Print this article    Email this article

High-level control
Asger Danielsen
Optimizing the efficiency of WTE plants

The more stable the operation in a waste-to-energy (WTE) facility, the more able the plant is to process waste and produce energy, with less wear on process equipment and less environmental load. Implementation of advanced control measures is an important step in the evolution of WTE plants, says Asger Danielsen

The key to increasing the efficiency of waste-to-energy (WTE) plants is to obtain more stable operation, independent of the actual waste quality, boiler cleaning status and other bottlenecks that appear during the various process stages. The goal is to obtain the highest possible waste throughput and energy output while minimizing the environmental impact.

To exemplify this point, experience shows that for each 10% improvement in process stability, energy production goes up by 2%—4%, and emissions are reduced by 8%—12%. And if the furnace temperature varies, then removal of HCl and SO₂ will be considerably less efficient and the expenses on lime will be unnecessarily high.

At most WTE plants only a few very experienced operators can handle the complexities of constantly obtaining optimal operation efficiency. Most operators consider it a very demanding and troublesome task.

A black-box approach adds to theproblem

Most WTE plants consist of a complex arrangement of different technologies from independent suppliers, each keeping its own specialties about optimal operation as proprietary know-how in a ‘black box’ setup. As a result, the WTE plant operators are left with limited options for keeping the overall plant optimized.

Loop controllers are common features at a WTE plant, but standard loop controllers make control actions at short intervals, which can be too frequent for the long delays (long time constants) often seen in WTE combustion systems.

Some control systems come under the banner of automatic combustion control (ACC). Unfortunately, only a few of these systems cope with the full complexity of the problems. Very often these systems are just another ‘ black box’ hiding the details of the optimization activities. These systems tend to be inefficient when modification is needed to cope with the changing environment and new ideas for more efficient operation at the WTE plant. Consequently such systems can be expensive or impossible to modify and might require additional advanced instrumentation that is difficult to keep in operation in the rough environment of a WTE plant.

In these circumstances, a high-level control system that operates the plant in the same way as the most experienced operator can be a good option, and it can do the work to the same standards 24 hours/day.

The waste incinerator control challenge

Various conventional methods are normally used for implementation of automatic control of the incinerator in the control system. Functions between process measurements like temperature, steam production and oxygen may be used to generate set points for the amount of waste and for the airflows. Normally rather narrow intervals are defined, which prevents the conventional control scheme from generating inappropriate waste and airflow set points. The result is that very often the conventional control scheme ends up at one of the limits, after which control has to be taken over by the furnace operator. Another main problem with conventional control is that the dynamic behaviour of the control strategy is implemented by PID controllers. (PID controllers describe the proportional, integral and differential loop control techniques applied in conventional control systems for — for example — steam production control.)

The operation of a waste incinerator is characterized by:

1. Very different and fast-varying operation of the furnace, caused by the fluctuations in waste composition
2. Long time delays and time constants
3. Varying process dynamics depending on the actual state of the furnace.

Click here to enlarge image Figure 1. FuzEvent control versus PID dynamics

A control strategy based on PID dynamics is not the best way to cope with the above-mentioned process characteristics of a waste incinerator. To optimize the operation of a waste incinerator, the control strategy must hold the following features:

1. To cope with the fluctuations in waste composition, the control strategy must be able to react fast, but without overreacting
2. The control strategy must be able to make efficient adjustments, and then wait until the process has responded. In addition, the control algorithm must consider previous control actions when evaluating the next adjustment.
3. The control algorithm must adapt its dynamic behaviour to the actual state of the process. If, for instance, the furnace operates at reduced production due to lack of waste, low energy needs or a dirty boiler surface, the control dynamic must be adjusted accordingly to cope with the new process behaviour.

above Typically for Danish waste-to-energy operations, the plant at Aars is well integrated into the city’s infrastructure

Based on decades of practical work with control of waste incinerators, Denmark-based Dublix Engineering A/S has developed a control scheme that fulfils the above-mentioned requirements for the optimization of waste incinerators.

The main principles of the Dublix controlscheme

1. The control strategy is based on rules that are derived from the expertise of process engineers and process operators.
2. The control dynamics are derived from manual control of the process in terms of waiting for the process to respond to previous adjustments, and by consideration of the previous actions.

left to right Like most waste-to-energy plant in Denmark, the Aars plant is close to the city centre and supplies district heating The new optimization project at Core, Sesto san Giovanni, Italy

Figure 1 (above) illustrates the difference in dynamic behaviour between normal PID control and the Dublix control algorithm implemented in the FuzEvent System. (FuzEvent is a registered product name that combines fuzzy logic methods with other event-driven mechanisms in a high-level control system. Fuzzy logic has the advantage that the solution to the problem can be cast in terms that human operators understand, so that their experience can be used in the design of the controller. This makes it easier to mechanize tasks that are already successfully performed by humans.) Like the experienced operator, the FuzEvent system is set to operate in due time and when the process behaviour has shown maximum information about its previous behaviour.

Implementation of the Dublix control strategy

The software package, named SoftWtE, for configuration of the Dublix Waste Control Algorithms runs under Windows NT or newer versions of Windows. The package includes an OPC client, which facilitates communication with the OPC server of the existing control system. OPC is a standardized communication system available in any modern DCS control system. SoftWtE may run in a server or client PC of the existing control system, or it may be implemented in a separate computer, which communicates with the OPC server via the plant control Ethernet.

Implementation of the FuzEvent system

At AVR Afvalverwerkung B.V. in Rozenburg, the Netherlands, the FuzEvent optimization system is implemented on one of the plant’ s seven waste incineration lines. The operation has been improved with 25% more stable steam production.

FuzEvent has, in this case, been installed in a separate Siemens WinCC server, which includes an OPC server through which FuzEvent can communicate with the process database. The SoftWtE software package is an open system where the user has access to all control parameters and to all calculations and control logic.

The control strategy is composed of individual software components named EventX. Behind each EventX component is an open script language, which enables the user to configure the calculations and the logic of the component.

Click here to enlarge image

The EventX components may enter a priority system, which makes it easy to cope with conflicting control objectives. For a waste incinerator, the so-called Priority Management System is used to switch between different control schemes depending on the actual process situation. If an abnormal situation occurs, for instance because of waste that is difficult to burn, then SoftWtE automatically activates the higher priority EventX components, which are designed to cope with this abnormal situation, and the Priority Management System deactivates the lower priority components for control under normal operation conditions.

left to right The crane at the Aars plant can be seen in action from the control room Inside the furnace at the Core plant, Sesto san Giovanni, Italy Crane operator in the Aars control room

The EventX philosophy is to detect automatically special or abnormal situations, but an EventX can also be activated by an operator-driven action, if the operator detects a special situation, such as unburnt slag or difficult new waste coming in. In this case the operator interacts with the system and his observations are used in order to bring the system in control of the new situation as quickly as possible.

The details in the operator interaction are defined during the design meeting, and will vary from plant to plant; the main aims are to keep the operators in charge of the operation, make the work done by operators much more efficient and to reduce routine interactions.

Practical implementation of this system took less than four months along the following schedule:

1. Feasibility meeting; two to four days at the plant in order to define and qualify the precise quarantines and success criteria of the optimized implementation.
2. Design meeting: one week defining the control strategy for the plant.
3. Plant management and operators’ discussion about the best possible operation took place during the design meeting and the operator teams and management and specialists had to agree on the best possible operation of the plant.
4. The optimization system operated the plant online covering approx 50% of the set-points connected to the FuzEvent system in automatic mode within three to four weeks.
5. During the remaining implementation period further parts of the plant were brought into control from the high-level control system.
6. Following the implementation, the AVR management requested two different performance test periods. These performance tests focused on the steam stability at two different steam production set-points.

Conclusions

In conclusion, it is evident that any WTE plant can benefit from high-level control, and that extensive production efficiency can be obtained; typically 20%—40% reduction in the variation of the steam flow often resulting in 3%—5% increase of steam flow set-point and steam production. But it is important to note that the process of implementing high-level control must be closely linked to organizational aspects. For example, the implementation team needs to address and solve any ‘natural’ conflicts between the different groups, such as operators, specialists, technology suppliers and the automation teams.

Asger Danielsen is Director of Dublix Engineering A/S, Denmark

---

---