By James Dennis, John Goldring, and Lawrence Berg

Conventional fossil fired boilers across the globe are being subjected to ever more stringent emissions limitations, despite the majority of those installations having been built during an era when the requirement to control such emissions was not a factor in their design. The combination of this departure from the original design intent of the firing system and the aging of associated plant systems, creates unique challenges when these boilers are modified for emissions reduction to allow continued operation.

Careful consideration of the capacity and performance of each plant system is required to ensure that a retrofitted firing system will not impose new operational limitations on the boiler. For example, a new burner must be designed to allow sufficient air-flow through the secondary air register, at the correct velocity to maintain both momentum balance and stoichiometry in the flame.

A detailed analysis is required of each individual unit to consider both the original design and limitations arising due to age related plant deterioration. Failure to complete this detailed analysis and incorporate these factors into the design of an emissions reduction retrofit can lead to critical failures, such as destruction of boiler tube walls and pendants due to increased flame length, limited ramp rates due to steam temperature excursions, large amounts of unburnts leaving the lower furnace or, in the most extreme cases, unstable and potentially unsafe combustion.

This paper explores the considerations taken by RJM combustion engineers during the design phase of a NOx reduction project, giving examples of some of the challenges seen and the methods employed to successfully overcome those challenges.

Combustion Emissions Reduction

Emissions of nitrogen oxides and carbon monoxide can be managed through proper control of combustion conditions inside the furnace; unlike coal ash particulate matter and sulphur dioxide, which require post combustion flue gas treatment such as precipitators or FGD. We describe these systems of emissions control during combustion as “primary measures”, with “secondary measures” referring to the flue gas treatment plant. NOx reduction through primary measures, when properly executed, has the potential to cut the cost of NOx compliance ten-fold when compared to the capital and operating costs of secondary measures.

Primary and secondary measures can be used in combination, to achieve the most stringent emissions requirements. For example, a low NOx firing system can be used in series with an SNCR system. The success of this combination relies on the two systems being designed such that they do not compete. If the low NOx firing system drives for the lowest possible NOX emissions, this can create CO concentrations at the SNCR level, which inhibit the reagent efficiency. The two systems must therefore be designed to complement each other, such that minimal reagent is used.

Primary measures for NOx and CO control include staging of combustion air, by reducing the air available for combustion at the burner, introduction of flue gas to reduce peak combustion temperatures and controlling the peak temperatures within the combustion chamber. Recognising the importance of these parameters relies of an understanding of the NOx formation mechanisms at work within the combustion chamber.

NOx Formation

Nitrogen oxide formation occurs via 3 independent mechanisms: fuel NOx, thermal NOx and Prompt NOx.

In the case of coal combustion, the majority of NOx produced is from nitrogen, stored in the fuel, which is liberated during devolatilization (the process by which the particle heats and the volatile content is released into the furnace). Once these nitrogen radicals are released into the combustion chamber, they will react with available oxygen atoms to form NOx. However, where oxygen is unavailable, these nitrogen radicals are forced to revert to their diatomic state, which can be safely released to atmosphere. NOx formation can therefore be controlled by limiting the oxygen available whilst these volatiles are reacting. This is the reason that air is staged in low NOx coal firing systems; to reduce the oxygen available during the first stages of the combustion. Over Fired Air is then employed to offset this low stoichiometry by injecting large amounts of oxygen post-devolatilization, to oxidise the remaining carbon and thereby preserving combustion efficiency, though the extent of that preservation relies on the quality of the design.

When considering gas combustion, thermal NOx takes the most significant role. This occurs where nitrogen in the combustion air is broken down, to create radicals for reaction with the available oxygen. This formation rate is therefore controlled by the energy available to deconstruct nitrogen molecules in the air, making the temperature of the combustion the key parameter to control where fuel-bound nitrogen is not present. Therefore, the firing system must be designed to distribute the heat release evenly across the furnace, as demonstrated in Figure 1.

Prompt NOx occurs early on during the combustion process. Partial oxidation of hydrocarbons leads to highly reactive radicals, which react with atmospheric nitrogen.

These nitrogenated compounds then react with oxygen in a similar manner as fuel NOx. This pathway is typically only considered at very low (less than 0.03 lbs/MMBtu NOx) NOx emissions on gas burners.

Air Flow Limitations

When designing a staged firing system, the first limitation that must be considered is the limitations of the existing draft plant. Re-distributing the air vertically across the boiler can create additional pressure drops within the system, which the FD fan may be unable to overcome (not least the high pressure drops required to achieve penetration from a Over Fire Air (OFA) system). Similarly, the ID fan may not be capable of accommodating an increase in the excess oxygen level, measured at the exit of the boiler. The impacts of these limitations can be numerous and severe.

Reducing the stoichiometry at the burner requires installation of OFA to maintain combustion efficiency. This creates an additional pressure drop in the secondary air supply system given the OFA is likely to require a much higher velocity than a typical burner, if it is to achieve penetration across the entire furnace. This additional pressure demand must be accommodated by the FD fan, unless additional booster fans are fitted at significant cost.

Additionally, in the case of a coal burner where the stoichiometry at the burner is lowered, the secondary air flow rate is reduced. In contrast, primary air to fuel ratio, and therefore primary air flow rate, must be maintained for safe conveying of the required mass of fuel to achieve full load. Therefore, in order to maintain the ratio of combustion air to primary air/fuel momentum within the flame, a higher velocity is required in the secondary air register, to compensate for the reduced mass flow of air. This additional pressure demand must be accommodated by the FD fan, unless additional booster fans are fitted at significant cost.

In the case of a recent RJM project, a 200 MW forced draft unit, which had been designed for a turbulent combustion system (without emissions control), and operated at 2.5 percent excess oxygen at full load. Based on a 34 percent unit efficiency and a standard fuel specification, this unit would require approximately 430,000 CFM of combustion air. Assuming the unit is served by two identical FD fans, the required flow per fan would be 215,000 CFM. This analysis assumes that the FD fan is operating as per its design and has not deteriorated.

A typical burner might require 5” wg of pressure drop to maintain design flame dynamics, adding an assumed system resistance through the secondary air supply duct work, associated air heaters and control surfaces produces a total system resistance of 25” wg at full load. This is plotted as point 1 in Figure 2.

Based on typical furnace dimensions for this size of unit, an estimate of 20” wg is made for the required OFA duct to furnace differential pressure to provide the OFA jets with sufficient velocity to penetrate the flue gas stream and achieve the required level of flue gas / air mixing. The magnitude of this pressure requirement can be affected by the depth of the combustion chamber, the internal cross-section of the furnace and subsequent flue gas stream velocity and the accessibility to install OFA ports on multiple wall of the boiler. Where access is only available at the front wall, the velocity (and therefore pressure) demand may be much higher.

The addition of the OFA therefore, moves the system resistance curve from the orange, to the yellow position as shown on Figure 2. This is because, with the modification of the flow areas into the boiler and the increased air velocity required to achieve that flow rate through the system, a proportionately higher amount of pressure is required to realize a given flow rate.

Assuming no change in unit efficiency, the amount of fuel required for full load remains constant and therefore the required air flow for 2.5% excess oxygen at full load remains at 215000 CFM. Referring back to the fan performance curve in Figure 2, we see that the required pressure to achieve this flow rate (point 2) is now above the pressure limit of the fan for that flow rate and the fan is therefore incapable of meeting the operational requirement and instead, the maximum air flow rate is reduced to approximately 205000 CFM. This is a 4% reduction on boiler air flow, and would limit the full load excess oxygen to 1.7%; which is unlikely to be enough to maintain complete combustion.

Role of Secondary Air in Flame Dynamics

Not only would the limitation on pressure and flow affect the penetration of the OFA across the flue gas stream, as shown in Figure 3d, but it would also affect the burner’s flame dynamics. As outlined above, the momentum balance in the flame is essential for maintaining flame stability, structure and emissions performance.

Correct flame dynamics position the combustion within the depth of the boiler and retain the visible flame within the combustion chamber, or below the furnace nose, as shown in Figure 3a and Figure 3c. A weaker secondary air momentum resulting from reduced flow or velocity will produce a longer, thinner flame as the axial component of the burners momentum ratio, which is contributed mainly be the primary air, will dominate; as shown in Figure 3b. The thermal profile of these older furnaces, originally designed for turbulent burners, is centred on a large heat release near the front wall of the furnace. Longer, thinner flames push that heat release onto the rear wall of the boiler and ultimately carry flame out of the combustion chamber and up into the pendants, as shown in Figure 3d. As well as elevating temperatures at the pendants, this tube wall impingement rapidly degrades the interior tube walls of the combustion chamber.

This weakening of the secondary air momentum is exactly the case in a project resolved by RJM International in Europe. Here, a low NOx firing system had been installed, with limited fan capacity and space for only one bank of OFA ports on the front wall. Failure to account for this impact on flame dynamics had resulted in long flames, with combustion carrying into the convective passes of the boiler (Figure 4). The heat release amongst the superheater pendants raised peak superheater temperatures by 75° above design; causing repeated tube failures, significantly lowering plant availability in a market where load factor would otherwise have been high.

Limited superheater temperature control also inhibited the unit’s ability to respond rapidly to changes in load demand, given such transient conditions would result in further deterioration in flame structure. In the absence of strong flame dynamics, the ignition front and fuel/air mixing are compromised when the momentum ratio is varied, such as during a load ramp when air and fuel velocities may vary independently of one another; thus exacerbating the firing systems limitations.

Through intelligent burner design and utilising advanced CFD techniques, RJM were able to custom-engineer a solution which returns flame structure and actually improves combustion efficiency within the limitations of the existing draught plant (Figure 5). Rebalancing the furnace thermal profile in this way brought superheater temperatures back to design levels and subsequently ceased tube failures. In addition, RJM delivered a 45 percent reduction in NOx performance through primary measures, improved combustion efficiency through a 1 percent LOI reduction and enhanced load flexibility as flame structure was made stable even through aggressive load ramps.

Flame Dynamics for Emissions Performance

Controlling flame dynamics in this way and ensuring complete combustion within the furnace, even at low stoichiometries, not only balances the thermal profile, but also raises the CO breakpoint; accommodating the NOx performance demanded by modern emissions legislation. As the stoichiometry in the flame reduces, flame dynamics become increasingly important as the fuel particle must be retained in the flame for as long as possible, maximising exposure to the limited oxygen available. Figure 6 illustrates this relationship between NOx performance and combustion efficiency. The effect of advanced burner dynamics is to move the CO curve to the left and to slide the Ideal Operating Range further down the NOx emissions curve.

As outlined above, though stoichiometry is the key parameter to control for coal combustion, in the case of gas combustion, the critical parameter to consider is the flame temperature. Assuming no change in the heat liberation per megawatt generated, the total thermal heat release for full load must also remain constant. Temperature is a function of the heat released per unit volume. Therefore, to reduce the flame temperature, the heat release must be spread as evenly as possible across the available space. RJM achieves this through advanced computational modelling, looking for hot spots in the combustion which are associated with localised elevations in NOx formation.

For gas combustion, Flue Gas Recirculation can also be employed to significantly reduce flame temperature, whilst maintaining the flame dynamics required for stable combustion. This is because of the reduced oxygen concentration of the flue gas. When atmospheric composition combustion air is replaced with low oxygen flue gas, the effect is to average the oxygen content proportionately. For example, if 90,000 cubic feet of atmospheric air (at 20.9% oxygen) is mixed with 10,000 cubic feet of flue gas with only 2.5 percent oxygen, the result mixture will be 100,000 cubic feet of gas with an oxygen concentration of 19 percent. Reducing the available oxygen has the same effect as a reduction in fuel; it limits the reaction rate. However, flue gas recirculation has further benefits.

The flue gas serves to temper the combustion, as the higher volume of inert gas absorbs heat from the flame.

It is using these mechanisms that RJM has been able to deliver up to 60 percent NOx reduction with the incorporation of flue gas recirculation. RJM’s CleanAir Burners are now achieving emissions performance as low as 0.02 lb/MMBTU incorporating FGR with advanced combustion dynamics.

Enhanced Flame Dynamics for Minimum Load Reduction

The same principles of flame dynamics can also allow for enhanced stability at low load. The critical parameter with regard to flame stability is maintaining the flame front at the burner, to ensure fuel entering the furnace is ignited on entry. Attention to near burner dynamics will allow for enhanced stability ranges and in a recent project has allowed a reduction in minimum stable generation on a 165MWe front wall fired gas unit; from an original MSG of 35 MWe, down to 13 MWe. 

Operationally, though the plant was required to stay online, acting as spinning reserve to provide back-up for wind generation. However, economics dictated that generation be as minimal as possible to conserve fuel consumption.  To accomplish the reduction in MSG, among the issues that required resolution, was ensuring that the burners would remain stable as fuel pressure was lowered. Near burner flames are compromised when either the local mixture falls outside stability limits, or when the fuel/air velocity exceeds local flame speeds. 

“Combustion control in conventionally fired boilers is paramount for maintaining combustion efficiency, emissions performance and boiler integrity.”

For this project, a particularly stable near-burner aerodynamic region was engineered. This was accomplished using RJM’s proprietary technology; which allows for a constant aerodynamic stability region to be created. This region is of nearly constant relative strength across a significant variation of secondary air velocities. The distribution of fuel and air at the burner face actually accommodates an emissions reduction across the load range.

Conclusion

Combustion control in conventionally fired boilers is paramount for maintaining combustion efficiency, emissions performance and boiler integrity. In the absence of comprehensive consideration of all parameters which can affect and be affected by the combustions system, all of these performance parameters are jeopardised. Equally, advanced combustion control can offer emissions reduction at significantly reduced cost over more expensive flue gas treatment processes.

A key consideration in the design of a staged firing system, is the availability of pressure and flow from the existing draught plant. Modifications to burner hardware and the additional pressure required for OFA jet penetration of the flue gas can quickly increase the required pressure beyond that capacity of the existing fans.

As well as the requirement for OFA jet penetration, pressure and flow must be maintained at the burner to ensure flame structure is held at all times. This is particularly difficult during transient conditions when variation in air and fuel flows can move independently of one another. Failure to maintain flame structure produces longer flames and can allow combustion to carry into the convective section, causing irreparable damage to the heat transfer surfaces.

RJM International has demonstrated through multiple successful projects that firing systems can be designed to meet even the most challenging performance requirements, under some of the most challenging conditions. Key to that success is a detailed, comprehensive analysis of the plant and intelligent design to mitigate those limitations.

Authors

James Dennis is lead combustion engineer at RJM International. John Goldring is managing director at RJM International. Lawrence Berg is vice president of Engineering at Systems Analyses and Solutions Inc.

A growing investment in renewable energy and distributed energy resources (DER) is creating opportunities also for ongoing modernization of the grid – and the effort shows no signs of letting up. System upgrades are pressing on worldwide to support the increase in renewable energy while making infrastructure smarter and more resilient.

This move from centralized to distributed power generation is spurring utilities to digitize, reinforcing the need for data collection and analysis to ensure proper operation and maintenance of these often unmanned assets. Along the way, it is increasing the opportunity for intelligent operations and planning.

According to survey results from Black & Veatch’s 2018 Strategic Directions: Smart Cities & Utilities Report, utilities rank improving reliability, integrating DER, and improving both efficiencies and analytics (tied) as the top three challenges facing distribution systems today, underscoring the need for upgrades that will address those issues and more (Figure 1).

To assimilate DERs into the generation mix, utilities are turning to the Internet of Things (IoT), using advanced communications networks, sensors and remote monitoring capabilities to see what’s happening in real time, and respond quicker. Through the integration of communications, technology and data analytics, utilities can make intelligent decisions and use automation to become more efficient. Overall, lessening disruption while enhancing optimization, reliability, security and safety.

UNLOCKING EFFICIENCY WITH DATA

Today, the plummeting costs of embedded sensors encourage increased use. Placing signal devices in a myriad of things — electric meters, field equipment including capacitor bank and recloser controls, and solar and wind controls — directly informs when repair and replacement is needed by allowing operations and maintenance personnel to collect data such as vibrational anomalies or spikes in energy usage or changes in voltage. Sensors also allow vendors to remotely update equipment with warranty data or recall information, allowing for greater control and accessibility.

This proliferation of sensors have advanced to enable utilities to collect massive amounts of data from their renewable installations, gracing operators with a far wider understanding of their systems and chunks of information to create actionable insights. It also opportunistically guards against disruption.

Operators of wind turbines, for instance, have been able to leverage data from those towering machines to prevent component failures and make minor design tweaks that make them more efficient, often at no additional expense. Such departures from traditional applications hold the promise of paring operational costs and furthering sustainability by using data analytics to better understand our environment and drive greater decisions about costs, materials and the amount of energy used – or more importantly, wasted.

Utilities are interested in improving grid reliability and efficiency through the use of automation and a deeper understanding of grid operations. To achieve that, according to the Black & Veatch’s report, one-third of utilities plan to implement advanced metering infrastructure. Thirty percent expect to put in place a distributed energy resources management system, with an equal percentage looking at outage management systems. Twenty-eight percent are considering new modeling tools for distribution planning (Figure 2).

Survey results show that the majority of utilities (80 percent) are planning to add smart assets as a critical component of their repair/replacement programs. Of that number, 41 percent plan to add smart assets in critical areas guided by cost/benefit analysis; 25 percent are looking at deploying smart assets as part of a coordinated data monitoring strategy, linked to asset life cycle plans; and 22 percent say they will implement smart infrastructure projects as existing assets are retired. Some 17 percent are planning for wholesale replacement of existing assets with smart assets.

Twenty percent of respondents are not including smart infrastructure implementation as part of their repair/replacement programs or capital plans.

PUTTING STRATEGY INTO PRACTICE

An emerging trend among utilities operating in the DER space is the narrowing focus on determining the appropriate level of monitoring and the corresponding economic investment. The question becomes: What is the right level of monitoring from a cost and production standpoint?

Let’s look at an example of solar. Typically, plants monitor at the inverter level because inverters are the most common point of on-site failure, and it’s the first level of observation. Inverter data can provide critical performance metrics such as power lost during DC to AC conversion, or fault code information for remote troubleshooting.

At the utility scale, the growing trend is to build large plants using smaller inverters (string inverters). For example, instead of using one 2MW inverter, a plant would use multiple 100kW inverters. Building large plants using smaller inverters offers the ability to communicate on a much smaller level while also delivering operations and maintenance (O&M) and production benefits. Failure on smaller inverters also means smaller losses. However, from the financial perspective, it does not always make sense to repair each individual inverter — rather, it is more economical to repair inverters in batches. The fix has to pay for itself.

Other levels of monitoring take an even closer look. Subarray monitoring on larger systems takes it one step deeper, providing greater visibility to identify issues at the string or module level. Specific subarray or zone monitoring breaks apart an inverter’s array into smaller metered arrays, allowing for easier identification of strings and areas of the array without investing in costly site visits and/or analysis.

String level monitoring offers even more granularity. Used by utility-scale PV plants to detect low production and potential energy loss, string monitoring combines 20 to 30 modules per string. These groups of strings then are combined in “smart combiners,” which monitor each string or pair of strings independently to collect data and send it to the monitoring system for software analysis. This level offers the most detailed level of surveillance and can identify underperforming strings.

Although these levels of monitoring offer minute details, large-scale plants might not need that level of information. A large solar PV plant, for instance, could have 300,000 modules. One individual module is 1/300,000th of the system. So, does it make sense to monitor each individual module? By comparison, small DER installations such as a residential solar system rely on micro inverters that provide input on every single module in the system.

ASSET MANAGEMENT CRITICAL TO INTEGRATION

While more utilities are looking into comprehensive programs that outfit equipment with sensors and remote monitoring systems, this year’s survey results reflect an industry still determining how technology and networks will work together in the modern grid. Many survey respondents were unsure of how all the elements will play together. Many utilities, for example, already have enhanced monitoring in place but are not using it to its full potential.

Data from enhanced monitoring can be used to make informed business decisions; today, only a meager 17 percent of utilities are not using smart monitoring to track asset condition or performance in any capacity (Figure 3).

Planning and preparation will be critical before starting any type of integration. Enhanced or “smart” monitoring still is relatively new to the utility industry, and it will require a firm commitment to asset management efforts.

Asset management programs are designed to align people, processes and technology to improve operations and create cost efficiencies, enhance service levels and extend asset life. Formal asset management efforts use data to inform business decisions on the basis of risk, or the likelihood and impact of potential asset failure. However, data is only useful if it provides the right information to the right people. Data must be actively used to monitor current operations and identify trends.

Although integration is specific to the utility’s needs and generating source, utilities across the board need to ensure that all assets speak the same language and are on a common, compatible platform. For example, when one system uses fiber optics, another goes with Ethernet, and a third RS 485, integration can cost more than the value of the information received. Furthermore, the method and reliability of data communication requires careful consideration.

Before integration, utilities will need to thoroughly understand their assets, and ask themselves some critical questions:

– What is the configuration of the DER system?

– Where is this DER asset located, relative to the electrical distribution and communications systems?

– What are the critical assets?

– How old are the assets?

– What is the condition of the assets?

– How are budgets for these assets best spent?

Once asset management is in place, utilities will need to understand how to analyze the data. Today, new combinations of innovative and traditional technologies such as Black & Veatch’s ASSET360® data analytics platform are giving utilities the ability to manage data with greater planning and operational complexity.

The cloud-based ASSET360® platform captures, integrates and analyzes data from infrastructure systems, assets and devices. By providing this type of analysis, platforms such as this allow utilities to make quick, actionable insights to improve operations and support future planning.

ADDRESSING CHALLENGES

The path to grid modernization is no doubt challenging. The ability to support a high penetration of ubiquitous sensors, automated controls and DER within the current distribution system requires focus on many levels. Survey results show that 49 percent of respondents report system stability as the biggest obstacle, followed by troubles with analyzing DER load flows, business modeling and standardizing interconnections.

A Data Tsunami

It’s well recognized that IoT can drive efficiencies and optimize processes. That said, IoT’s ability to capture huge blasts of data through sensors leads to one major challenge — how does an organization structure that data, and more importantly, distill something usable from what is effectively a “Data Tsunami”?

This is a significant, even daunting proposition, and utilities vary with their level of data integration abilities. Some have developed detailed plans to direct data to specific areas within the utility; others struggle with the concept of using operational data for something other than where it is coming from. For example, data can bring about a turf battle between IT and OT. Operational data (e.g., percent of energy remaining) is collected by battery sensors through SCADA systems. This data must then pass to those responsible for making energy purchase decisions, and then to billing in the IT systems department.

Utilities are in various stages of interconnecting these technologies, but those that master this union and nimbly collect and process vast amounts of data hold the edge. This challenge demands a strategy for structuring information, applying analytics and extracting knowledge to harness data’s value.

The Communications Gap

As these data collection systems evolve and become more efficient, they require reliable, robust communications systems to send information back to the central office or SCADA system. Unfortunately, the move to more advanced grid technologies can be hampered by the capacity and capabilities of current communications systems.

The equipment may have aged and become outdated, making renewable integration, distribution automation or network convergence difficult to implement without more.

Survey results show that more than 58 percent of utilities believe their current communications network is inadequate, while nearly 5 percent of that group admitted not knowing where to start (Figure 4).

Protecting Against Cyber Threats

Utility leaders clearly understand the importance of maintaining security across IT and OT networks. Survey results show that 70 percent of utilities see cybersecurity and physical security as a growing “must have” to protect the influx of wireless technology and other communication devices.

But the rapid evolution of electric grids and communications networks can make it difficult to plan safeguards. As efficiency efforts drive IT and OT systems to converge, hackers can gain access to the OT infrastructure via an IT route, fueling a need for utility leaders to assess, plan and implement protection strategies for critical assets.

This position represents a fundamental shift in approach to security; where security was once set up under IT only or as a separate shop, it is now being integrated into a broader IT/OT function. Asset managers must assess their risks and adopt responsible security measures that are flexible and scalable.

Committing Human and Capital Investments

When looking to upgrade, utilities appear ready to invest the needed capital. While most utilities say they plan to invest less than $50 million in their electric distribution systems over the next three years, a quarter of respondents will spend more than $100 million (Figure 5). In stark contrast, 70 percent of the previous year’s respondents reported they were planning to invest less than $20 million.

A major challenge for utilities is that these new data analytics systems are mandating a new set of skills, requiring a different type of staffing than what was needed in the past. Two-thirds of utilities anticipate a shortage in skilled professional, technical and/or labor resources within the next 10 years.

To combat this, the majority of utilities are working to attract, retain and train new staff (78 percent), use advanced technology to automate operations and processes (54 percent), outsource specific job functions (47 percent), process and/or organizational re-engineering (46 percent), and implement knowledge management systems (42 percent).

PLANNING FOR THE FUTURE

Increasing connectivity and a growing embrace of renewable energy are driving deep change in the utility industry. While there is a growing understanding that “smart” monitoring systems deliver unparalleled levels of scrutiny and increased awareness, utilities must be aware that these networks are best served with a holistic and integrated systems approach, rather than focusing on the separate components of the generation and distribution process.

Although system reliability and efficiency remain the top priorities, utility managers in the U.S. are planning accordingly and reimagining these priorities in a much more distributed paradigm.

Authors

John T A Miller is the solar technology manager for Black & Veatch. He has spent the last 15 years focused on sustainable technology, energy and infrastructure. Doug Young is a project manager and lead engineer with Black & Veatch. His specialties are in SCADA design, automation, protective relaying and electric metering.

Editor’s Note:

This article is based on a paper presented at POWER-GEN International 2017 in Las Vegas, Nevada.

Renewable energy has followed a process common to disruptive technologies. For a variety of reasons, renewable energy has seen popular support sustain policy incentives that have improved the technology, developed markets, led to economies of scale which facilitate private investment and lead to a broad dispersion of the technology. The rapidly falling cost of renewable energy technology and grid integration along with the spread of renewable energy to developing countries are two clear indicators of the maturation of renewable energy.

To procure the growing fleet of renewables, two prominent approaches are reviewed; renewable energy auctions and wholesale power markets combined with capacity and ancillary services markets. The two procurement strategies are compared in the context of the application in varying global contexts. Finally, techniques to facilitate grid integration and increase the overall flexibility of the generation and demand side resource mix are reviewed as a way forward to continue the process of renewable energy technology and grid integration maturation.

Incentives: Setting the Foundation

Renewable energy is an emerging and rapidly maturing technology. Renewable energy has followed a process common to many disruptive technologies. In the early stages of technological development, the support from incentives to catalyze markets and eventually capture economies of scale are critical. The sustainability of incentives through this early stage development hinges on a foundation of popular support and political will.

In the power sector, this popular support is impacted by several variables. One key correlation is level of income (GDP) per capita in a country. Analyses by the International Energy Agency (International Energy Agency, 2017) has found a 99 percent probability of significant correlation between GDP per capita and RE deployment levels. This study has also found a 97 percent probability of significant correlation for net energy importers and relatively high levels of RE deployment.

However, it is important to remember that no variable is perfectly correlated with support for RE incentives, there are always exceptions to the rule. In summary, high income

countries and countries that are net energy importers tend to favor RE and are likely to sustain incentives leading to increased deployment. Other factors such as the presence of a high RE resource capacity (sun and wind resource), the existing power generation mix and its ability to support increased grid flexibility and the presence of an environmental ethic are examples of factors that can lead to sustained incentives to develop RE technology and markets.

Throughout the process of developing a disruptive technology, it is important to structure incentives to realize maximize public benefit at each stage of development. In the earliest stages, where many competing ideas have small probabilities of significant breakthroughs, government often funds the critical role of basic research and development (R&D). As the technology matures, venture capital, which is often supported by a form of risk mitigation, becomes a more important source of finance. As economies of scale develop and technologies approach market parity, public equity and credit markets take an increasing role leading to the eventual commercialization of the technology and often a diminished role for publicly financed incentives. While continued government R&D support is less often associated with this late stage of market deployment, recent literature suggests that continued innovation, as measured through patent activity, is an important factor in continued cost improvements. (Noah Kittner, 2017).

The process of commercializing renewable energy technology has led to one very clear result —the sharply falling cost of RE technology. From 2009 to 2015 the levelized cost of wind energy declined by 61 percent and the cost of utility scale solar declined by 82 percent. (Lazard, 2015). This remarkable cost decline and increased experience with grid integration methods has led to greater diffusion of RE among developing countries.

In 2015, investment in renewable energy was higher in developing economies than in developed countries for the first time. (United Nations Environment, Bloomberg New Energy Finance, Frankfurt School, 2017). In developing countries, renewable energy investment by 2014 had grown by a factor of 17 over the equivalent figure for 2004. A large part of the renewable investment growth in developing countries can be attributed to rapid growth in China, India and Brazil. However, other developing countries also increased RE investment in 2015 by 30 percent in just 2016 to an all-time high, increasing investment over 20014 by a factor of 12. Meanwhile, renewable energy investment in the developed world has declined consistently since the peak in 2011.

The shifting geography of RE investment and deployment is indicative of a later stage in the process of commercializing a disruptive technology. Higher income countries (especially energy importers) have implemented the research and incentive structure to foster technological improvement, develop markets which can support private investment piloted programs for grid integration and flexibility. All of this has led to sharp declines in RE cost for both technology and grid integration procedures. The lower costs and development of flexibility capacity has allowed developing countries to increase their investment in RE while developing countries investment has slowed with technological and market maturation evidenced by decreased incentives.

Investment: Alternative Methods Tailored to Local Circumstances

Myriad techniques have developed to deploy RE investment. These techniques have emerged to best fit widely varying circumstances among utility structures, national goals and consumer preferences. This article focuses on two investment frameworks that have emerged and grown over time; renewable energy auctions and wholesale power markets structured on bid-based short run marginal costs. As a broad generalization (with many exceptions) auctions have been used for new resource procurement more often in relatively integrated utility systems which also tend to occur in developing countries. Unbundled systems in larger developed countries have moved more toward wholesale power markets for all new resource development.

In the past few years, the use of auctions has grown along with increased renewable energy investment in developing countries. The number of countries that have adopted auctions for renewable energy increased from 6 in 2005 to at least 67 by mid-2016. (International Renewable Energy Agency, 2015). RE auctions have several advantages for more bundled and smaller systems in developing countries. Auctions, managed by central regulator, are by their nature amenable to integrated systems. Renewable auctions that result in Power Purchase Agreement (PPA) from a central regulator mitigate one of the chief risks of generation finance in developing countries, off-taker credit risk. Auctions lead to a formalized agreement between parties that can clearly evaluate PPA terms and counter party risk.

In addition, the resulting legally binding contract can minimize the political risk in the event of institutional change. Finally, developing countries are more likely to use energy policy to for development objectives which can be reflected in auction design.

RE auction design can be summarized by four fundamental elements. First, auction demand specifies the overall volume to be procured, specific technology requirements such as a quota for wind or solar PV and the parameters around project size. Qualification requirements set the criteria for bidders. These can include credit and finance requirements, local employment requirements, regional experience in RE deployment and demonstration of site selection and grid access. Sellers’ liabilities include the commitment to contract signing, and the settlement rules and penalties for underperformance or delays. Finally, the selection component includes the winner’s selection criteria, final contract terms and conditions including contract schedule and the payment profile.

In developed countries the growing prominence of wholesale power markets based on short run marginal cost dispatch has created a different set of issues concerning RE. In a wholesale power market prices are driven by marginal cost and demand. There is no contractual agreement that all costs will be recovered in any given dispatch period. As the dispatch order is driven by variable costs, RE with near zero variable cost tends to be dispatched first, putting additional pressure on fixed cost recovery for all generators. This uncertainty over fixed cost recovery can dampen incentives for investment in new generation — an issue exacerbated by the low variable cost and dispatch priority of renewable energy. This phenomenon, also known as the missing money problem, is illustrated in Figure 1.

Figure 1 shows various generating resource types arranged in order by variable cost. Also shown for each resource is the levelized fixed cost. (U.S. Energy Information Administration, 2017). Two simplified price ranges are shown. These prices are for the Pennsylvania-Jersey-Maryland (PJM) territory and show the highest and lowest (On-Peak, Off-Peak) monthly average prices for 2016. Even at the higher monthly average price, only part of the fixed cost are recovered for most generators. Actual prices vary with much greater frequency and range, but the issue of partial recovery of fixed cost is an issue for wholesale power markets that can be exacerbated by renewables.

In addition to capital recovery issues, renewable energy can add pressure on ancillary services due to increased variability affecting the scheduling and pricing of those services. Design and experimentation of capacity and ancillary services markets to address these issues are well under way. Early capacity markets were instituted in the United States in 2006 and 2007. These markets allow generators to bid their services into supplemental markets in a process that share similarities with RE auctions described earlier. Demand for services, qualification requirements, seller’s liabilities are often specified prior to the selection which often focuses on prices after preliminary necessary conditions are met.

Integration: Shifting the Special Case to Business as Usual

During the initial stages of RE deployment, system operators can manage variability issues through conventional methods, such as dispatch of fossil-based load-following generators. As the RE share increases, these conventional methods can become costlier, negating some of the benefits from the new clean resources. Alternatively, the system operator can begin implementing more fundamental reforms towards flexibility.

The California Independent System Operator (CAISO) illustrates this transition from conventional to flexible practices. In 2007, when RE accounted for about 11% of generation, CAISO undertook a significant study to examine the feasibility of a 20% renewable portfolio standard. This report concluded that the 20% level would require “several additions to the operational practice” in order to avoid “significant effects on the market clearing prices and unit commitment costs.” (California Independent System Operator, 2007) Ten years later, following multiple market and operational reforms, CAISO now integrates nearly 30% RE, with goals of 50 percent by 2025 and a further target of 100 percent by 2050 under consideration. In this high-RE future, the CAISO envisions that “Renewables supply an increasing share of Essential Reliability Services, including Primary Frequency Response, regulation, voltage support, and spinning reserves, all of which had previously been supplied by fossil, nuclear, and hydroelectric power.” (California Independent System Operator Corporation, 2017)

These market and operational reforms seek to increase network ‘flexibility’ to better accommodate RE into existing grids. Flexibility is the capability to maintain generation and load balance under uncertainty and is key to integrating RE. Network flexibility is composed of three related dimensions; power range (MW), ramp rate (MW/min) and duration (MWh). These three dimensions are summarized below in Figure 2.

Often some of the most cost-effective procedures involve relatively simple operational adjustments. For example, most North American power markets integrate wind power into their economic dispatch process, allowing the dispatch of wind plants along with conventional power plants. Using wind forecasting can reduce reserve costs. Most regions now achieve wind forecast accuracy in the 90-95 percent range. As another example, U.S. organized markets must now pay resources based on 5 minute intervals, rather than an hourly basis which was an earlier norm. Shorter intervals will create price signals for generators to move up or down based on system needs, encouraging and rewarding flexibility.

“Using wind forecasting can reduce reserve costs. Most regions now achieve wind forecast accuracy in the 90-95 percent range.”

Beyond operational adjustments, changes to certain market structures have also provided a financial signal to motivate supply and demand resources to operate in a manner that facilitates RE integration. As an example, FERC Order No. 755 required system operators to incorporate a resource’s speed and accuracy into a performance-based payment. Prior to FERC Order 755, even though some resources can provide higher performance for frequency regulation they were not compensated for this capability. These new performance-based payments not only created a business case for storage, but created a more efficient means to procure frequency regulation service. Over time, batteries have displaced other resources in providing frequency regulation, and now provide almost half this service in PJM. As a result of these performance-based payments, independent developers built 200 MW of storage in PJM between 2013 and 2015. PJM may have been too successful as the region may now have more storage than necessary for its current frequency regulation requirements.

In the longer-term, updates to network management can have significant impacts on RE integration. A key to improved operation is to facilitate larger balancing areas. This can be accomplished through pooling and other forms on cooperation and integration. Larger balancing areas can help reduce flexibility requirements. The random variations in both loads and variable generation resources smooths out from aggregation and become a proportionally smaller share of the total system demand as the geographic area grows. For example, the Western Energy Imbalance Market (EIM) integrates CAISO dispatch with other western balancing authorities. Since 2014, the EIM has grown from California to include Balancing Areas from 6 other states. After integration, the EIM has realized savings from reduced reserve, dispatch, and curtailment costs, totaling $173 million.

Taken together, these operational improvements help facilitate RE procurement, whether by auction or market. Aggregation of variability across large geographies and reduction of forecast error both help reduce reserve costs, whether allocated to the RE producer or consumer. Steps that reduce RE curtailment improve project economics, which further improves the competitiveness of these new resources. System operators that shift from conventional to innovative practices can maximize the benefits of RE.

Conclusion

Renewable energy has followed a path common to disruptive technologies. This development can be summarized in three phases; incentives which act to spur technological improvement, investment growth in which the bulk of investment shifts increasingly from the public to the private sector and finally more widespread integration through increased grid flexibility. The tangible result of this phased development can be seen in the sharp decline of capital costs, the spread of renewable technology to developing countries and the phasing out of some incentives in developed countries. The combination of incentives, investment and integration acting together over time to lower cost and increase the competitiveness of RE is summarized in Figure 3.

Authors

Robert Anderson is the Director of Energy at Millennium Challenge Corporation. Eric Hsieh is Director of Energy Finance and Incentives Analysis in the Energy Policy and Systems Analysis Office at the U.S. Department of Energy.

A key enabler in integrating large amount of PV generation into the electric power grid, is the capability of utility-scale PV plants to address grid reliability and stability concerns. PV plants with “grid-friendly” features such as voltage regulation, active power controls, ramp rate controls, fault ride-through, frequency droop control and others have alleviated these concerns.

The viability of PV plants to provide important ancillary services to the grid was recently demonstrated in a test conducted with NREL and CAISO on a 300MW utility-scale PV plant. The results showed that the PV plant value can be extended from being simply an energy source to provide services such as spinning reserves, load following, ramping, frequency response, variability smoothing & frequency regulation. The results showed that a PV plant can regulate to 4-second Automated Generator Control (AGC) signal 24-30 points more accurately than even fast gas turbines.

With an increasing share of variable generation on the grid, traditional power generation resources equipped with automatic generation control and automatic voltage regulation controls are being displaced. To support grid stability and reliability, deployment of PV power plants that incorporate advanced grid services capability becomes increasingly essential [1-4]. Recognizing this need, CAISO with support from NREL & First Solar, ran a set of tests on a 300MW utilityscale PV plant to measure its ability to provide ancillary services to the grid.

“Deployment of PV power plants that incorporate advanced grid services capability is becoming increasingly essential.”

Test Results

CAISO published the test results quantifying the performance of a utility-scale PV plant to provide services that range from spinning reserves, load following, voltage support, ramping, frequency response, variability smoothing and frequency regulation to power quality [5]. Specifically, the tests conducted included various forms of active power controls such as automatic generation control (AGC) and frequency regulation, and droop response, as well as reactive power/voltage/power factor controls.

As an example, Figure 1 demonstrates the plant’s nearly perfect ability to follow the CAISO’s four-second AGC dispatch signal within its selected regulation range of 30 MW, or 10 percent of rated plant power. The plant was commanded to curtail its production to a lower level (red trace) that was 30 MW below its available power. The AGC signal was then fed to the power plant controller (blue trace), so the plant output (orange trace) was changing following the set point while the plant output was increasing during the morning period.

The Figure 2 shows the plant’s behavior during the midday period. In both the figures it is quite clear that the plant’s response is very close to the AGC signal demonstrating high level of performance.

CAISO measures the accuracy of a resource’s response to EMS signals during 15-minute intervals by calculating the ratio between the sum of total 4-second set point deviations and the sum of AGC set points. As shown in Figure 3, the solar PV resource performed with much higher accuracy than conventional resources in following 4-second regulation dispatch signals − about 24-30 % points better than fast gas turbine technologies. This performance is a reflection of the underlying fast responding power electronics technology deployed in PV inverters along with well-designed power plant controllers. The data from these tests will be used by the CAISO in future ancillary service market design for determining the resource-specific expected mileage for the purposes of awarding Regulation Up and Regulation Down capacity.

Conclusions

There has been considerable discussion on the capabilities of wind and solar plants to provide ancillary services and a call to action for all new resources to have capabilities to support grid reliability. And to great extent, these changes are well underway. The results described here clearly demonstrate that utility-scale PV plants are not only capable of providing essential reliability services that can enhance system flexibility and reliability without incurring carbon emissions but can perform even better than conventional plants. With ever decreasing cost of solar electricity, the opportunity cost associated with some curtailment to provide services like up regulation will reduce as well.

Authors

Mahesh Morjaria is vice president of PV Systems Development at First Solar. Vladimir Chadliev is director of Global Grid Integration at First Solar. Vahan Gevorgian is principal engineer at the National Renewable Energy Laboratory. Clyde Loutan is principal, Renewable Energy Integration at California ISO.

References

1. North American Electric Reliability Corporation. Integration of Variable Generation Task Force Report (Washington, DC: 2012).

2. North American Electric Reliability Corporation. Essential Reliability Service Task Force Measures Framework Report, December 2015

3. V. Gevorgian, B. O’Neill, Advanced Grid-Friendly Controls Demonstration Project for Utility-Scale PV Power Plants, NREL Technical Report, January 2016, www.nrel.gov/docs/fy16osti/65368.pdf

4. M. Morjaria, D. Anichkov, V. Chadliev, and S. Soni. “A Grid-Friendly Plant.” IEEE Power and Energy Magazine May/June (2014)

5. C. Loutan, M. Morjaria, V. Gevorgian, et al. “Demonstration of Essential Reliability Services by a 300-MW PV Power Plant”, NREL report, March 2015, https://www.nrel.gov/docs/fy17osti/67799.pdf

The Westside Energy Station can generate 47 MW of power for customers in and around Rochester. Photo courtesy: Boldt Corp.

The real power behind the new generation station in Rochester, Minnesota has nothing to do with megawatts and everything to do with collaboration and transparency. The Westside Energy Station can generate 47 MW of power for customers in and around Rochester, and it is a study in teamwork and pre-planning.

When Rochester Public Utilities (RPU) decided to replace its 1949-vintage Silver Lake coal-fired plant, part of the plan was to add a new generation station designed for additional peaking power capacity, increased efficiency, and significantly lower carbon emissions. Design engineering of this new plant began in spring of 2016 and construction began that fall; the $62.6 million facility was fully operational in early 2018. It is the largest power development project ever undertaken by RPU.

“This natural gas power plant will help meet RPU’s capacity obligation with reliability and reduced emissions, and keep prices low by being able to hedge against the risk of volatile electricity prices,” said Jeremy Sutton, Director of Power Resources at RPU.

“We had to develop a level of trust and rapport with the owner and Wartsila.” – Tony Densmore, Boldt

The project was on an aggressive timeline. Engineering firm Sargent & Lundy and construction firm The Boldt Company formed a joint venture called The Westside Energy Partners to fast track planning. The project faced construction issues early on; later crews had to base all planning on the shipment date of five 20V34SG natural gas fired reciprocating internal combustion engines (RICE) supplied by technology group Wärtsilä.”

Before one shovel broke the dirt, the contract between owner, engineer, and construction company was written to meet unique challenges. The agreement fixed prices for design and general conditions on the project, it provided a fixed margin structure for the work, and it provided an open-book approach for the balance of the plant delivery. Under this agreement, the owner allowed flexibility in how work was performed and materials were sourced to best meet the project requirements.

“We had to develop a level of trust and rapport with the owner and Wärtsilä,” said Tony Densmore, Boldt project manager. “We had a budget and within that, incremental savings were used to create contingency funds which were used to provide additional scope to enhance the facility.”

Construction and engineering teams executed the project with maximum transparency. “We had an obligation to provide clarity,” said Nelson Rosado, engineering project manager for Sargent & Lundy. “When we made decisions on design features and materials, we were completely transparent and gave the owner a chance to weigh in.”

Construction and design collaboration: start early

Pre-construction planning began early with weekly meetings between Sargent & Lundy engineering, Boldt construction management, engine manufacturer Wärtsilä, and the owner’s management team leaders. All parties reviewed balance of plant engineering, procurement, constructability issues – and all players made decisions in a transparent environment.

“Early collaboration is key,” Rosado said. “It allows us to clarify the technical issues and incorporate the owner’s requirements for local siting and permitting, sustainability, reliability, and ongoing operation of the facility. At the end, the owner gets what they want and need.”

The collaboration was tested early on when the owner indicated the site was on a floodplain and needed to be elevated and stabilized. Boldt and Sargent & Lundy crews persuaded the owner to start construction six months early in order to meet the end deadline that required stabilizing the earth under equipment foundations and working through a complicated local permitting process. Crews also had to beat oncoming winter weather. Boldt teams raised the site five feet using more than 900 gravel piles to support the weight of the equipment across the power island footprint. Construction closed 2016 with earthwork complete and resumed work in spring of 2017, thus saving the owner time and money.

Construction involved complicated scheduling and critical path planning spearheaded by Boldt/Sargent & Lundy, but was entirely dependent on delivery of the five 150-ton engines. The RICE units are designed to provide quick-start power generation in regions needing energy on short notice and ancillary power services to maintain grid stability. Their flexibility is becoming a high value characteristic for power customers due to the increasing reliance on intermittent renewable energy.

“They have the ability to be at full load extremely quickly and come off-line rapidly.” – Josh Klopp, Wartsila, on the flexibility of reciprocating engines.

“They’re the ultimate, flexible technology,” said Josh Klopp, business development manager for Wärtsilä. “They have the ability to be at full load extremely quickly and come off-line rapidly. Reciprocating engines are designed to be maintained on site and are an excellent complement to renewable energy.” The units installed at RPU can go from sitting idle to providing power to the grid in less than ten minutes, but in other installations engines have achieved full load in two minutes.

The foundations and balance of plant tie-ins had to be completed quickly to set the massive engines when they arrived in June of 2016.

“You need close cooperation between Wärtsilä, the construction company, and the owner to ensure there’s a proper laydown area and proper coordination,” Klopp said. Each engine arrived fully assembled at the site on flatbed transports and supporting equipment for each engine arrived at the site in 12, 40-foot conex shipping containers.

Construction built on transparency

“One of the best aspects of this project was the owner created a culture that allowed us to get the job done as a team,” Densmore said.

It was a tight time frame to get the building completed enough to set the engines. Densmore managed more than 25 subcontractors, but had the flexibility for Boldt crews to self perform most of the construction. Boldt teams did concrete work, erected steel, installed the exhaust system, installed all major equipment systems, and built out the control room and admin building. Key subcontractors installed the electrical systems for the engines, and provided insulation and specialty mechanical tie-ins. Constant review of the work process between Boldt, Sargent & Lundy, and the owner helped save time and money.

“By self-performing work we had fewer subs to manage and flexibility in shifting resources,” Densmore said. “We could identify the work that made the most sense on any one day and juggle workers and schedule tasks to meet our goals without waiting for subcontractors.”

The exacting work of installing the RICE units required adhering the to the tight tolerances of the manufacturer and Wärtsilä technical advisors were on site to oversee all work.

“Allowing a good engineering and construction team to create an open, collaborative environment allows the owner to have input on the overall design and make key decisions during construction,” Rosado said.

The project also has these unique features:

– The structure features a building envelope around the engine structure that reduces sound impact to the adjacent community.

– There is a 60-kilowatt solar array adjacent to the engine building to provide renewable energy to the plant office

Increased collaboration allowed the engineering and construction teams to complete the project two months ahead of schedule and RPU saved more than $2 million through construction cost efficiency. Teams say the project “moved at the speed of construction” primarily because of the transparent, collaborative environment created before any crews moved on site.

Author

Mary Schmidt is an independent writer and communications consultant specializing in construction, manufacturing, and business topics.

By Ravi Krishnan

Iberdrola’s 890-MW Topolobampo II combined cycle plant is scheduled for commercial operation in January 2019. Topolobampo II is a high-efficiency natural gas-fired plant. The plant’s heat recovery steam generator is illustrated. Source: Iberdrola

Significant reforms to Mexico’s Comisión Federal de Electricidad (CFE), the state-owned conglomerate that controls the country’s power generation and transmission infrastructure, began in 2013. The new electricity market mirror global wholesale markets having unlocked the electricity sector to privatization, foreign investment, and foreign party operations through service contracts. Dr. Fatih Birol, executive director of the International Energy Agency (IEA) commented on its 2016 report on Mexico’s energy reform in its Word Energy Outlook series, “This is not a reform, it’s a revolution on an unprecedented scale.” In September 2015, Mexico’s Program for Development of the National Electric System established new market operating rules, particularly with respect to renewables integration.

Today, Mexico’s installed capacity is over 70 GW with about 75% from thermal plants, 19% non-fossil-fuel generation (principally hydropower), followed by nuclear and wind. Supply side generation is dominated by natural gas and use of that fuel continues to grow. CFE generates, distributes, and markets electricity to more than 35 million customers and is growing at the rate of about one million new customers each year.

Electricity demand in Mexico has doubled over the past 20 years. Mexico expects its electricity demand to continue to grow 2.4% per year through 2040 (2014 data baseline) which equates to 85% growth over that time period. Consequently, installed electricity generation must double, from about 70 GW to almost 160 GW in 2040. However, averages can be misleading. According to CFE, “energy demand in the Western region of Mexico will grow at an average of 4.8% annually.” Mexico expects half that capacity growth to be supplied by IPPs building gas-fired combined cycle plants, particularly in the western part of the country.

It’s not surprising that CFE expects natural gas will fuel 44% of all energy additions to Mexico through 2030. According to the IEA, there are currently over 40 private companies now active in gas and electricity projects with independent power producers (IPPs) currently generating around 30% of the nation’s electricity.

Iberdrola Generación México is the largest privately-owned electricity producer in Mexico today, with over 6,000 MW currently in service. Over the past 18 years, Iberdrola has been expanding its investment in Mexico and by 2020 the company expects to have installed more than 10,000 MW of new capacity, which equates to the electricity needs of about 20 million customers. Iberdrola currently has several major new power projects under development in Mexico including three combined cycles, two cogeneration, 325MW of new wind power, and two medium-scale (totaling 270MW) solar energy plants. With these projects, Iberdrola has invested more than US$5 billion into energy facilities in Mexico, with about US$3.5 billion in projects initiated after approval of Mexico’s energy reform in 2013.

Build the Best

CFE awarded Iberdrola a contract for the design, assembly, testing, commissioning, operation and maintenance of the 890-MW Topolobampo II (also called Noroeste) combined cycle plant in April 2016. Iberdrola must also install the power lines and electrical substation in order to interconnect the plant with the Mexican electricity grid. The plant is located midway down the west coast in the municipality of Ahome at the Topolobampo port in Mexico’s Sinaloa state and will provide reliable power to about three million residents.

The report price tag of Topolobampo II is approximately US$400 million and the plant is scheduled to enter commercial service in January 2019. The entire electricity production from the plant will be sold to CFE under a 25-year power purchase agreement with fixed capacity charges.

CFE is installing a $1 Billion underwater transmission line that will cross the Sea of Cortez west to connect Baja California Sur to replace diesel-fueled generation with efficient, combined cycle generation and to provide new renewable capacity under development in Baja California Sur and entry point into the national grid.

Iberdrola subsequently selected the popular Mitsubishi Hitachi Power Systems (MHPS) 501 J-series combustion turbine (CT), nominally rated at 327MW. The plant is arranged as a 2 x 2 x 1 combined cycle. Each CT is coupled to a CMI Energy natural circulation heat recovery steam generation configured as triple-pressure with reheat to produce steam for the plant’s single MHPS TC4F steam turbine. When operated in a one-on-one configuration, the 60-Hz plant will deliver 470MW (gross) at 61.5% thermal efficiency (LHV). When operated two-on-one the plant produces 942.9MW (gross) with a plant thermal efficiency of 61.7% (LHV) Natural gas will be supplied through the El Encino-Topolobampo line extension from Mexico’s north-northwest pipeline system.

Hotter is Better

The HRSG steam conditions are state-of-the-art (see the Table). In comparison, EDF’s Bouchain plant, commissioned in June 2016 in northern France, was recently recognized as the world’s most efficient combined cycle plant with a 62.22% thermal efficiency under ISO conditions. The plant operates at 50Hz. That plant’s CMI Energy HRSG produces design superheat and reheat temperatures of 585C, that is, 20C higher than steam produced by HRSGs connected to the latest generation of F-class CTs, yet 17C lower than Topolobampo II. At these temperatures, the correct metallurgical selections are critical.

The increase in steam temperature is made possible by moving upscale from the conventional choice of SA231 P91, which allows up to 650C at 100,000 operating hours per the ASME Code. However, at these pressures and temperatures tube wall thickness must increase due to maximum allowable stress limitations.

The tradeoff is thicker wall tubes are more susceptible to damage from thermal cycling. Austenitic stainless steel tubes are commonly found in ultra-supercritical steam boilers but use in HRSGs subject to cycling has been a concern. CMI Energy, in collaboration with research institutions in Germany and Belgium and steel suppliers have thoroughly examined the use of austenitic stainless steels in high temperature and pressure HRSG applications, particularly with respect to maximum allowable stress, thermal expansion, cycle fatigue, and weldability (including dissimilar welds).

Dissimilar welds are required when joining ferrous alloy steel headers with the austenitic stainless steel tubes.

Welding of dissimilar materials is necessary in high temperature steam applications, either on tubes upstream of the header or on connector downstream of the collector (Figure 1).

Each method has been successfully used in practice. Welding connectors (6-inch diameter) was used on the Bouchain project in close communication with GE. For the Topolobampo II project, CMI elected to weld tubes to P91 header tube stubs while keeping the header in P91 material.

“The IEA states that to accomplish Mexico’s energy reform agenda, investments of as much at $240 billion will be required through 2040…”

Welding dissimilar metal tubes was preferable on this project because there is less thermal expansion, weld stresses are lower, and the process is less expensive because there is reduced usage of expensive Incoloy transition pieces.

CMI Energy selected austenitic stainless steel SA213 S30432 tubes (commercially marketed as Super 304H) for the first few rows of the superheater and reheater tube sheets where the temperatures are the highest to take advantage of its increased resistance to stress corrosion and stress relaxation cracking and because of technical concerns related to steam oxidation above 605C with P91. These headers and critical dissimilar metal welding were performed at the CMI Welding Expertise Center in Belgium. The completed headers were then shipped to CMI Energy’s Mexican subsidiary where the superheater and reheater fabrication was completed.

The HRSG design is modular, not unlike that of most combined cycle plant designs. The modular nature of the design speeds installation and commissioning of the completed HRSG on the job site. Each Topolobampo II HRSG consists of 12 prefabricated modules that are arranged in a three wide by four deep configuration. Each HRSG has 11,575 tubes totaling 2,173 meters long. The heat transfer surface is 330,000 square meters.

Made in Mexico

With Mexican energy reform measures firmly in place, indigenous gas supply and transmission facilities under construction, and a world-leading energy growth rate predicted, Mexico has become a very attractive country in which to do business.

Pedro Joaquin Coldwell, Mexico’s Secretariat of Energy, has likened the country’s power sector reform to an “economic competitiveness reform” where access to adequate and reliable electricity will give Mexican businesses a competitive advantage that will increase jobs and overall economic growth.

The IEA states that to accomplish Mexico’s energy reform agenda, investments of as much as $240 billion will be required through 2040 plus another $130 billion investment in energy efficiency.

CMI has long experience with HRSGs with installations across the globe, including over 20 combined cycle projects in North America alone, many with multiple HRSGs, during the past decade alone.

The Mexican power market is one of CMI Energy’s largest customers, with 15 HRSGs supplied on six projects totaling over 4GW of capacity in recent history, the latest of which is Topolobampo II. Located 30 km north of Monterrey, CMI Energy’s Mexican subsidiary, CMI Energy Mexico (CEM) has been fabricating pressure parts, structures, stacks, and providing inspection and repair services since 2005.

In 2014, CMI Energy invested in an expansion of the fabrication facilities of CEM to meet the local content requirements of the North American Free Trade Agreement.

In September 2017, the HRSGs for Topolobampo II were completed by CEM, the company’s first HRSGs that were 100% manufactured in Mexico. CEM has since increased its fabrication capacity to 4 to 6 HRSGs every year, each consisting of approximately 10,000 tubes. The shop is fully ASME certified (R, PP, U, S).

Mexican energy market reforms promise to dramatically increase use of natural gas-fired combined cycle plants.

From a strategic perspective, the largest HRSG market in North American will remain in Mexico and the U.S. for the foreseeable future. From a geographic perspective, CEM has the advantage.

The company’s fabrication plant is centrally located to the U.S. (only 125 miles from the southern border of the U.S.), Canadian, and Latin America markets.

The National Renewable Energy Laboratory released a video demonstrating its wind turbine blade testing capabilities.

The facility conducts structural research with servo-hydraulic equipment and data acquisition systems to ensure blades live up to the International Electrotechnical Commission 61400-23 standard. The facility can validate blades and other components as small as one meter to more than 50 meters in length.

NREL can evaluate structural properties such as shape, mass and inertial and stiffness properties. The facility can measure static strength by applying quasi-static bending movements to validate design parameters, including extreme loads applied by cranes, hydraulic actuators and servo-electric winches.

Additionally, the facility conducts fatigue research by applying millions of cycles of fatigue loads either single-axis by direct load application or biaxial loading with flapwise and lead-lag directions simultaneously. Use this link to view the video: https://youtu.be/rEmJBAWKYJo

The January 25, 2018 EPA memo on “Reclassification of Major Sources as Area Sources” significantly changes an Environmental Protection Agency (EPA) policy in place since 1995. The EPA and State Agencies have been operating on a Once-In-Always-In (OIAI) policy for emissions of hazardous air pollutants (HAPs) for the past 23 years.

OIAI established that once a source crossed the major-source thresholds for HAPs (10 tons per year for a single HAP or 25 tons per year for combined HAPs) and was required to comply with the Maximum Achievable Control Technology (MACT) standards, it would never be able to become an area source (i.e. a minor source) for HAPs by reducing the facilities potential to emit for HAPs. The EPA was concerned that by allowing a facility to take an enforceable limit for their PTE and to become an area source for HAPs, sources would then limit their potential to emit via control technologies that were less-stringent than the MACT standard. The EPA regulators were thought this would cause an increase in HAP emissions and the emission reductions mandated by congress would not be achieved.

In January of 2018, the EPA issued updated guidance that restored the “plain language” reading the statute and rescinded the OIAI policy across the country. Today, a facility or source at any time can take an enforceable limit in a permit to reduce the PTE below the 10/25 tpy for single/combined HAPs and be reclassified as an area source that is no longer required to meet the facility’s MACT limitations or the industry-specific NESHAP limits.

This reclassification helps many facilities that have lowered their HAP emissions due to evolution in process and throughput. For example, a decade ago, a facility that operates several painting booths projected that their 12-month rolling emissions of toluene would exceeded the 10 tpy threshold. Because of this projection, the facility was classified as a major HAP sources and was then subject to the industry-specific NESHAP standard, as a part of the MACT determination. The facility’s 12-month rolling emissions of toluene only exceeded the single HAP threshold for 3 months. Since this time, the facility has begun using a reformulated paint with a lower HAP content and has reduced production at that site. The actual HAPs for the source is now around 1 tpy toluene. Under OIAI, they were not able to get an enforceable permit limit in order to be released from the NESHAP. Since OIAI has been revoked, the facility can now obtain a permit limitation for 10/25 tpy for a single/combined HAP, eliminating the recording keeping and documentation requirements for the facility.

“If your facility is classified as a major source of HAPs, but your actual emissions are below 10/25 tpy, you should re-evaluate your current permits to see if your facility could benefit from this change in policy.”

Similarly, converting fuels can cause operational changes that can reduce the emission of HAPs, for example a coal-fired boiler that was repowered to burn only natural gas. Combusting coal, the boiler was a major HAP source and subject to MACT and NESHAP standards. After switching to natural gas, HAP emissions are well below the HAP tpy thresholds. Under this new policy, the facility can request a site-wide enforceable limit of 10/25 tpy for single/combined HAPs. This will remove the NESHAP compliance requirements the facility was required to maintain that were only relevant when combusting coal.

Environmental groups have argued that HAP emissions will increase now that OIAI has been revoked. However, OIAI never prevented an area HAP source from increasing the source PTE to become a major source, and facilities that are redefined will be legally required to maintain HAP emissions at or below thresholds for area sources. On the other hand, industry claimed that OIAI policy discouraged facilities from voluntarily making changes to their processes that would reduce HAP emissions (e.g. installing emission controls, switching to a solvent with a reduced HAP content, implementing innovative technologies or processes, etc.), and it is probable that many facilities will now be applying for construction or operating permits that incorporate many of these changes, while reducing the recordkeeping requirements for the sites.

If your facility is classified as a major source of HAPs, but your actual emissions are below 10/25 tpy, you should re-evaluate your current permits to see if your facility could benefit from this change in policy.