Skip to main content

Simulation Enhanced Integrated Systems for Model-based Intelligent Control(s)

Final Report Summary - EINSTEIN (Simulation Enhanced Integrated Systems for Model-based Intelligent Control(s))

The objectives of the EINSTEIN project is to formulate and integrate a number of state-of-the-art building control strategies to test their effect on improving the performance of buildings beyond traditional control approaches typically employed.

The three control strategies developed in the project relate to:
1. Fault Detection and Diagnosis (FDD)
2. Building Performance Prediction
3. Building Performance Optimisation

While Fault Detection and Diagnosis are generally the first step in “correcting” issues in a building, the second and third algorithm form part of what is described as a Model Predictive Control (MPC) solution, which essentially predicts and dynamically optimises the building performance beyond typical rule-based state-or-the-art control.

To implement an MPC control strategy a Reduced Order Model which is representative of the building’s key performance parameters needs to be incorporated. The ROM is developed and used in conjunction with the algorithms in real-time to continuously predict and optimise the dynamic performance of the building and its systems to meet a set of objectives and constraints. These objectives and constraints can be multi-faceted, such as minimise energy cost while maintaining comfort based on the predicted performance of the building over the next number of hours. In this way, the MPC approach is predictive and pro-active in managing the building’s performance, as opposed to re-active control found in current building control practices, such as proportional control.

The key objectives of the EINSTEIN project was to research, develop and test FDD and MPC algorithms to determine any benefit over traditional control strategies. The algorithms developed in the project are tested on both data acquired from real buildings or models for FDD, and on Advanced Calibrated Models (ACMs) based on existing buildings for MPC (ACMs are full dynamic building simulation models developed in IES’s IESVE building simulation software, which are calibrated to ensure the models accurately effect the real performance of the building they represent).

To achieve these EINSTEIN project objectives, work was carried out in several areas:
• Training was required for all seconded and recruited fellows for the project, which focused on building modelling using the IESVE, MPC control strategies, fault detection and diagnosis. This training was carried by the EINSTEIN fellows mainly during their secondment periods between TCD and IES.
• Research was undertaken in relation to state-of-the-art MPC and FDD algorithms, integration requirements, the use of MPC in other industries, building user requirements and appropriate hardware, software and communications protocols for building data acquisition and control.
• An EINSTEIN Architecture was developed ahead of integration of the solution’s elements
• The EINSTEIN Algorithms were developed where initial prediction, optimisation and fault detection algorithms were then tested. For MPC this also focused on testing of an Economic-MPC (E-MPC) implementation where minimisation of cost for variable tariffs was considered.
• Model calibration was investigated, and the existing approach for calibrating IESVE models to create the ACMs was improved through EINSTEIN
• Reduced Order Models were created and an approach was developed to easily generate ROMs from the high-fidelity IESVE ACM models, which allows for scalability of the solution
• User Interfaces were developed to display the results and to communicate with the outputs of the algorithms developed back to a user
• Dissemination activities were carried out throughout EINSTEIN to promote the work/findings

These EINSTEIN developments were tested and refined during the project for a range of demo sites, which included:
- MPC climate control tested for a residential building in Dublin, Ireland
- E-MPC for a residential building located in Findhorn, Scotland
- An open loop system ID demo for a system identification ROM in TCD, Dublin
- E-MPC for a commercial office building in Dublin, Ireland
- MPC and E-MPC for an apartment building in Louth, Ireland
- Signal based fault detection for heating equipment for a typical two zone residential building
- Rule-based fault detection testing on real data from a University room based in Galway Ireland

All of the tests carried out resulted in improved performance of the building’s in terms of energy consumption and/or cost while (critically) maintaining user comfort. Tested on both residential and commercial buildings, the (E-)MPC and FDD algorithms show promise as practical solutions to improve the performance of buildings across the EU building stock. Testing of the MPC algorithms on the models developed resulted in demonstrated energy savings in the range of 15-18%, with energy cost savings associated with E-MPC algorithms ranging from 18-40%. Although difficult to associate a definite energy and cost saving with the FDD algorithms implementation, all of the tests successfully resulted in the automated identification of faults, which in itself leads to the avoidance of energy and cost waste and would likely reduce the possibility of user discomfort due to mechanical faults.

Considering that building’s consume almost 40% of energy in the EU, better management of operational buildings will positively affect society both socially and economically through the reduction in energy production required, reduced requirements on the grid infrastructure to meet this demand and the minimisation of operational energy costs, all while maintaining/improving user comfort. The successful prototyping of the algorithms developed in EINSTEIN opens up new and real opportunities to advance beyond classical building control approaches by deploying solutions that are “Smarter”. The key benefits of the demonstrated EINSTEIN approach are that it:
- Addresses the complexity of modern buildings in terms of the systems and dynamics involved, including external factors such as weather;
- Takes advantage of the current technical evolution and progression in meters/sensors and smart devices to provide more building and user data at lower costs
- Aim’s to connect into existing building control infrastructure to reduce the cost and effort of installation of the solution
- Results in better and more efficient building management on a continuous basis, and greater flexibility for control through Building Management Systems (BMS)

You can find out more information about the EINSTEIN project and its results at or by contacting Catherine Conaghan at