory and advisory control capabilities should be used to control and optimize in real-time various parameters throughout the system. Using optimization algorithms, the user could program the power management system (i.e., assist energy consumers by automatically operating their system to minimize system losses, reduce peak load consumption, or minimize control adjustment). For energy producers, this energy management system could be set up to minimize generation fuel costs, and optimize system operation.
Demand response or management
Another significant cost component of operations is demand charge of the energy bill. The demand charge is 40% to 60% of the bill for sites without peak shaving generation. A single unmanaged demand charge can produce a very large hike in the power bill each month and with “ratcheting” demand charges, the effect can linger for an entire year.
An intelligent power management system should provide the current and predicted demand for each day thus managing peak demands on a continuous basis. Loads can be shed manually or automatically, peak-shaving generators can be started, load startup can be postponed or sequenced, while maintaining and preserving vital processes.
A classic example is an industrial plant that purchases a new large motor and decides to perform a full-run test for 30 minutes. At the end of the month, the purchasing department notifies the operations manager because their operating bill is up by 7%, or nearly $20,000. Since their contract was a “ratcheting” contract, the peak demand charges they paid that month lingered for 10 months, making the motor test a $200,000 mistake, which could have been avoided if system demand had been monitored, simulated with system operating conditions and therefore fully understood.
In a study performed for a large industrial facility (150MVA), advanced optimization algorithms, native to the energy management system, were utilized to reduce real and reactive power losses. Assuming a conservative power loss reduction of only 0.1% at an average electrical energy cost of $0.13/kWh, an energy management system would yield savings of more than $135,000 per year and would pay for itself through the immediate realization of savings in operating and maintenance costs.
Intelligent proactive load shedding
A major disturbance in an electrical power system may result in certain areas becoming isolated and experiencing low frequency and voltage, which can result in an unstable operation. A model based power management system can easily determine the electrical subsystems (areas of the network that are electrically isolated) including islanding detection. The system should then provide optimal “fast load shedding” based on the actual operating condition of the system including type and location of the disturbances.
Fast response time of less than 40ms (duration between disturbance to tripping non-essential load) is achievable. High-speed load shedding has become a necessity because the stability margins for today’s complex systems are razor thin. The longer it takes to shed load during a disturbance, the more load that must ultimately be shed.
By adding intelligence and proactive calculations, fast response times can be achieved. By speeding up the load shedding process, the actual amount of load that is shed is far less than that of using the conventional methods such as frequency relay and PLC-based schemes; both of which rely on fixed logic, lack logic validation using stability analysis and post process the disturbance adding to the response time delays.
The power management system should have the intelligence to initiate load shedding based on a user-defined Load Priority Table (LPT) and a pre-constructed Stability Knowledge Base (SKB) in response to electrical or mechanical disturbances in the system. Load shedding schemes by conventional frequency relays are generally a static control with fixed frequency settings. Based on Neural Networks, a power management system would be able to adapt to all real-time situations and provide a true dynamic load shedding control. This would allow the operator to optimize load preservation, reduce downtime for critical loads, and simulate/test the load shedding recommendations.
The Intelligent Load Shedding built into a model based solution offers yet another advantage; logic verification. Logic verification is a necessary step during Factory Acceptance Test (FAT) as well as Site Acceptance Test (SAT) performed during system commissioning. With an integrated model based solution, it is easy to capture operating data and evaluate results of Intelligent Load Shedding using dynamic stability.
This offers analytical and instantaneous visible confirmation that the load shedding logic will perform as expected without going through commissioning issues that are sometimes seen with hardware only based systems. Every scenario can be checked by using the live system data and a transient stability program. Since the model may be expanded to accommodate additional substations or manufacturing trains, the load shedding model is also updated in the process and the entire system can be retested in a fraction of time it would take for conventional hardware based systems (See Figure 2).
Real-time power management
A standard power management system evaluates collected data in a non-electrical system environment without recognizing the interdependencies of equipment. Extending the power monitoring system by equipping it with an appropriate electrical system context, simulation modules, and playback routines will provide the system operator and engineer with a powerful new set of tools. Using these tools, the user can accurately predict the behavior of the electrical system in response to a variety of changes.
The playback of recorded message logs into the simulator-equipped monitoring system provides the operator with an invaluable means of exploring the effects of alternative actions during historical events. Simulation techniques readily extend into power system control and can be used to perform system automation and load shedding functions.
Finally, all of these capabilities should be included in one application with the flexibility and compatibility that allows plants to expand and upgrade as their needs grow. Real-time power management has become a plant manager’s secret weapon in assuring reliable plant operations 24/7.
Shervin Shokooh, is Sr. Principal Electrical Engineer and Tanuj Khandelwal is Principal Electrical Engineer for ETAP.