| Current Operational Demands on Coal Units |
 |
Ray Johnson
VP, Business Development NeuCo, Inc. |
Wednesday, October 14, 2009
Coal-fired generating units are currently operating in ways that have never been experienced by the industry before or at least not for many years. The economic downturn has lowered demand significantly for many of our customers. Natural gas prices have fallen making many gas-fired plants competitive or lower cost than coal-fired generation. In several areas of the US, installation of wind turbines has resulted in previously base-loaded coal plants running at low loads for long periods of time. When all these issues are combined, the operations dynamics vary dramatically.
Recently I was talking with one of our customers and the two large units at his station had been off line for three weeks due to lack of demand. These units are in a central location on the grid and have normally experienced capacity factors greater than 85% year after year. This customer is located in the eastern third of the US and the stated cause of this downtime was all three of the factors stated above. Most people would not expect wind turbines to be a significant factor in this region, but due to the growth of available wind power in the last year the varying load has become a significant factor in satisfying the current demand.
When we talked, the units were running but at loads significantly lower than normal. The dispatch group was using all the power available from the wind turbines, running the combined cycle plants at high loads, and using these coal units to make up any shortfalls in demand not supplied by the other two sources. Under this scenario, efficient operation at very low loads and ramp rates become very important, much more so than under more traditional methods of operations. The most surprising fact was that at night the units were generating only 50 mw. The units cannot run at that low load. To accomplish this level of generation, the units were making steam for approximately 135 mw and bypassing the turbine with the excess. This resulted in the incremental cost of generation for those 50 mw very high.
This customer uses CombustionOpt for the typical improvements achieved with combustion optimization systems: NOx reduction, heat rate improvements, O2 balancing, megawatt maximization, steam temperature control, and others. However, the systems were configured to operate at normal load ranges of about 50% to 100% load. Due to the amount of time the unit is running below the normal minimum load, the customer is interested in modifying the CombustionOpt configuration to assist operations with the difficult task of operating at these very low loads. There are only a small number of combustion optimization systems that have been configured to assist operations in this manner. The reason being is that few coal-fired units have operated in this manner in the past. That said, there is no technical limitation on CombustionOpt that precludes the system from running at these load loads. As long as the DCS is run in automatic, the closed-loop features of CombustionOpt are just as applicable at 50 MW as they are at 100% load.
This month we are meeting with the customer to define the requirements, constraints, and limitations of the units while operating in this mode. NeuCo engineers, working with customer personnel, will modify the optimizer to help operators with this very difficult task.
CombustionOpt, being a closed-loop system, will provide assistance and reduce the operator workload for this low-load scenario just as it does for high load operations. This additional work will be covered under the Annual Maintenance and Support agreement in place for these units so the customer will not have to pay additional fees for this modification. Situations like these are one of the reasons NeuCo’s customers continue to achieve benefits from the optimizers for years after initial installation. As requirements change, NeuCo is there to help the customer change with the times.
|
Post Comment | Forward to a Friend
|
 |
 |
 |
At 3:43 PM on Friday, 10/16/2009, Peter Spinney said... |
 |
 |
Ray, I am also getting a lot of questions about how our optimizers can adapt to and even help to mitigate some of the pressures resulting from the new “low-load paradigm”.
While the particular processes for which better control is helpful vary: e.g. fan amps for limits at the high end, pulverizer stability (especially for ball tube mills) or drum level control at the lower end, the nature of the process control challenges is essentially the same.
A few recent examples of how we’ve worked with a customer to adapt the optimizers when their units began to ramp more and run at lower loads, include:
--To ensure SootOpt doesn’t over-clean at low loads due to chasing unachievable RH steam temperature targets, an “achievable curve” was created as a function of load that could also be used as a feedforward input in reheat temperature control.
--Because mill starts and stops were becoming more frequent, we created individual models for different mill combinations and set up a mechanism to swap models automatically based on the active mill combination
--Supplementing CombustionOpt’s neural networks and expert rules with the dynamic control capabilities of model predictive control (MPC).
So far I think we have only scratched the surface of what can be done in these areas – for instance with the use of MPC for tighter control and greater stability (whether at the upper or lower limits of the load range for a particular unit.) The biggest challenge for us in expanding the envelope around these benefits likely lies in the customer's willingness to try advanced multivariate control for these problems as opposed to existing univariate PID methods.
|
|
 |
 |
|
|