Something in the air
One of the main topics at last week's CIBSE Building Performance Conference and Exhibition, indoor air quality is fast being recognised as a priority concern in the world of wellbeing. But in a field where marginal gains are everything, being on top of your data is very important Arie Taal from the Department of Mechanical Engineering at The Hague University has produced research into eliminating faults in HVAC using a BMS
Carbon Dioxide based demand control ventilation (DCV) can reduce heating/cooling loads by up to 30% and fan power consumption by up to 35%. DCV maintains the CO2 concentration in a room within an appropriate range by adjusting the supply air flowrate. CO2-based DCV is the most commonly used control method with CO2 sensors installed in the main return air duct. Nowadays, the increased requirement for smart buildings, combined with a decrease of CO2 sensor prices, has resulted in buildings being equipped with more sensors.
A common issue occurs when one of the CO2 sensors encounters a fault. This can be down to a lack of maintenance or incorrect sensor placements in rooms. In a DCV system, a fault can mean that the estimated energy savings and air quality is not guaranteed. In 1993 the Automatic Background Calibration (ABC) method was developed to calibrate CO2 sensors with the idea that CO2 levels would drop outside normal levels in buildings that are not occupied on weekends or weekday evenings. However, placement of sensors can become a problem as rooms on the inner side of a building or rooms with well-sealed windows may never drop outside of these baseline levels.
Alongside Dr Yang Zhao and Prof Wim Zeiler at the Department of the Built Environment, Eindhoven University in the Netherlands, Mr Taal has been working toward a systematic method of diagnosing faults in CO2 sensors. Using automatic fault detection, diagnosis and self-correction in CO2 sensors would be a proactive method in air conditioning systems to solve this problem. The premise of Mr Taal’s study has been to show how the automatic commissioning of CO2 sensors in air conditioning systems is achievable using benchmark values obtained in one of two methods.
In conventional methods, sensor faults are detected by comparing their measurements with benchmark values. These values can be obtained manually, measured by technicians, or calculated automatically using other available measurements. The latter is more common because it can be done automatically in the building management system (BMS). Practical issues arise in air conditioning systems because there are no sensors equipped to measure the CO2 generation rate, CO2 concentration in the supply air and the flow rate of the supply air in m3/s. In the development of models for CO2 sensor fault detection, the lack of information poses a real challenge.
In an effort to eliminate the threat posed by this lack of information the idea is to perform one of two test methods under specific operating conditions to ascertain the required benchmark levels.
The first is to recycle air without adding any outdoor air for between one and two hours to create 100% return air ventilation. By closing all windows, doors and fresh air dampers in air handling units the measurements of all CO2 sensors should theoretically be the same. The second test is full outdoor ventilation, to supply fresh air into the building without any recycling for between one and two hours. Again, at the end of the time period all of the CO2 sensors should be the same and equal the CO2 concentration of the ambient air.
Faulty sensors will be detected if their readings are different from the assessed benchmark values. A faulty sensor can be detected if its measurements are obviously higher or lower than other sensors.
In the first method, the degree of fault is then measured from the difference between the defective sensor and the average measurement of the other faultless CO2 sensors. The second compares the faulty sensor reading to the ambient CO2 concentration both looking for a negative or positive bias in CO2 levels when measured against the benchmark.
Self-correction is the final step in the process where all of the information is taken from the faulty sensor for adjustment. Using the assessment results from the fault diagnosis the CO2 bias can be corrected. The results of the detection, diagnosis and self-correction will then be reported to technicians for reference.
Together with his team, Mr Taal produced a simulation of their works on the first floor of a school building at The Hague University in Delft. In their experiment, nine rooms were used with a CO2-based DCV applied to control the amount of supply air to each room in order to keep the CO2 measurements within the benchmark. Separate experiments were conducted to simulate different conditions. The first simulates a fault free operation and a second introduces faulty sensors to show the impacts of automatic fault detection system.
Using two operating methods to obtain CO2 benchmarks, 100% return air ventilation and full outdoor air ventilation; faulty sensors can be detected, diagnosed and self-corrected using a BMS. From the simulations, results show that after 45 minutes there are obvious differences between functional sensors and those that are faulty. After an hour and a half the positive or negative bias can be accurately measured.
Theoretically, the proposed methods are effective ways to detect faulty CO2 sensors, effectively diagnoses the state of failure and to automatically remove the fault. The ability to automatically detect, diagnose and repair faults is vital to the effective running of DCV systems.
Carbon Dioxide based demand control ventilation (DCV) can reduce heating/cooling loads by up to 30% and fan power consumption by up to 35%. DCV maintains the CO2 concentration in a room within an appropriate range by adjusting the supply air flowrate. CO2-based DCV is the most commonly used control method with CO2 sensors installed in the main return air duct. Nowadays, the increased requirement for smart buildings, combined with a decrease of CO2 sensor prices, has resulted in buildings being equipped with more sensors.
A common issue occurs when one of the CO2 sensors encounters a fault. This can be down to a lack of maintenance or incorrect sensor placements in rooms. In a DCV system, a fault can mean that the estimated energy savings and air quality is not guaranteed. In 1993 the Automatic Background Calibration (ABC) method was developed to calibrate CO2 sensors with the idea that CO2 levels would drop outside normal levels in buildings that are not occupied on weekends or weekday evenings. However, placement of sensors can become a problem as rooms on the inner side of a building or rooms with well-sealed windows may never drop outside of these baseline levels.
 |
| Properly controlled ventilation can have a big impact on a buidling's energy consumption |
In conventional methods, sensor faults are detected by comparing their measurements with benchmark values. These values can be obtained manually, measured by technicians, or calculated automatically using other available measurements. The latter is more common because it can be done automatically in the building management system (BMS). Practical issues arise in air conditioning systems because there are no sensors equipped to measure the CO2 generation rate, CO2 concentration in the supply air and the flow rate of the supply air in m3/s. In the development of models for CO2 sensor fault detection, the lack of information poses a real challenge.
 |
| It is important to test the sensors without outdoor air, and with solely outdoor air |
The first is to recycle air without adding any outdoor air for between one and two hours to create 100% return air ventilation. By closing all windows, doors and fresh air dampers in air handling units the measurements of all CO2 sensors should theoretically be the same. The second test is full outdoor ventilation, to supply fresh air into the building without any recycling for between one and two hours. Again, at the end of the time period all of the CO2 sensors should be the same and equal the CO2 concentration of the ambient air.
Faulty sensors will be detected if their readings are different from the assessed benchmark values. A faulty sensor can be detected if its measurements are obviously higher or lower than other sensors.
In the first method, the degree of fault is then measured from the difference between the defective sensor and the average measurement of the other faultless CO2 sensors. The second compares the faulty sensor reading to the ambient CO2 concentration both looking for a negative or positive bias in CO2 levels when measured against the benchmark.
Self-correction is the final step in the process where all of the information is taken from the faulty sensor for adjustment. Using the assessment results from the fault diagnosis the CO2 bias can be corrected. The results of the detection, diagnosis and self-correction will then be reported to technicians for reference.
Together with his team, Mr Taal produced a simulation of their works on the first floor of a school building at The Hague University in Delft. In their experiment, nine rooms were used with a CO2-based DCV applied to control the amount of supply air to each room in order to keep the CO2 measurements within the benchmark. Separate experiments were conducted to simulate different conditions. The first simulates a fault free operation and a second introduces faulty sensors to show the impacts of automatic fault detection system.
 |
| A potentially serious problem with an HVAC system can be detected simply using a BMS |
Using two operating methods to obtain CO2 benchmarks, 100% return air ventilation and full outdoor air ventilation; faulty sensors can be detected, diagnosed and self-corrected using a BMS. From the simulations, results show that after 45 minutes there are obvious differences between functional sensors and those that are faulty. After an hour and a half the positive or negative bias can be accurately measured.
Theoretically, the proposed methods are effective ways to detect faulty CO2 sensors, effectively diagnoses the state of failure and to automatically remove the fault. The ability to automatically detect, diagnose and repair faults is vital to the effective running of DCV systems.
Comments
Post a Comment