Cybernetics can offer both energy and cost savings in CHP plants Credit: Alstom


The Anaconda project in Borgaro Torinese, Italy, demonstrates how forecast and optimization can be integrated into SCADA systems at CHP plants to improve energy and cost efficiency, report Faddy Ardian and Francesco Vallone.

The idea of energy efficiency has been around for some time but has only recently assumed a central position on the political agenda and in the marketplace. The current decade has witnessed a growth in initiatives to find a new paradigm to cut primary energy consumption, reduce anthropic pollutants and raise energy production from renewables sources or fuels such as natural gas.

Many small and medium-sized power producers that use combined heat and power (CHP) have been challenged in generating power, while raising their energy efficiency so that the cascade heat, distributed by a district heating network, can serve a near-by community. At the same time, they must leverage electronics and modern communication tools to provide an intrinsic and independent intelligence to the system. It is also important for these producers to consider technical and economic constraints in running their power plants.

Unfortunately, in many such companies, these electronics and communication tools are yet to be fully exploited to provide intelligence to the system. This challenge has driven the idea of applying ‘cybernetics’ as a way to overcome the current energy issues. It enhances communication between power plants and end-users by means of external variables and real-time feedback via communication tools and electronics in the power plant system. All together it can cut costs and emissions, and simultaneously raise energy efficiency.

The following is an example of a successful project conducted by the Cogenpower R&D department that has led to the development of a cybernetic CHP power plant that is able to raise energy efficiency, as well as cost efficiency.

The Anaconda project

Cybernetics is a broad field of study, but the essential objective is to understand and define the functions and processes of systems that have goals and that participate in circular, causal chains, which move from action to sensing to comparison with a desired goal and again to action.

In engineering, these systems are also well-known as ‘feedback error-based’ systems. By definition, the cybernetic CHP power plant in this project is a system that has been set up to automatically reach its goal through collecting external data (feedback), such as local temperature or electricity price, then making real-time adjustments based on feedback. The system has to become a smart power system that can run or switch itself off according to external inputs.

In brief, it is our opinion that the current challenge in the cogeneration world is going to be overcome through cybernetic CHP power plants.

The project has been named Anaconda and the system is comprised of a natural gas-fuelled 3 MWe CHP power plant (thermal power of 3 MWth), three back-up boilers to meet peak demand, with a maximum thermal power of 4 MWth for each boiler, a 15 km district heating pipeline enriched with a fibre-optic network and heat-accumulation facility, 42 heat exchangers substations, an interconnected system of electronic devices able to constantly exchange data between the production site and the consumption sites, and a SCADA system to control all the elements.

Data exchange through the fibre-optic network has the objective of maximizing the energy efficiency of the whole system, creating a truly Smart Grid. These systems have been continuously operated by Cogenpower to provide heat and electricity to buildings in the Italian municipality of Borgaro Torinese. But difficulties have arisen in recent years, particularly in the operational field because it was not always obvious how to run the power plant at full efficiency. This is quite typical for other small and medium-sized independent power projects that use CHP.

The key challenges can be summarized as follows:

  • Decide hourly, throughout the year, the right amount of energy to be produced and delivered both to the power grid and to the district heating network, based upon economic input coming both from the Italian Power Exchange (IPEX) and from heat tariffs, other than green economy incentives such as CV-TLR and penalties relating to carbon dioxide (CO2) allowances.
  • Manage the heat consumption profile according to the weather, and independently decide daily consumption patterns.

Key operational goals

The main idea of this project has always been to provide the system with enough intelligence to make the right decision about energy production.

Cogenpower therefore enabled the SCADA unit to capture the weather forecast at least one day ahead and to combine this information with the consumption pattern to predict heat demand and, based on mathematical models, make decisions on how much heat to produce using computer algorithms.

No final unique parameter has been established to determine the behaviour of the power plant because the number of hours the CHP engine should produce electricity depends on the spot market price of the day-ahead IPEX in a complex way, requiring optimization of several parameters.

Further constraints include the limitation of the CHP unit to one start per day to avoid shortening the average life of the machine. Such constraints had to be balanced with the heat power consumption and the penalties incurred by the introduction of CO2 into the atmosphere by CHP power plant, as well as gas consumption.

Furthermore, quotas of CO2 are daily exchanged on various exchanges or by means of bilateral contracts between market operators. One of the original ideas, later discarded, was to monitor, together with the power spot prices of the day-ahead market, the CO2 quotas on some exchange. It was discarded because estimates of the desirability of producing heat with the CHP unit or with alternative boilers were skewed towards the CHP unit by the CV-TLR.

The CV-TLR are environmental certificates, also known as ‘green certificates’, awarded to CHP plants with district heating networks, initially as a government incentive to promote CHP power plants that can use district heating networks to cut pollution in cities. These incentives grow with the amount of heat the CHP unit produces and delivers through the district heating network to the end-users. Unfortunately, it was an error to call them green certificates, as usually the fuel used in these plants is natural gas, not a renewable source.

In any case, subsequent legislattion has meant the phase out of these incentives by 2016.

To overcome these operational challenges, Cogenpower aims to have a well-prepared production plan for the next few hours incorporated into SCADA at all times. SCADA will then run the power plant accordingly. It is therefore necessary to have the SCADA system as the main controlling unit, with mathematical algorithms that can automatically optimize the production plan and the configuration of the power plant system.

The algorithms consist of forecasting algorithms to predict heat demand and electricity price, alongside optimization algorithms to decide the best production plan according to current data. These two sets of algorithms provide the backbone of auto-optimization in the SCADA system. With the objective of raising energy efficiency, this auto-optimization is Cogenpower’s solution, leading to the Cybernetic CHP power plant concept.

Hence, the main key objectives are:

  • Building and integrating forecasting algorithms for electricity price and heat demand;
  • Building and integrating optimization algorithms.

The forecasting algorithm

The first essential algorithm must forecast future electricity price and heat demand, and decide, based on an additional algorithm, whether or not to run the CHP and other elements in the system.

It is important to note that the electricity price would not be a deciding variable if a bilateral contract on a fixed price was applied. However, in a liberalized electricity market environment, different hours can have different prices on the spot market, as in Cogenpower’s case where the sale price is based on the IPEX spot price. Prices at certain times under marginal electricity generation costs are unfavourable, unless other variables such as incentives and peak heat demand make it profitable to turn the engines on.

As a result, an algorithm to forecast electricity spot price has be incorporated into the SCADA system. Hence, another mathematical algorithm is needed to forecast this particular variable. As a consequence, well-known statistical models such as SARIMA (Seasonal Auto Regressive Integrated Moving Average) are used to elaborate these mathematical algorithms.

Even without auto-optimization, the upper closed loop system basically established a cybernetic CHP power plant. But the loop was aimed only at stabilizing and controlling the power plant. Without incorporating auto-optimization, the only way to increase energy and cost efficiency is through manual changes by operators, who would have to make judgments based solely on experience and knowledge – and not always with good preparation.

Algorithms that build auto-optimization therefore have considerable added value in raising both energy and cost efficiency. The two main added values from integrating the algorithms are:

  • Higher overall energy and cost efficiency through continuous optimization;
  • Fewer possibilities for human error in optimization decisions.

To sum up the proposed cybernetic system, the forecast algorithm will detect the future changes, and then the optimization algorithm will decide anticipatory actions based on forecasted changes, while the controller will balance (stabilize) the real time operations. Then, together, they create a communication link between the power plant and end-users to create a more compact cybernetic CHP power plant that can raise the energy efficiency and cost efficiency. As a result, the Cybernetic CHP power plant can successfully achieve these key performances:

  • Lower cost, with a total saving of 25%;
  • Optimized emissions, with a 3% reduction in CO2 emissions;
  • Optimized fuel consumption, with a reduction of 3%.

The future of cybernetic CHP

Cogenpower has reached the state of the art where a cybernetic CHP power plant is used to raise energy efficiencies and cut emissions merely through algorithms. The incorporated algorithms have achieved auto-optimization in the power plant system, helping Cogenpower overcome its operational issues. The objectives of the project have been successfully reached but more development, such as bottom-up calculation on forecasting algorithms, is needed.

However, it is expected that future cybernetic CHP power plants will have greater intelligence that enables them to operate all the time in the most efficient way, thus creating a true Smart Grid in distributing energy such as heat and power. Clearly, Cogenpower is in the process of reaching this stage through integrating these algorithms in the SCADA system. Future research is now being planned to expand this idea along with many other technologies to support it.

One approach is to integrate AI (artificial intelligence) in the SCADA system instead of the logical algorithm it currently incorporates. Another idea is to integrate batteries or energy storage to store electricity to prevent energy losses and reduce costs in the system. In any case, these efforts are opening the door for new challenges that will lead to future cybernetic CHP power plants that can tackle current world energy problems.

Faddy Ardian is an intern student at Cogenpower Spa and received a Beasiswa Unggulan Scholarship from the Indonesian Ministry of Education.

Francesco Vallone is the President and CEO of Cogenpower Spa;

More COSPP Articles
Past COSPP Articles