Method, air handling unit, system and computer program

Abstract

The present disclosure relates to a method (100) for prediction of a maintenance need of an air handling unit, AHU, comprising at least one power component. The method obtains energy consumption, EC, profiles for the AHU as a whole as well as the at least one power component and compares the profiles with current EC behavior in order predict the maintenance need. The present disclosure also relates to corresponding air handling units, systems and computer programs.

Description

[0001] METHOD, AIR HANDLING UNIT, SYSTEM AND COMPUTER PROGRAM

[0002] TECHNICAL FIELD

[0003] The present disclosure relates to an air handling unit and associated methods, systems and computer programs.

[0004] BACKGROUND ART

[0005] When any component of an air handling unit, AHU, does not operate in optimal condition due to malfunction or maintenance requirement, the AHU may consume more energy to maintain a desired set point, have a component brake or cause a safety related issue, such as risk of fire. It is therefore highly desirable to detect potential problems as soon as possible, which alleviates or eliminates the above- mentioned problems. There is thus a need in the art for improved identification of a maintenance need of an air handling unit.

[0006] SUMMARY OF THE DISCLOSURE

[0007] The present disclosure relates to a method for prediction of a maintenance need of an air handling unit, AHU, comprising at least one power component, each power component having a corresponding range of operational states. The method comprises obtaining a first energy consumption, EC, baseline profile of the AHU. Said first EC baseline profile is obtained as a total power consumption of the AHU when each of the at least one power component operates at a predetermined operational state. The first EC baseline profile of the AHU comprises a start-up subset relating to start-up of the AHU and / or a steadystate subset relating to operation after start-up of the AHU. The method further comprises, for each of the at least one power component, obtaining a respective set of second EC baseline profiles. The set of second EC baseline profiles is obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component. Each second EC baseline profile comprises a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component. The method also comprises obtaining a current EC profile for the air handling unit. The method additionally comprises determining a first discrepancy based on a comparison between the obtained current EC profile for the AHU and the obtained first EC baseline profile of the AHU. The method further comprises triggering a predictive fault detection procedure when the determined first discrepancy meets a predetermined first criterion. The predictive fault detection procedure comprises, for each of the at least one power component, obtaining a respective set of current second EC profiles. The set of current second EC profiles is obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component. Each second EC baseline profile comprises a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component. The predictive fault detection procedure further comprises determining a second discrepancy based on a comparison between the obtained sets of current second EC profiles and the obtained set of second EC baseline profiles. The predictive fault detection procedure additionally comprises predicting a maintenance need in one of the power component(s) and / or other component of the AHU based on the determined second discrepancy.

[0008] The disclosed method thereby enables prediction of when the air handling unit, AHU, or any of its components are likely to experience failure, require maintenance and / or is not operating in an optimal condition, which affects the performance and energy consumption. Using one energy meter for the whole AHU, the present disclosure proposes an approach to store the energy consumption profile of the unit's components during the startup and steady state operating. These profiles, together with the whole AHU's energy consumption profile provides a base line for detecting deviations from a normal operating state. Using this information, stand alone or in combination with other sensor data, the disclosed method can localize the component that requires maintenance and predict equipment failures and / or maintenance needs or any failure before it happens. In particular, the second set of EC baseline profiles enable detailed diagnosis of individual power components of the AHU with respect to potential failure, maintenance requirement and / or operating under suboptimal condition. The second set of EC baseline profiles further enables diagnosing complex relations between powered components and non-powered components, e.g. filters, of the AHU. In particular, the second set of EC baseline profiles enable predictive maintenance of non-powered components.

[0009] According to some aspects, the method further comprises obtaining current and optionally historical sensor data relating to the operational state of the at least one power component, and determining whether the determined first discrepancy meets the predetermined first criterion. The first discrepancy meets the predetermined first criterion when the cause of the first discrepancy is not identifiable from the obtained sensor data. The method additionally comprises, when the predetermined first criterion is not met, predicting the maintenance need in one of the powered component(s) and / or other component of the AHU based on the determined first discrepancy and the obtained sensor data. This has synergistic effects in that it simultaneously enhances the predictive capabilities of the AHU and enables early termination of the maintenance prediction procedure if possible, thereby potentially avoiding interruption of the operational use of the AHU.

[0010] According to some aspects, predicting the maintenance need is performed using a neural network. According to some aspects, determining the first discrepancy is performed using a neural network. According to some aspects, determining the second discrepancy is performed using a neural network.

[0011] Neural networks enables capturing complex interdependencies between power components and nonpower components, such as filters, and can therefore be used to predict maintenance needs in AHU:s with complex interdependencies between its components. Neural networks further excel at capture relevant features that relate to the determination of the first and / or second discrepancy. In other words, neural networks enable discrepancy detection beyond simple methods such as deviations from baseline exceeding a predetermined threshold. The use of multiple neural networks can also achieve synergistic effects in that a neural network can first identify features that are relevant for discrepancy determination, but hard for a human to specify by a simple set of rules, which will influence downstream neural networks prediction of the maintenance need.

[0012] According to some aspects, the method further comprises obtaining permission from a user to trigger the predictive fault detection procedure, wherein the predetermined first criterion further comprises having obtained said permission.

[0013] This allows the user to control whether the operational use of the AHU should be interrupted to run diagnostics.

[0014] According to some aspects, the method further comprises notifying a user, the notification relating to preventive maintenance measures in the identified powered component(s) and / or other component of the AHU based on the predicted maintenance need.

[0015] The notification thereby removes the need for a user to perform complex analysis of which component(s) need maintenance; the analysis is performed automatically for the user.

[0016] According to some aspects, wherein during obtaining the first EC baseline profile of the AHU, the predetermined operational state of each power component is a maximum of the corresponding range of operational states of said power component.

[0017] By obtaining the EC baseline profiles at the respective maximums of each component, EC baseline values for operational states below the maximum can be interpolated. According to some aspects, wherein obtaining the first EC baseline profile further comprises transmitting first control signals to a server. The first control signals are configured to cause the server to transmit second control signals to the AHU. The second control signals are configured to cause the AHU to obtain the first EC baseline profile, and transmit the first EC baseline profile to the server. According to some aspects, obtaining the respective set of second EC baseline profiles further comprises transmitting third control signals to a server, the third control signals being configured to cause the server to transmit fourth control signals to the AHU. The fourth control signals are configured to cause the AHU to obtain the respective set of second EC baseline profiles, and transmit the respective set of second EC baseline profiles to the server. According to some aspects, obtaining a current EC profile for the AHU further comprises transmitting the current EC profile to a server. According to some aspects, determining the first discrepancy is performed at a server having received a transmission of the current EC profile.

[0018] This transfers the computational burden of predicting the maintenance need of the air handling unit to the server; the AHU only needs to collect relevant energy consumption and sensor data and provide it, directly or indirectly, to the server.

[0019] The present disclosure further relates to an air handling unit, AHU, comprising at least one power component, each power component having a corresponding range of operational states. The AHU is configured to receive energy consumption, EC, baseline profile acquisition instructions, said EC baseline profile acquisition instructions being configured to cause the AHU to

[0020] • obtain a first energy consumption, EC, baseline profile of the AHU, said first EC baseline profile being obtained as a total power consumption of the AHU when each of the at least one power component operates at a predetermined operational state, said first EC baseline profile of the AHU comprising a start-up subset relating to start-up of the AHU and / or a steady-state subset relating to operation after start-up of the AHU,

[0021] • for each of the at least one power component, obtain a respective set of second EC baseline profiles, the set of second EC baseline profiles being obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component, each second EC baseline profile comprising a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component, and

[0022] • transmit the obtained first EC baseline profile and the sets of second EC baseline profiles to a server. The AHU is further configured to obtain a current EC profile for the AHU. The AHU is also configured to transmit the current EC profile for the AHU to the server. The AHU is additionally configured to, for each of the at least one power component, obtain a respective set of current second EC profiles, the set of current second EC profiles being obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component, each second EC baseline profile comprising a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component, and transmit the sets of current second EC profiles to the server.

[0023] The AHU is thereby configured to enable the disclosed method while off-loading the computational burden to an external server.

[0024] According to some aspects, the AHU is configured to obtain the current EC profile for the AHU and transmit the current EC profile for the AHU to the server after each startup of the AHU and / or during operational use with fixed intervals.

[0025] The AHU thereby enables real-time monitoring of the AHU and also provides statistics to the server, which may be used to train neural networks that are used in the prediction of a maintenance need.

[0026] According to some aspects, the AHU is further configured to receive predictive fault detection procedure instructions, the predictive fault detection procedure instructions being configured to cause the AHU to

[0027] • for each of the at least one power component, obtain a respective set of current second EC profiles, the set of current second EC profiles being obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component, each second EC baseline profile comprising a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component, and

[0028] • transmit the sets of current second EC profiles to the server.

[0029] The sets of second EC profiles enable the server to analyze the complex interdependencies of the components of the AHU, both power and non-power components.

[0030] The present disclosure also relates to a system for prediction of a maintenance need of an air handling unit, AHU, the system comprising at least one AHU according to any of the claims 13-15, and a server configured to transmit and receive control instructions and data to / from the at least one AHU, wherein the system is configured to carry out the method according to any of the claims 1-12. The system provides all the technical effects and advantages as the method and AHU according to the present disclosure.

[0031] The present disclosure also relates to a computer program comprising computer program code which, when executed, carries out the disclosed method, as described above and below.

[0032] BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figures la and lb illustrate aspects of the disclosed method;

[0034] Figure 2 illustrates aspects of the disclosed air handling unit; and

[0035] Figure 3 illustrates aspects of the disclosed system.

[0036] DETAILED DESCRIPTION

[0037] Figures la and lb illustrate aspects of the disclosed method 100 method for prediction of a maintenance need of an air handling unit, AHU. Figure la outlines the disclosed method and figure lb illustrates the disclosed method in the context of cloud based predictive maintenance, wherein the AHU communicates with an external server. The AHU comprises at least one power component, each power component having a corresponding range of operational states. Examples of operational states include revolutions per minute, RPM, of a fan and / or a rotating drum, temperature of a radiator, output-voltage of a power component configured to generate an output-voltage based on the power driving the power component.

[0038] During operational use, both the AHU as well as its individual power components will have specific respective time-dependent energy consumptions, herein referred to as energy consumption profiles. The energy consumption will look different during start-up and steady state. During start-up, when a power component (or the AHU as a whole), is powered up, the energy consumption will rise, and sometimes temporarily overshoot, the steady state power level associated with a set point of a corresponding operational state. A main idea of the present disclosure is to initially log energy consumption, EC, profiles for the AHU as a whole, as well as its power components for a set of different operational state set points, wherein the logged EC profiles will serve as baseline profiles characterizing what normal operation of the AHU looks like when there are no problems and no need for maintenance, and then monitor the energy consumption of the AHU during operational use and compare the EC usage with the EC baseline profiles to detect potential problems indicating a maintenance need. This can be further augmented with incorporation of additional sensor data, as described further above and below. Thus, the method 100 comprises obtaining S100 a first energy consumption, EC, baseline profile of the AHU. Said first EC baseline profile is obtained as a total power consumption of the AHU when each of the at least one power component operates at a predetermined operational state. The first EC baseline profile of the AHU comprises a start-up subset relating to start-up of the AHU and / or a steady-state subset relating to operation after start-up of the AHU.

[0039] Another idea of the disclosed method is that, according to some aspects, a server external to the AHU may be used to perform most or all of the analysis relating to the prediction of the maintenance need of the air handling unit. This removes the need for computational circuitry at the AHU by relying on an external server. According to some aspects, the AHU comprises a receive and transmit receive module configured to transmit and receive data and control signals to / from the server, either by communicating directly with the server or indirectly via a node configured to communicate with the server, such as a wireless remote controller.

[0040] Thus, according to some aspects, wherein obtaining S100 the first EC baseline profile further comprises transmitting S102a first control signals to a server. The first control signals are configured to cause the server to transmit second control signals to the AHU. The second control signals are configured to cause the AHU to obtain S100 the first EC baseline profile, and transmit S104a the first EC baseline profile to the server. According to some aspects, the first control signals are transmitted S102a via a user interface, Ul, by a user of the AHU. The Ul may be comprised in the AHU and / or be part of an external device, such as a mobile phone or a computer.

[0041] The method further comprises, for each of the at least one power component, obtaining S110 a respective set of second EC baseline profiles. The set of second EC baseline profiles is obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component. Each second EC baseline profile comprises a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component. According to some aspects, during obtaining the first EC baseline profile of the AHU, the predetermined operational state of each power component is a maximum of the corresponding range of operational states of said power component. EC baseline profiles for an operational state not previously obtained can be approximated by interpolating between two known EC baseline profiles. Obtaining EC baseline profiles of the AHU when each power component is operating at their respective maximum of range of operational states enables a maximum interpolation range for each EC baseline profile of the power components.

[0042] According to some aspects, obtaining the respective set of second EC baseline profiles further comprises transmitting third 102b control signals to a server, the third control signals being configured to cause the server to transmit fourth control signals to the AHU. The fourth control signals are configured to cause the AHU to obtain S110 the respective set of second EC baseline profiles, and transmit S104b the respective set of second EC baseline profiles to the server. According to some aspects, the third control signals are transmitted S102b via the user interface, Ul, by the user of the AHU.

[0043] Further details of aspects of the obtaining S100, S110 of the first EC baseline profile and the sets of second EC baseline profiles are illustrated in Figure lb. According to some aspects, the transmitting of S102a, S102b first and third control signals to a server is triggered SlOOa by a user via the Ul. The first and third control signals thus comprises instructions that initiate an EC profiling procedure for the AHU operating under normal conditions. The second and fourth control signals transmitted from the server to the AHU are configured to start the EC profiling procedure, which may comprise a series of activities as part of the obtaining S100, S110 of the first EC baseline profile and the sets of second EC baseline profiles, as described further below.

[0044] Upon receiving the second and fourth control signals, the AHU initiates S103a the data collection required for obtaining S100, S110 of the first EC baseline profile and the sets of second EC baseline profiles. According to some aspects, the method further comprises obtaining S400 current and optionally historical sensor data relating to the operational state of the at least one power component. According to some further aspects, obtaining S400 sensor data comprises obtaining a set of sensor data profiles, wherein each sensor data profile comprises a time series of sensor data. The time series of sensor data is preferably obtained within a time interval that overlaps with the startup and / or steady state operation of a power component and / or the AHU.

[0045] For each of the at least one power component, a respective set of second EC baseline profiles is then obtained S110 as follows.

[0046] For a first predetermined operational state, a second EC baseline profile is obtained S103b by having a power component of the AHU being powered up and the EC profile for startup and steady state operation is logged and subsequently transmitted to the server. According to some aspects, the first predetermined operational state is a maximum operational state of the power component. According to some aspects, corresponding sensor profiles are also transmitted to the server.

[0047] The power component is subsequently powered off. A second operational state, different from the first predetermined operational state, is chosen from a set of predetermined operational states. Another second EC baseline profile is obtained S103c by having a power component of the AHU being powered up and the EC profile for startup and steady state operation is logged and subsequently transmitted to the server. This obtaining S103c of second EC baseline profiles is repeated for all operational states of the set of predetermined operational states. According to some aspects, corresponding sensor profiles are also transmitted to the server.

[0048] The process of obtaining S103b, 103c second EC baseline profiles is repeated S103d for all power components of the AHU.

[0049] According to some preferred aspects, the second EC baseline profiles of a power component are obtained S103b, 103c, 103d with all other power components being turned off.

[0050] The calibration process further comprises obtaining S103e a set first EC baseline profiles for the AHU. The AHU is power up and allowed to reach steady state for a set of predetermined operational states of the power components of the AHU, and corresponding first EC baseline profiles of the AHU are logged and transmitted to the server. According to some aspects, corresponding sensor profiles are also transmitted to the server.

[0051] According to some aspects, the method further comprises notifying S700 the user, the notification relating to preventive maintenance measures in the identified powered component(s) and / or other component of the AHU based on the predicted maintenance need. Specifically, notifying S700 the may comprise notifying S700a the user that the EC profiling procedure has finished.

[0052] The method also comprises obtaining S200 a current EC profile for the air handling unit. According to some aspects, obtaining S200 a current EC profile for the AHU further comprises transmitting S210 the current EC profile to a server. The current EC profile enables an initial determination if there are any maintenance needs. Real-time monitoring of the AHU can be obtained by continuously monitoring the current energy consumption of the AHU and comparing it with the baseline EC profile for the AHU. Thus, according to some aspects, the AHU is configured to obtain S200 the current EC profile for the AHU and transmit the current EC profile for the AHU to the server after each startup of the AHU and / or during operational use with fixed intervals. In summary, according to some aspects, obtaining S200 the current EC profile comprises running S200a the AHU under normal operating conditions and transmitting S200b the current EC profile for the AHU to the server after each startup and during said normal operating conditions with fixed intervals. Additionally, by continuously collecting EC data, the server can use the data to improve on the method, in particular aspects of the method using neural networks.

[0053] The method additionally comprises determining S300 a first discrepancy based on a comparison between the obtained current EC profile for the AHU and the obtained first EC baseline profile of the AHU. According to some aspects, determining S300 the first discrepancy is performed at the server having received a transmission of the current EC profile. In case the determination S300 of the first discrepancy does not detect any significant deviation of the obtained current EC profile from the first EC baseline profile, the user may be notified S700 of this. Thus, according to some aspects, the method comprises notifying S700b the user that the AHU is running as if it is operating under normal operating conditions.

[0054] According to some aspects, determining S300 the first discrepancy is performed using a neural network.

[0055] The method further comprises triggering S600 a predictive fault detection procedure when the determined first discrepancy meets a predetermined first criterion 1C.

[0056] The determination S300 of the first discrepancy can be enhanced by also taking additional sensor data into consideration. The sensor data can be correlated with the first EC baseline profile in order to determine a maintenance need based on component interdependencies associated with the sensor data.

[0057] For instance, sensor data may comprise differential pressure across a filter, a non-power component. If the filter is getting clogged, a downstream fan, i.e. a power component, may need to increase its energy consumption in order to compensate for the filter getting clogged. Since the fan increases its energy consumption, the total energy consumption also goes up. Thus, by combining sensor data, in this example differential pressure, with the first discrepancy, in this example an increase in current EC of the AHU, a maintenance need can potentially be determined.

[0058] Thus, according to some aspects, the method further comprises obtaining S400 current and optionally historical sensor data relating to the operational state of the at least one power component, and determining S410 whether the determined first discrepancy meets the predetermined first criterion 1C. The first discrepancy meets the predetermined first criterion 1C when the cause of the first discrepancy is not identifiable from the obtained sensor data. The method additionally comprises, when the predetermined first criterion 1C is not met, predicting S420 the maintenance need in one of the powered component(s) and / or other component of the AHU based on the determined S300 first discrepancy and the obtained S400 sensor data. According to some aspects, predicting the maintenance need is performed using a neural network.

[0059] According to some aspects, the method further comprises obtaining S500 permission from a user to trigger the predictive fault detection procedure, wherein the predetermined first criterion further comprises having obtained said permission. According to some aspects, the user is notified S700d and the notification is configured to request said permission. The predictive fault detection procedure comprises, for each of the at least one power component, obtaining S610 a respective set of current second EC profiles. The set of current second EC profiles is obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component. Each second EC baseline profile comprises a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component. The predictive fault detection procedure further comprises determining S620 a second discrepancy based on a comparison between the obtained sets of current second EC profiles and the obtained set of second EC baseline profiles. According to some aspects, determining the second discrepancy is performed using a neural network. The predictive fault detection procedure additionally comprises predicting S630 a maintenance need in one of the power component(s) and / or other component of the AHU based on the determined second discrepancy. According to some aspects, predicting the maintenance need is performed using a neural network. Just like in the above example with sensor data and the total energy consumption of the AHU, the sets of current second EC profiles can be combined with sensor data to improve the prediction of a maintenance need of the AHU. Examples of sensor data includes temperature, vibration, flow and pressure, including differential pressure.

[0060] According to some aspects, the predictive fault detection procedure additionally comprises transmitting S602 fifth control signals to the server, the fifth control signals being configured to cause the server to transmit sixth control signals to the AHU, wherein the sixth control signals are configured to cause the AHU to obtain S610 the respective set of current second EC profiles, and transmitting the obtained S610 respective set of current second EC profiles to the server. The predictive fault detection procedure further comprises determining S620 the second discrepancy based on a comparison between the obtained S610 sets of current second EC profiles and the obtained S110 set of second EC baseline profiles, and predicting S630 the maintenance need at the server.

[0061] By transmitting the sensor data, including the energy consumption to the server, the method enables analyzing the data in real time, predict equipment failures or maintenance needs or any failure before it happens using correlation analysis and / or machine learning and notify the user in advance. Thus, to summarize, using one energy meter for the whole AHU, the disclosed method proposes an approach to file the energy consumption profile of the AHU's components during the startup and steady state operating. These profiles, together with the energy consumption profile of the whole AHU provides a base line for detecting deviations from normal operation of the AHU. Using this information, stand alone or in combination with other sensor data, the method enables localization of the component that requires maintenance and predict equipment failures or maintenance needs or any failure before it happens. According to some aspects, the method further comprises notifying S700, S700c a user, the notification relating to preventive maintenance measures in the identified powered component(s) and / or other component of the AHU based on the predicted maintenance need. According to some aspects, the notification further relates to a completion of the server having received the EC baseline profile of the AHU and the sets of second EC baseline profiles.

[0062] Figure 2 illustrates aspects of the disclosed air handling unit 200. The air handling unit 200, AHU, comprises at least one power component 210, each power component 210 having a corresponding range of operational states. The AHU 200 is configured to receive energy consumption, EC, baseline profile acquisition instructions. Said EC baseline profile acquisition instructions is configured to cause the AHU to obtain a first energy consumption, EC, baseline profile of the AHU, said first EC baseline profile being obtained as a total power consumption of the AHU when each of the at least one power component operates at a predetermined operational state, said first EC baseline profile of the AHU comprising a start-up subset relating to start-up of the AHU and / or a steady-state subset relating to operation after start-up of the AHU. The EC baseline profile acquisition instructions is further configured to cause the AHU to for each of the at least one power component, obtain a respective set of second EC baseline profiles, the set of second EC baseline profiles being obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component, each second EC baseline profile comprising a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component. The EC baseline profile acquisition instructions is also configured to cause the AHU to transmit the obtained first EC baseline profile and the sets of second EC baseline profiles to a server. The AHU 200 is also configured to obtain a current EC profile for the AHU. The AHU 200 is additionally configured to transmit the current EC profile for the AHU to the server. The AHU 200 is further configured to, for each of the at least one power component, obtain a respective set of current second EC profiles, the set of current second EC profiles being obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component, each second EC baseline profile comprising a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component, and transmit the sets of current second EC profiles to the server.

[0063] According to some aspects the AHU is configured to obtain the current EC profile for the AHU and transmit the current EC profile for the AHU to the server after each startup of the AHU and / or during operational use with fixed intervals. According to some aspects the AHU is further configured to receive predictive fault detection procedure instructions, the predictive fault detection procedure instructions being configured to cause the AHU to, for each of the at least one power component, obtain a respective set of current second EC profiles, the set of current second EC profiles being obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component, each second EC baseline profile comprising a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component, and transmit the sets of current second EC profiles to the server. Figure 3 illustrates aspects of the disclosed system 350 for prediction of a maintenance need of an air handling unit, AHU. The system comprises at least one AHU 300 according to the present disclosure, as described above and below, and a server 320 configured to transmit and receive control instructions and data to / from the at least one AHU, wherein the system 350 is configured to carry out the disclosed method, as described above and below.

[0064] The present disclosure also relates to a computer program comprising computer program code which, when executed, carries out the disclosed method, as described above and below.

Claims

CLAIMS1. A method (100) for prediction of a maintenance need of an air handling unit, AHU, comprising at least one power component, each power component having a corresponding range of operational states, the method (100) comprising obtaining (S100) a first energy consumption, EC, baseline profile of the AHU, said first EC baseline profile being obtained as a total power consumption of the AHU when each of the at least one power component operates at a predetermined operational state, said first EC baseline profile of the AHU comprising a start-up subset relating to start-up of the AHU and / or a steady-state subset relating to operation after start-up of the AHU, for each of the at least one power component, obtaining (S110) a respective set of second EC baseline profiles, the set of second EC baseline profiles being obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component, each second EC baseline profile comprising a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component, obtaining (S200) a current EC profile for the air handling unit, determining (S300) a first discrepancy based on a comparison between the obtained (S200) current EC profile for the AHU and the obtained (S100) first EC baseline profile of the AHU, triggering (S600) a predictive fault detection procedure when the determined first discrepancy meets a predetermined first criterion (1C), the predictive fault detection procedure comprising• for each of the at least one power component, obtaining (S610) a respective set of current second EC profiles, the set of current second EC profiles being obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component, each second EC baseline profile comprising a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component,• determining (S620) a second discrepancy based on a comparison between the obtained (610) sets of current second EC profiles and the obtained (S110) set of second EC baseline profiles, and• predicting (S630) a maintenance need in one of the power component(s) and / or other component of the AHU based on the determined second discrepancy.

2. The method (100) according to claim 1, further comprising obtaining (S400) current and optionally historical sensor data relating to the operational state of the at least one power component, and determining (S410) whether the determined first discrepancy meets the predetermined first criterion (1C), wherein the first discrepancy meets the predetermined first criterion (1C) when the cause of the first discrepancy is not identifiable from the obtained sensor data, and wherein when the predetermined first criterion (1C) is not met, predicting (S420) the maintenance need in one of the powered component(s) and / or other component of the AHU based on the determined (S300) first discrepancy and the obtained (S400) sensor data.

3. The method according to claim 2, wherein predicting (S420, S630) the maintenance need is performed using a neural network.

4. The method (100) according to any of the preceding claims, further comprising obtaining (S500) permission from a user to trigger (S600) the predictive fault detection procedure, wherein the predetermined first criterion further comprises having obtained (S500) said permission.

5. The method (100) according to any of the preceding claims, further comprising notifying a user (S700), the notification relating to preventive maintenance measures in the identified powered component(s) and / or other component of the AHU based on the predicted maintenance need.

6. The method according to any of the preceding claims, wherein during obtaining (S100) the first EC baseline profile of the AHU, the predetermined operational state of each power component is a maximum of the corresponding range of operational states of said power component.

7. The method according to any of the preceding claims, wherein determining (S300) the first discrepancy is performed using a neural network.

8. The method according to any of the preceding claims, wherein determining (S620) the second discrepancy is performed using a neural network.

9. The method according to any of the preceding claims, wherein obtaining (S100) the first EC baseline profile further comprises transmitting (S102a) first control signals to a server, the first control signals being configured to cause the server to transmit second control signals to the AHU, wherein the second control signals are configured to cause the AHU to• obtain (S100) the first EC baseline profile, and• transmit (S104a) the first EC baseline profile to the server.

10. The method according to any of the preceding claims, wherein obtaining (S110) the respective set of second EC baseline profiles further comprises transmitting (S102b) third control signals to a server, the third control signals being configured to cause the server to transmit fourth control signals to the AHU, wherein the fourth control signals are configured to cause the AHU to• obtaining (S110) the respective set of second EC baseline profiles, and• transmit (S104b) the respective set of second EC baseline profiles to the server.

11. The method according to any of the preceding claims, wherein obtaining (S200) a current EC profile for the AHU further comprises transmitting (S210) the current EC profile to a server.

12. The method according to any of the preceding claims, wherein determining (S300) the first discrepancy is performed at a server having received a transmission (S210) of the current EC profile.

13. An air handling unit (200), AHU, comprising at least one power component (210), each power component (210) having a corresponding range of operational states, the AHU (200) being configured toreceive energy consumption, EC, baseline profile acquisition instructions, said EC baseline profile acquisition instructions being configured to cause the AHU to• obtain a first energy consumption, EC, baseline profile of the AHU, said first EC baseline profile being obtained as a total power consumption of the AHU when each of the at least one power component operates at a predetermined operational state, said first EC baseline profile of the AHU comprising a start-up subset relating to start-up of the AHU and / or a steady-state subset relating to operation after start-up of the AHU,• for each of the at least one power component, obtain a respective set of second EC baseline profiles, the set of second EC baseline profiles being obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component, each second EC baseline profile comprising a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component, and• transmit the obtained first EC baseline profile and the sets of second EC baseline profiles to a server, obtain a current EC profile for the AHU, transmit the current EC profile for the AHU to the server, for each of the at least one power component, obtain a respective set of current second EC profiles, the set of current second EC profiles being obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component, each second EC baseline profile comprising a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component, and transmit the sets of current second EC profiles to the server.

14. The air handling unit according to claim 13, wherein the AHU is configured to obtain the current EC profile for the AHU and transmit the current EC profile for the AHU to the server after each startup of the AHU and / or during operational use with fixed intervals.

15. The air handling unit according to claim 13 or 14, wherein the AHU is further configured to receive predictive fault detection procedure instructions, the predictive fault detection procedure instructions being configured to cause the AHU to• for each of the at least one power component, obtain a respective set of current second EC profiles, the set of current second EC profiles being obtained as a total power consumption of the power component when operating at a plurality of operational states within the range of operational states of the power component, each second EC baseline profile comprising a start-up subset relating to start-up of the power component and / or a steady-state subset relating to operation after start-up of the power component, and• transmit the sets of current second EC profiles to the server.

16. A system for prediction of a maintenance need of an air handling unit, AHU, the system comprising at least one AHU according to any of the claims 13-15, and a server configured to transmit and receive control instructions and data to / from the at least one AHU, wherein the system is configured to carry out the method according to any of the claims 1-12.

17. A computer program comprising computer program code which, when executed, carries out the method according to any of claims 1-12.

Images