A state machine approach in modelling the heating process of a building

Jouko Pakanen (Corresponding Author), Sami Karjalainen

    Research output: Contribution to journalArticleScientificpeer-review

    1 Citation (Scopus)


    Process models and their applications have gradually become an integral part of the design, maintenance and automation of modern buildings. The following state machine model outlines a new approach in this area. The heating power described by the model is based on the recent inputs as well as on the past inputs and outputs of the process, thus also representing the states of the system. Identifying the model means collecting, assorting and storing observations, but also effectively utilizing their inherent relationships and nearest neighbours. The last aspect enables to create a uniform set of data, which forms the characteristic, dynamic behaviour of the HVAC process. The state machine model is non-parametric and needs no sophisticated algorithm for identification. It is therefore suitable for small microprocessor devices equipped with a larger memory capacity. The first test runs, performed in a simulated environment, were encouraging and showed good prediction capability.
    Original languageEnglish
    Pages (from-to)528-533
    Number of pages6
    JournalEnergy and Buildings
    Issue number5
    Publication statusPublished - 2009
    MoE publication typeA1 Journal article-refereed


    • Buildings
    • control
    • heating
    • identification
    • memory
    • models
    • nearest neighbour
    • state machine


    Dive into the research topics of 'A state machine approach in modelling the heating process of a building'. Together they form a unique fingerprint.

    Cite this