WO2022049378A1

WO2022049378A1 - Automated sensor calibration device

Info

Publication number: WO2022049378A1
Application number: PCT/GB2021/052263
Authority: WO
Inventors: Gary Barnett; Johnathan CHURCHWARD-STEEL
Original assignee: Airsensa Products Limited
Priority date: 2020-09-01
Filing date: 2021-09-01
Publication date: 2022-03-10
Also published as: GB2598719A; GB202013707D0

Abstract

This invention relates to a device that is capable of automated calibration of sensors. The device is capable of precise yet efficient calibration of large quantities of sensors. The present invention also contemplates a method for calibrating sensors using a device of the present invention. The calibration device for performing sensor calibration comprises: one or more analyte sources (12); a mixing chamber (14); and a chamber (16) capable of accepting one or more sensors to be calibrated (18), wherein the chamber comprises an exhaust (20). The device further comprises one or more reference instruments (22) in fluid communication with the chamber; and a control unit (24), wherein the control unit is adapted to receive a signal produced by the one or more sensors in response to the analyte concentration and the control unit is adapted to be responsive to said signal.

Description

Automated Sensor Calibration Device

[0001] This invention relates to a device that is capable of automated calibration of sensors. The device is capable of precise yet efficient calibration of large quantities of sensors. The present invention also contemplates a method for calibrating sensors using a device of the present invention.

BACKGROUND

[0002] Sensors are used in numerous applications within modern day life to measure and monitor physical and environmental phenomena in order to better understand the environment and determine the best response to changes in that environment. Sensors vary in their design, construction, and purpose from basic sensors which may be as simple as a switch, to complex sensors that can accurately measure the concentration of toxic gasses in air.

[0003] The ability to accurately measure environmental phenomena like air quality, pollution levels, and the presence of toxic or irritant compounds in the air is growing in importance as our awareness of the health and economic impacts of poor air quality develops. Poor air quality has a massive economic cost and has a strong bearing on the health and welfare of the people who experience poor or marginal air quality.

[0004] This desire to better understand the impact of air quality on public health has led to an increased demand for accurate and timely data on air quality, so its impacts can be better understood and appropriate steps can be taken to mitigate the effects of poor air quality.

[0005] However, in order to properly establish the impact of air quality, and to prompt active measures to improve it (or mitigate its effects) we need to monitor air quality in a much more localised way - In offices, homes, factories, and public spaces. This need has driven demand for lower-cost air quality monitoring solutions and has spurred innovation in the design and manufacture of sensors.

[0006] We also need to monitor the appropriate analytes; in urban air quality monitoring the key analytes are Ozone (03), Nitrogen Dioxide (NO2), particles (at varying sizes), sound, vibration, temperature, and humidity. In offices the focus is on Volatile Organic Compounds (VOC), particles, Carbon Dioxide (CO2), sound, and vibration. In domestic premises the most interesting analytes are Carbon Monoxide (CO), smoke, Volatile Organic Compounds, particles (at varying sizes), sound, vibration, temperature, and humidity.

[0007] The precise measurement of these analytes has traditionally involved costly and complex technology and processes, however in the past few years a range of new “low- cost” technologies have emerged which purport to be capable of measuring different phenomena accurately and inexpensively.

[0008] While the manufacturers of these sensors make confident claims about the precision of their inexpensive sensors, it is well known that their precision varies significantly, and many sensors degrade over time, causing their readings to drift with use. In order to make effective use of these sensors it is vital, therefore, to establish a reliable and cost-effective means of evaluating, benchmarking, and calibrating low cost sensors in order to establish a degree of confidence in the data they produce.

[0009] Sensors are only as good as the process by which they have been calibrated. An imprecisely calibrated sensor will fail to accurately detect the entity to be detected; this is of particular concern with sensors that measure gasses such as ozone, carbon monoxide, carbon dioxide, sulfur oxides, nitrogen oxides, volatile organic compounds, hydrocarbons and more. It is also an issue with the accurate measurement of particulates. Therefore, the calibration of sensors is of paramount importance when producing, selecting, or deploying sensors. In addition, it is crucial that a sensor maintains an accurate reading throughout its lifetime. Therefore, monitoring and recalibration is also an important aspect in the lifecycle of a sensor.

[0010] Calibration of sensors generally proceeds by exposing a sensor to an entity to be measured, referred to herein as an analyte. The analyte may be a pollutant. The sensor then provides a measurement for the analyte and the device can be modified by a calibration parameter so that the measurement made by the sensor can be adjusted to provide the correct measurement. Most calibration devices rely on the addition of precisely known quantities of analytes which adds to the cost and complexity of the solution. Other calibration devices rely on a significant amount of manual effort to perform the calculations that are necessary in order to derive the calibration parameters.

[0011] Accordingly, calibration devices and methods known in the art suffer from the problem of requiring highly specific calibration environments which can be challenging and expensive or, where simpler approaches are adopted have a tendency to produce inaccurate results. Prior art devices and methods also suffer from the problem of requiring the performance of time consuming and challenging manual calculations to derive the calibration parameters. Furthermore, the accuracy of prior art calibration techniques is lacking. Embodiments of the present invention intend to solve at least one of these problems. For example, the present invention aims to greatly reduce the effort required to perform calibrations on large numbers of sensors, while also providing a greater level of precision in calibration.

[0012] The current approaches to sensor calibration are complex, inefficient, and costly, moreover some of the stated best practices suffer from significant shortcomings which render them impractical to implement on a large scale and unlikely to produce meaningful results. Current calibration approaches fall into three categories; lab-based calibration, field-based calibration, and statistical calibration.

[0013] Lab-based calibration typically involves the removal of the sensors from a device so that the individual sensors can be placed into a calibration chamber into which is then filled with a calibration gas made up of a mixture of an inert gas and the gas that the sensor measures. Calibration gasses are bought from external suppliers who provide them in preset concentrations which are certified as accurate. Calibration gasses are expensive to procure, and unless the chamber is designed specifically for a particular sensor or device a great deal of it will be consumed in ensuring the chamber is completely filled. So, while the final results of a typical lab-based calibration can be very precise, it is a costly and labour intensive process. This is why manufacturers of external air quality monitors do not recommend this approach.

[0014] Many manufacturers of external air quality monitors recommend that their monitoring devices be calibrated every few months by collocating them with a public “reference station” (typically a highly accurate monitoring station used by public authorities to monitor air quality to meet statutory obligations) this approach suffers from several shortcomings; it can be difficult to get the site owner/operator’s permission to collocate sensors with their equipment. The first issue is a practical one, even where site owner/operators are willing to allow sensors to be co-located with their equipment, they are very unlikely to agree to the colocation of large numbers of sensors with their equipment. The second key shortcoming of this approach is that environmental conditions such as wind or rain can result in the reference station experiencing significantly different levels of pollution by comparison to the devices that are being compared with it, even when they are very close to one another. [0015] The third class of calibration approach - statistical calibration is presented by some organisations as a solution to the previously laid-out problems. This approach seeks to apply statistical models to the calibration of devices in the field, where the historical behaviour of devices is characterised, in combination with known external factors like wind direction and speed, rainfall, humidity, and temperature to identify correlations between sensors that may be some distance apart. Statistical analysis approaches depend on the factors affecting pollution levels remaining stable, and in an urban environment have shown to be incapable of producing verifiable data.

[0016] The approach applied in this invention takes the best characteristics of Laboratory based calibration and addresses first the cost and complexity of laboratory based approaches by applying automation, and addresses the concerns about lab-based approaches not effectively mimicking real-world conditions by actively recreating real-world conditions within the chamber.

[0017] Embodiments of the present invention seek to solve the problems associated with the prior art calibration techniques. For example, the present invention removes the need for highly controlled and expensive calibration methods requiring specific analyte concentrations.

BRIEF SUMMARY OF THE DISCLOSURE

[0018] In accordance with the present invention there is provided a calibration device for performing sensor calibration (optionally automated sensor calibration), wherein the device comprises:

1) one or more analyte sources;

2) a mixing chamber comprising an ambient air inlet and being in fluid communication with the analyte source;

3) a chamber capable of accepting one or more sensors to be calibrated, wherein the chamber comprises an exhaust and a first inlet, the first inlet being in fluid communication with the mixing chamber;

4) one or more reference instruments in fluid communication with the chamber; and

5) a control unit optionally comprising an interface, wherein the control unit is adapted to receive a signal produced by the one or more sensors in response to the analyte concentration and the control unit is adapted to be responsive to said signal. [0019] In embodiments, the device further comprises a scrubber. The scrubber may be situated at the ambient air inlet and/or situated in between the analyte source and the mixing chamber. Alternatively, the scrubber may be situated at the ambient air inlet and/or situated in between the mixing chamber and the chamber.

[0020] In an embodiment the device is adapted to allow for introducing an analyte into the chamber by exposing the chamber to ambient air. This may be achieved by allowing continuous open fluid communication between the one or more analyte sources the mixing chamber and the chamber capable of accepting one or more sensors. The device does not require a precise concentration of analyte. The present invention is able to provide sensor calibration without a precise concentration of analyte. Preferably, the analyte source and mixing chamber are configured to allow for a continuously changing concentration of analyte within the chamber capable of accepting one or more sensors.

Analyte source

[0021] The device of the present invention can utilise any analyte source for which there is a sensor to measure the analyte. By way of non-limiting examples, the analyte source may be: ambient air, bottled gas, UV ozone generator, electrostatic ozone generator, a combustion unit, fabric pads that have been impregnated with a compound that releases gas or particles for analysis, or a combination thereof. The reference gas may be the same as the gas being measured or may be a “proxy” which is known to cause the sensor to react in the same way as it would in the presence of the analyte the sensor is intended to measure.

[0022] The bottled gas may be selected from: carbon dioxide, carbon monoxide, sulphur oxides, nitrogen oxides, volatile organic compounds, particulates and more etc. The gas may be introduced at varying rates via a mass flow controller which is controlled by the control unit. The mass flow controller is designed to precisely control the flow of gasses or liquids and is used to apply varying doses of a given analyte to the mixing chamber. The mass flow controller is an electrically controlled valve which, when actuated, will allow gas to flow at a rate that is set according to the electrical signal it receives. Accordingly, in embodiments the device further comprises a mass flow controller. The mass flow controller may be located in between the analyte source and the mixing chamber.

[0023] The rate of flow controlled by the mass flow controller is not intended to produce a specific concentration of the target analyte in the chamber, instead it is adjusted to elevate the concentration of the target analyte to within between 4 and 8 predetermined ranges. For NO2 the ranges are as follows: 0-10ppb, 15-20ppb, 25-40ppb, 40-60ppb, 60-80 ppb, 90- HOppb. For Ozone the ranges are as follows 0-10ppb, 15-20ppb, 25-40ppb, 40-60ppb, 60- 80 ppb, 90-1 lOppb. This adjustment process takes the form of a feedback loop which is controlled by the controller. For a given calibration protocol the controller will measure the concentration of the target analyte in the chamber, and then introduce additional quantities of the target analyte, gradually increasing the flow via an electronic signal to the flow controller until the desired concentration range is achieved. This process is described in Figure 4.

[0024] In embodiments the analyte source may be a UV or electrostatic ozone generator. The current device makes use of an electrostatic ozone generator which generates ozone by ionizing air which causes O2 molecules to split, and re-bind with other O2 molecules to form O3 molecules (Ozone). The amount of ozone generated will be controlled by the control unit using a technique known as pulse-width-modulation in which the ozone generator is actuated by the controller for very short periods of time after which there is a variable delay before it is reactivated - the length of the delay between actuations and the duration of the actuations will govern the amount of ozone that is introduced to the chamber.

[0025] In embodiments the analyte source is a combustion unit, for example a gas heater which may take the form of a heated coil which causes a gas or petrochemical fuel to oxidise or a burner similar to burners used as pilot lights in gas boilers. The combustion unit may produce gases and particulates resulting from combustion of a fuel source, such as gas, oil, coal or a combination thereof. The gases and particulates generated by the combustion unit may be supplied to the mixing chamber as a crude mixture or the gases and particulates generated from the combustion unit may be purified so that gases or particulates are excluded from the mixing chamber. The combustion unit may be controlled by the control unit.

Mixing Chamber

[0026] The mixing chamber combines ambient air with analytes introduced from the analyte source. In embodiments the mixing chamber may further comprise mixing apparatus for mixing ambient air with an analyte supplied from the analyte source. The mixing apparatus may be: one or more contoured baffles; one or more fans; or a combination thereof.

Chamber [0027] In embodiments the chamber further comprises an entrance to allow the one or more sensors to enter the chamber. The entrance may be in an open position to allow the one or more sensors to enter the chamber or in a closed position, for example when calibration is taking place. The sensors may also exit the chamber via the entrance. In embodiments the chamber further comprises an exit. The exit may be in an open position to allow the one or more sensors to exit the chamber or in a closed position, when calibration is taking place. When in a closed position the entrance and exit may be air-tight to prevent unwanted mixing of the fluid contents of the chamber with ambient air.

[0028] The chamber (also referred to herein as the calibration chamber) will be capable of being partitioned into two or more zones to allow different zones to be measured independently, this is particularly applicable to the analysis of devices that interact with pollution or which seek to filter pollution out of the air

Reference instruments

[0029] In embodiments the reference instruments are instruments that are used in statutory measurement which have the appropriate certifications and calibration certificates. These reference instruments provide highly accurate and verifiable measurements of the analytes under analysis. The reference instruments are highly accurate measurement instruments that meet the standards set out by MCERTS which is the Environment Agency’s Monitoring Certification Scheme which is used to approve instruments for use in the statutory reporting of environmental phenomena. The MCERTs certification scheme is aligned with EU regulations and operates alongside the US Environmental Protection Agency standards. Exemplary, instruments are manufactured by Ecotech - they are a Serinus 40 (NO2) and a Serinus 10 (Ozone), for particle measurement the system uses a Particles Plus 700 Series.

Control Unit

[0030] The control unit performs a number of functions. The control unit is configured to carry out certain features of the method. As such, the present invention contemplates a deice comprising a control unit (also referred to as a controller) that carry out steps of the method.

[0031] In embodiments the control unit or a processor comprised within the control unit is configured to carry out the calibration phase and/or the data analysis phase. [0032] In embodiments the control unit or a processor comprised within the control unit is configured to compare the one or more sensor signals to the one or more reference signals to produce one or more calibration parameters for each sensor.

[0033] In embodiments the control unit or a processor comprised within the control unit is configured to apply the calibration parameter to the one or more sensors.

[0034] The control unit may control the ingress of ambient air into the mixing chamber and ingress of air into the chamber. Ingress of ambient air into the chamber may be via the mixing chamber or may be via a further inlet comprised within the chamber. Accordingly, in embodiments the chamber comprises a second inlet for introducing ambient air into the chamber.

[0035] In embodiments the control unit controls the addition of the analytes into the mixing chamber. The control unit may introduce an unknown amount of analyte. However, the amount of analyte will be within the sensing range of the one or more sensors and the one or more reference instruments. In an embodiment the control unit introduces a volume of analyte or introduces flow of analyte at a flow rate for a specified period of time.

[0036] In embodiments the one or more reference instruments are controlled by the control unit.

[0037] In embodiments the control unit controls the sensors under operation. The sensors will be controlled via interface hardware which acts as a host for the sensor and provides an electronic interface to the control unit, or they will be controlled via interaction with the sensor device which contains the sensors - this interaction will either be via wireless network or via a wired connection. Where interface hardware is used, the hardware may be procured from the sensor manufacturer or be manufactured by AirSensa according to the application notes provided by the manufacturer. The interface hardware will typically connect to a logging device which has been programmed to interpret the signals sent by the interface hardware and then transmit the resulting data values to the control unit, for example AirSensa STORRM platform, via either WiFi, ethernet, or some other wired or wireless network technology. Where the sensor device is deployed within the chamber, it will transmit data to the control unit (for example the STORRM platform) via its standard communications channel, which may be WiFi, Cellular modem, serial connection, or some other wired or wireless network technology. [0038] The actual operation of the sensors depends on the design of the sensor itself. Some sensors use diffracted laser light to measure the presence of the relevant analyte while others detect the presence of the analyte based on an electrochemical reaction between the analyte and an electrochemical material.

[0039] Algorithms that perform comparisons between the data produced by the instruments currently being analysed and reference instruments that have been externally calibrated. The algorithms may also incorporate data from other instruments which exhibit similar characteristics (in terms of sensor location, sensor manufacturer and type, historical sensor behaviour, sensor manufacturing batch).

[0040] The calibration phase is designed to simulate typical levels of pollution in an urban environment. The calibration phase will comprise any number of analyte phases. However, generally the calibration phase comprises no fewer than four and no more than eight analyte phases. The first phase would typically be “ambient” - using just ambient air This may be called a background phase, discussed below in more detail, or an analyte phase. The second phase will add a small additional volume of the pollutants to the ambient air before it enters the chamber in order to bring the level of the pollutant into the range demanded by that phase (as described above). Each subsequent phase will add incrementally more pollutants to the point that the final phase will simulate an environment that exceeds the levels set by current regulations.

[0041] In embodiments the reference instruments are integrated with the control unit by means of either a wireless or wired connection so their operation and the data they record can be managed and captured. The reference instruments may be co-located with the chamber and sample air from the chamber by drawing a small sample from the chamber and passing it into their analysis instrument.

[0042] In a further aspect of the present invention there is provided a method of calibrating sensors, wherein the method comprises:

1) providing an automated calibration device comprising a chamber and one or more reference instruments adapted to produce a signal in response to an analyte concentration;

2) introducing one or more sensors into the chamber, wherein the sensors are adapted to detect the analyte concentration or to produce a signal dependent on the analyte concentration; 3) carrying out a calibration phase comprising: a. introducing an analyte into the chamber to produce the analyte concentration within the chamber; b. causing the one or more sensors to measure the analyte concentration to produce one or more sensor signals; c. causing the one or more reference instruments to measure the analyte to produce one or more reference signals;

4) carrying out a data analysis phase comprising: a. transmitting the one or more sensor signals and the one or more reference signals to a control unit, the control unit comprising a processor; b. the processor being adapted to receive the one or more sensor signals and the one or more reference signals, the processor comparing the one or more sensor signals to the one or more reference signals to produce one or more calibration parameters for each sensor; and c. applying the calibration parameter to the one or more sensors.

[0043] In embodiments the method may comprise one or more calibration phases. In embodiments the method comprises one or more data analysis phases, for example a number of data analysis phases equal to the number of sensors.

[0044] The step of introducing an analyte into the chamber may be exposing the chamber to ambient air. The step of introducing the analyte does not require a precise concentration of analyte. The present invention is able to provide sensor calibration without a precise concentration of analyte. Preferably, the step of introducing an analyte into the chamber allows for a continuously changing concentration of analyte within the chamber.

[0045] In embodiments, the calibration phase may further comprise one or more of the following:

• obtaining an analyte range from the controller;

• introducing analyte into the mixing chamber prior to introduction into the chamber;

• reading the analyte levels from the reference instruments;

• checking that the analyte levels are within the target range; and • optionally adjusting the analyte level to within the target range.

[0046] In embodiments, the steps of: a. introducing an analyte into the chamber to produce the analyte concentration within the chamber; b. causing the one or more sensors to measure the analyte concentration to produce one or more sensor signals; c. causing the one or more reference instruments to measure the analyte to produce one or more reference signals; are steps within an analyte phase. The analyte phase is a substep of the calibration phase.

[0047] In embodiments the calibration phase comprises a background phase and an analyte phase, wherein the analyte phase comprises the steps of a. introducing an analyte into the chamber to produce the analyte concentration within the chamber; b. causing the one or more sensors to measure the analyte concentration to produce one or more sensor signals; c. causing the one or more reference instruments to measure the analyte to produce one or more reference signals;

[0048] In embodiments the calibration phase comprises an analyte phase, wherein the analyte phase comprises the steps of a. obtaining an analyte range from the controller; b. introducing analyte into the mixing chamber prior to introduction into the chamber; c. introducing an analyte into the chamber to produce the analyte concentration within the chamber; d. reading the analyte levels from the reference instruments; e. checking that the analyte levels are within the target range; f. optionally adjusting the analyte level to within the target range; g. causing the one or more sensors to measure the analyte concentration to produce one or more sensor signals; and h. causing the one or more reference instruments to measure the analyte to produce one or more reference signals.

[0049] In embodiments, the method further comprises the step of querying whether further analyte phase is required. As the skilled person would appreciate, a further analyte phase would result in repeating of any steps associated with the analyte phase or all of the steps associated with the analyte phase.

[0050] In embodiments the calibration phase further comprises a background phase. In embodiments, the background phase comprises introducing ambient air into the chamber and causing the one or more sensors and the one or more reference instruments to carry out sensor background measurements (taken by the one or more sensors) and reference background measurements (taken by the one or more reference instruments) on the ambient air. The background phase is not essential to the calibration phase and is therefore optional. In addition, measurements on ambient air may constitute an analyte phase as ambient air contains certain analytes at a concentration that is suitable for calibration of certain sensors.

[0051] Optionally, the one or more sensors may be one to 50 sensors. For example, there may be ten to 40 sensors, 15 to 40 sensors, 20 to 40 sensors, or 25 to 35 sensors.

[0052] In embodiments the one or more sensor signals will be different for each of the sensors. As the skilled person will appreciate with there being one or more sensors, there may be n sensors, where n is an integer of 1 or more, (optionally 1 to 50) for example n may be 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 etc. Where there are n sensors there may be n sensor signals. Accordingly, there may be n calibration parameters.

[0053] Throughout the disclosure of the present invention, it is contemplated that a sensor may be referred to as a first sensor, second sensor, third sensor fourth sensor, . . .nth sensor. Accordingly, there will be a first sensor signal, second sensor signal, third sensor signal, . . .nth sensor signal. Equally, there may be a first calibration parameter, a second calibration parameter, a third calibration parameter, a fourth calibration parameter, . . .nth calibration parameter.

[0054] The one or more sensors may be individual sensors or they may be sensor units comprising one or more sensors. The units or sensors are placed within the calibration chamber, connected to power and linked to the control unit. The invention also contemplates the inclusion of one or more sensor units within the chamber. As such, reference to one or more sensors may refer to: one or more individual sensors; one or more sensor units; or a combination of one or more individual sensors and one or more individual sensor units.

[0055] In embodiments the method further comprises informing a database that a sensor has entered the chamber, optionally including informing the database that the sensor has entered a calibration phase. In embodiments the method further comprises the control unit allocating a unique identification number to each sensor that enters the chamber.

[0056] Accordingly, in embodiments the method comprises:

2) introducing one or more sensors into the chamber, wherein the sensors are adapted to detect the analyte concentration or to produce a signal dependent on the analyte concentration;

3) allocating each of the one or more sensors with a unique identifier;

4) informing a database of the unique identifier of the sensor that has entered the chamber;

5) carrying out a calibration phase comprising: a. introducing an analyte into the chamber to produce the analyte concentration within the chamber; b. causing the one or more sensors to measure the analyte concentration to produce one or more sensor signals; c. causing the one or more reference instruments to measure the analyte to produce one or more reference signals;

6) carrying out a data analysis phase comprising: a. transmitting the one or more sensor signals and the one or more reference signals to a control unit, the control unit comprising a processor; b. the processor being adapted to receive the one or more sensor signals and the one or more reference signals, the processor comparing the one or more sensor signals to the one or more reference signals to produce one or more calibration parameters for each sensor; and c. applying the calibration parameter to the one or more sensors.

[0057] In embodiments the step of comparing the one or more sensor signals to the one or more reference signals (the data analysis phase, discussed in more detail below later) further comprises using a standard regression analysis which compares all of the collected signals from each sensor being analysed and the signals received by the reference sensor. Specifically, this comparison takes the form of, firstly, a statistical regression analysis which produces a candidate slope (a multiplier) and an offset (a factor that is added to the sensor signal) which produces the closest mathematical fit to the data collected by the reference sensors. Further analysis might, optionally, include, performing the regression analysis on different averaging periods - where many sensor readings are aggregated and averaged to be compared with the same aggregation of the data collected by the reference instrument. Where this additional analysis is performed, the controller will select the slope and offset which produces results that fit most closely with the reference instrument data.

[0058] In embodiments the method comprises:

3) allocating each of the one or more sensors with a unique identifier;

BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Embodiments of the invention are further described here by means of example but not in any limiting sense with reference to the accompanying drawings, in which:

Figure 1 shows an apparatus of the present invention for calibration of sensors.

Figure 2 shows a mixing chamber of the present invention.

Figure 3 shows a flow diagram representing the steps of the method of the present invention.

Figure 4 shows a flow diagram of steps taken during the “perform calibration phases” step of Figure 3.

Figure 5 shows a flow diagram of steps taken dring the “perform analysis of data” step of Figure 3.

Figure 6 shows uncalibrated NO2 data from devices to be calibrated and reference instruments.

Figure 7 shows uncalibrated O3 data from devices to be calibrated and reference instruments.

Figure 8 shows NO2 data after calibration has been applied.

Figure 9 shows O3 data after calibration has been applied.

Figure 10 shows NO2 data for a sensor showing the raw data, reference data and the regressed data following calibration.

Figure 11 shows O3 data for a sensor showing the raw data, reference data and the regressed data following calibration.

Figure 12 shows PM 10 data for a sensor showing the raw data, reference data and the regressed data following calibration. Figure 13 shows PM2.5 data for a sensor showing the raw data, reference data and the regressed data following calibration.

Figure 14 shows NO2 data for a sensor showing the raw data, reference data and the regressed data following calibration.

Figure 15 shows O3 data for a sensor showing the raw data, reference data and the regressed data following calibration.

Figure 16 shows PM 10 data for a sensor showing the raw data, reference data and the regressed data following calibration.

Figure 17 shows PM2.5 data for a sensor showing the raw data, reference data and the regressed data following calibration.

Figure 18 shows NO2 data for a sensor showing the raw data, reference data and the regressed data following calibration.

Figure 19 shows O3 data for a sensor showing the raw data, reference data and the regressed data following calibration.

DETAILED DESCRIPTION

[0060] Referring to Figure 1, the automated sensor calibration device (10) of the present invention comprises one or more analyte sources (12). Figure 1 shows an embodiment of the invention where there is a single analyte source (12). Each analyte source (12) is in fluid communication with a mixing chamber (14). The mixing chamber (14) and analyte source (12) are in fluid communication via a valve. The valve may be placed in an open or closed state dependent on whether or not the mixing chamber and analyte source are required to be in fluid communication (open) or not (closed).

[0061] The mixing chamber (14) comprises an ambient air inlet (26). The ambient air inlet allows for air to be mixed with the analyte from the analyte source so that a desired amount of analyte can be achieved in the mixing chamber. The ambient air inlet also serves to allow air intake into the chamber (16) to flush the chamber or to carry out background measurements.

[0062] Chamber (16) can accept one or more sensors to be calibrated (18). The sensors will be placed within the chamber (16) via an appropriate opening within the chamber (16). Chamber (16) is in fluid communication with the mixing chamber (14), via a first inlet, and ultimately in fluid communication with the analyte source (12). The chamber (16), mixing chamber (14) and analyte source (12) may be separated by a valve, tap, or some other form of closure to allow for controlled fluid communication between each of the chamber (16), mixing chamber (14) and analyte source (12).

[0063] The chamber may also comprise an exhaust (20). The exhaust (20) may be used to evacuate the chamber after each completed calibration or in between each completed stage of the calibration.

[0064] The chamber (16) is also in fluid communication with one or more reference instruments (22). The one or more reference instruments (22) may be placed within the device at a location where the reference instruments can sample the fluid within the chamber (16). Thus, the reference instruments (22) may be connected to the chamber (16) by an opening, such as a pipe or tube.

[0065] The control unit (24) may comprise an interface for an operator to interact with the device. The control unit operates the calibration device.

[0066] The calibration device may also comprise a mass flow controller (27) situated between the analyte source and the mixing chamber (14).

[0067] Referring to Figure 2, ambient air inlet (26) is shown as being an opening in the mixing unit. The ambient air inlet (26) may also be or comprise a valve. The ambient air inlet may also comprise a valve and tubing.

[0068] The mixing chamber also comprises an analyte inlet (28). The analyte inlet is in fluid communication with the analyte source (12), optionally via the mass flow controller (27).

[0069] The mixing chamber comprises means to mix the analyte with a carrier gas. The mixing may be achieved by baffles (30) and/or fans (32). The exhaust of the mixing chamber (34) serves as the conduit to the chamber (16). The exhaust of the mixing chamber (34) may comprise a valve which is optionally controlled by the controller.

[0070] Operation of the Chamber

[0071] The device of the present invention may be operated in the following way.

[0072] One or more sensors (either induvial sensors, sensor units comprising a multiplicity of sensors or a combination) are loaded into the chamber of the device of the present invention, before, after or as they are loaded into the chamber the control unit is informed that the sensor has been loaded into the chamber (16). The control unit can be informed that sensors have been loaded by an operator, for example via means of either a touch sensitive screen into which the operator enters the serial number of the device, or by scanning a barcode on the device. Alternatively, the device may further comprise a barcode reader which automatically reads the barcode on a sensor and inputs the serial number or other identifier into the control unit. In so doing, the control unit then informs a database that the sensor is entering a calibration phase.

[0073] Once the sensors (18) have been loaded into the chamber (16) the device initiates the calibration phase (402). This may be accomplished automatically or by a user informing the control unit to enter the calibration phase, for example by clicking a button on the interface of the control unit.

[0074] The control unit commences a calibration phase shown as a flow diagram in Figure 4. There are two sub-phases of the calibration phase: a background phase and an analyte phase.

[0075] The background phase comprises introducing ambient air into the chamber (404) and causing the one or more sensors and the one or more reference instruments to carry out sensor background measurements (taken by the one or more sensors) and reference background measurements (taken by the one or more reference instruments) on the ambient air (406).

[0076] The ambient air may be introduced into the chamber via an inlet of the chamber or via the ambient air inlet of the mixing chamber (26). In the background phase the ambient air may be drawn through the chamber by way of a continuous flow entering and exiting the chamber. Ambient air may exit the chamber via the exhaust (20). Alternatively, ambient air is introduced into the chamber as a batch and the background measurements are taken on the batch of ambient air. After the background measurements have been taken the batch of ambient air exits the chamber (16) via the exhaust (20).

[0077] The background measurement from the one or more sensors under calibration (18) and the one or more reference instruments (22) are simultaneously recorded. The background measurements are transmitted to a database by the control unit (24). The device (10) takes the background measurements on ambient air for a set period (which may be adjusted over time). Measurements are taken at regular intervals, typically at a rate of one to 10 measurements every second to every 10 seconds for gas sensors and up to 1000 times per second for sound and motion sensors. Each individual reading may then be averaged over 60 second timespans. The data collection phase may have a duration of 1 minute to 1 hour (optionally 10 minutes to 1 hour, 15 minutes to one hour, 20 minutes to one hour, 30 minutes to one hour, 45 minutes to one hour). Optionally, the data collection phase lasts for an hour. Depending on the length of the data collection phase, there will be a number of one minute readings equal to the number of minutes for which the measurements were taken. In the event that the data collection phase is 1 hour long it will result in 60 one minute readings for each analyte from the reference sensors and each of the sensors/devices that have been placed in the unit.

[0078] The one or more sensor measurements and the one or more reference measurements are taken over a predetermined period of time in order generate a sufficient number of datapoints to perform the calibration. The period of time over which a cycle runs and the number of datapoints required will vary according to the nature of the sensors under analysis for gas sensors each phase will have a duration of between 1 and 2 hours to take into account the response times of gas sensors, while acoustic sensors and vibration sensors will require shorter intervals because of their highly responsive nature.

[0079] It is possible that the background phase will provide sufficient data to calibrate the sensors. As such, it is possible that the calibration can be completed following a single background phase.

[0080] However, the unit may then continue the calibration phase by transitioning into the analyte phase. The analyte phase may consist of one or more analyte phases. The analyte phases comprise introducing an analyte (408) into the incoming ambient air in the mixing chamber (14) from the analyte source (12), causing the sensors to measure the analyte concentration to produce one or more sensor signals; and causing one or more reference instruments to measure the analyte concentration (310). The reference instruments relay the measurement of the analyte to the controller (24) to indicate whether or not the analyte concentration is within an acceptable range (312).

[0081] Generally speaking, the amount of analyte that is introduced into the mixing chamber is intended to raise the levels of the analytes in the chamber by between 5 and 10 %. Each calibration phase will entail the introduction of steadily increasing levels of pollutants as set out above. The purpose of the phases is to provide a meaningful range of pollutant levels to provide a sufficiently wide spectrum of readings to simulate the typical levels of pollutants found in urban environments.

[0082] In the event that the analyte levels are within the acceptable range the controller initiates a data collection phase (314) with the one or more sensors and the reference instruments as described above for the background phase, to produce one or more reference signals.

[0083] In the event that the analyte levels are not within the acceptable the flow of analyte into the mixing chamber (14) from the analyte source (12) is modified by either reducing the flow of analyte (316) or increasing the analyte flow (318).

[0084] If the data collection phase has elapsed (320) then the controller will complete that data collection phase within the analyte phase and determine whether or not any further phases are to be run (322). In the event that there are further phases to run then a further analyte phase will start including:

• obtaining an analyte range from the controller;

• introducing analyte into the mixing chamber and subsequently the chamber;

• reading the analyte levels from the reference instruments;

• checking that the analyte levels are within the target range;

• optionally adjusting the analyte level to within the target range;

• initiating a data collection phase;

• completing data collection phase; and

• querying whether further analyte phase is required.

[0085] When sufficient analyte phases have been completed the controller (24) will complete the calibration phase (324).

[0086] For the avoidance of doubt, the multiple calibration phases may be the following: first background phase (ambient air is measured); first analyte phase (a first analyte level is measured); second analyte phase (a second analyte level is measured); third analyte phase (a third analyte level is measured); fourth analyte phase (a fourth analyte level is measured); etc. Please note that to be concise “measured” has been used to indicate the one or more sensors take a measurement and the one or more reference instruments take a measurement, the measurement being a of the ambient air or the analyte as appropriate. For the avoidance of doubt, the calibration phase may consist solely of a first background phase (ambient air is measured).

[0087] Once the requisite number of analyte phases are completed and the calibration phase has ended, the analysis of data (data analysis phase) is carried out as indicated in Figure 5. The data collected from the one or more sensors and the one or more reference instruments is conveyed to the controller (24) and analysed. Calculations are performed to determine the accuracy of the units being calibrated, and the optimal calibration parameters to be applied. This analysis will involve the application of multiple calibration algorithms, ranging from simple linear regression, to multivariate regression algorithms. This phase may also include the transmission of baseline data, and the recorded calibration data to external Al and Machine learning platforms for analysis. Once a set of candidate calibration equations has been established they will be applied to the historic data and the equation that produces the best fit will be selected by the algorithm.

[0088] Referring now to Figure 5, the first step of the data analysis phase is to obtain the recorded data (502). In this step the data obtained from the one or more sensors and the one or more reference instruments is conveyed to the controller (24). The controller then performs a comparison of the data (504). The comparison of data will be between the one or more sensors and one of the reference instruments or more than one of the reference instruments. For example, the one or more sensors may be a sensor device comprising more than one different sensor. As such, the sensor device will need to be compared to each of the reference instruments that is appropriate for each of the sensors comprised within the sensor device.

[0089] Alternatively, the one or more sensors within the chamber may be the same type of sensor. In this case, each of the sensors may be compared to a single reference instrument, appropriate for that sensor. In a further alternative where each of the sensors are the same, each sensor may be compared to one or more reference instruments. This might be appropriate for highly sensitive measurements.

[0090] After the data has been compared (504) the controller calculates the calibration parameter (506). This calculation process will involve a number of factors, including (but not limited to) the historical calibration settings that have been applied to the units under calibration, the batch numbers of the sensors contained within the units, as well as other environmental factors like humidity, pressure, and temperature. An example using linear regression to produce a calibration parameter is provided in Example 1.

[0091] The algorithms may incorporate machine learning processes to determine optimal strategies for the calculation of calibration parameters.

[0092] The new calibration parameters are applied to the one or more sensors that are being calibrated (508). [0093] The data analysis phase is then repeated (510) for each sensor within the calibration chamber. When no further sensors are to be calibrated the data analysis phase is completed.

[0094] On completion of the calibration process (including the calibration phase and the data analysis phase), the units can be removed from the chamber. The one or more sensors may then be allocated to a stock location by the controller, where they will be ready for deployment in the field.

EXAMPLES

[0095] Example 1

[0096] Below is described a method to calibrate multiple low-cost air quality monitoring instruments using an automated calibration chamber which incorporates highly accurate reference instruments as well as a controller to manage the process.

[0097] The particular instruments that were calibrated in this calibration exercise were Aeroqual AQY external Air Quality Monitoring Units. These units are manufactured by Aeroqual, in New Zealand.

[0098] The reference instruments against which the AQY instruments were calibrated were manufactured by Ecotech. An Ecotech Serinus 40 was used to provide reference measurements for nitrogen dioxide (NO2) and an Ecotech Serinus 30 was used to provide reference measurements for ozone (O3). The Ecotech instruments are certified for use in regulatory air quality monitoring applications and are regularly calibrated according to the manufacturer’s instructions as well as being independently calibrated at least once a year.

[0099] The reference instruments were connected to the controller unit via a serial connection, and they were connected to the calibration chamber via PTFE tubing which is connected at one end to the sample inlet port on the reference instrument and led through an aperture in the bottom of the calibration chamber so that the air within the calibration chamber could be sampled and analysed by the reference unit. In operation, the controller polled each reference unit every few seconds to obtain the current readings from its sensors, these readings were averaged over 1 minute intervals and were transmitted to the controller.

[00100] The AQY instruments were configured to transmit data to the controller. On configuration the AQY units are assigned an identity by the controller, in the case of this exercise the AQY units were allocated the following identities: CHAMBER001, CHAMBER003, CHAMBER004, CHAMBER005 and CHAMBER011. The specific identity allocated by the controller is arbitrary and has no impact on the analysis, it is merely used to identify the incoming data as belonging to one device or another.

[00101] The AQY instruments were placed inside the chamber and were connected to a power supply. On being connected to power the instruments immediately began sampling the air and recording the data gathered from the sensors they contain.

[00102] The controller sent a signal to the mixing chamber to start its fans so that ambient air is drawn from outside and flows in via the ambient air inlet (8).

[00103] The controller sets a checkpoint and begins recording the data produced by the reference instruments, storing the data on the controller, alongside the data being collected from the units that are to be calibrated.

[00104] In this example a single phase using ambient air was run, over a period of 24 hours. After 24 hours had elapsed the data analysis stage was initiated.

[00105] The second stage of the process is the data analysis stage, on completion of the data collection phase the data from the units under calibration and the reference units is analysed.

[00106] Figures 6 and 7 show the raw data produced by the AQY instruments and the reference instruments for NO2 and O3. While a correlation is visible from the graph, it can be seen that the instruments are responding differently to the presence of ozone and NO2 within the chamber. The data from each machine is compared to the data from the reference instruments using a regression analysis. The regression method chosen is the method of ordinary least squares, which computes the unique line that minimizes the sum of squared distances between the true data and that line. This provides an estimate of the conditional expectation (or population average value) of the dependent variable (from the unit being analysed) when the independent variables (from the reference instruments) take on a given set of values.

[00107] The result of the linear regression is a function is defined by this equation: y = bx + a where: x is a reading from the unit; y is a reading from the reference unit; a is the Y-intercept, which is the expected mean value of y when all x variables are equal to 0; and b is the slope of a regression line, which is the rate of change for y as x changes.

[00108] As Air Quality Standards are applied to averaged data, typically one hour averages or eight hour averages, the data from the reference units and the units to be calibrated is averaged into one hour averages before the regression analysis is applied. This provides the closest fit to the desired data output and also serves to filter out the impact of small differences in response times between the sensors in the units being calibrated and the reference instruments.

[00109] The regression analysis is conducted iteratively against the data from each unit, in each iteration the data from the reference instruments is requested from the control unit, along with the data from the instrument being calibrated. The regression function is then applied to the data, to derive the slope and offset for each unit.

[00110] The resulting slope and offset are applied as transformation parameters to the data produced by the units. Application of the transformation parameters results in the data shown in Figures 8 and 9 for NO2 and O3 respectively.

[00111] A further check is made to verify the degree of correlation between the units being calibrated and the data produced by the reference instruments. To do this the R squared (R²) value for the data from each individual unit is compared with the reference data using the Pearson product-moment correlation coefficient. Where an R² value is close to 1 a very strong correlation is indicated, as the R² value drops a lower correlation is implied. For the purposes of this exercise an R² value of less than .9 is regarded as an indication of a failure in the sensors, and any units which produce data that exhibits an R² below this level are excluded from calibration and set aside for further testing and remediation, which may include the replacement of the sensors within the unit.

[00112] Units where the R² is at 0.9 or above are regarded as successfully calibrated, and the calibration parameters, in the form of calculated slope and offset, are transmitted to the unit by the controller and subsequently applied to all new sensor readings produced by the unit. Once the new parameters have been applied to it, each unit is then unplugged from the power supply and placed into stock ready to be redeployed in the field.

[00113] On completion of the analysis the following R² values were observed for each of the units under calibration: [00114] Table 1 - R² Values for Devices Post Calibration

[00115] All of the units showed excellent correlation with the reference instruments for O3, an R² > 0.99 represents an extremely strong correlation. For NO2 however, one of the devices, CHAMBER003, exhibited a poor R², and as such was rejected and quarantined.

[00116] Example 2

[00117] Following the same methodology as Example 1 further sensors have been calibrated by the device of the present invention. The multiplier and offset for sensors measuring NO2 and O3 are shown in Table 2 below. After the multiplier and offset are applied to the sensor readings an R² value was calculated showing the correlation between the reference readings and the sensor readings.

Table 2

[00118] Calibration graphs showing raw data, regressed data (where the calibration has been applied) and reference data are shown in Figure 10 for NO2 and Figure 11 for O3.

[00119] Example 3

[00120] Following the same methodology as Example 1 further sensors were calibrated by the device of the present invention. The multiplier and offset for sensors measuring PM10, PM2.5, NO2 and O3 are shown in Table 3 below. PM 10 and PM2.5 refer to particulate matter pollutants. PM2.5 refers to particles with a size of 2.5 microns and PM10 refers to particles with a size of 10 microns. After the multiplier and offset are applied to the sensor readings an R² value was calculated showing the correlation between the reference readings and the sensor readings. Table 3

[00121] Calibration graphs showing raw data, regressed data (where the calibration has been applied) and reference data are shown in Figure 12 for PM10, Figure 13 for PM2.5, Figure 14 for NO2 and Figure 15 for O3.

[00122] Example 4

[00123] Following the same methodology as Example 1 further sensors were calibrated by the device of the present invention. The multiplier and offset for sensors measuring PM 10, PM2.5, NO2 and O3 are shown in Table 4 below. After the multiplier and offset are applied to the sensor readings an R² value was calculated showing the correlation between the reference readings and the sensor readings.

Table 4

[00124] Calibration graphs showing raw data, regressed data (where the calibration has been applied) and reference data are shown in Figure 16 for PM10, Figure 17 for PM2.5, Figure 18 for NO2 and Figure 19 for O3. [00125] Throughout the description and claims of this specification, the words “comprise” and “contain” and variations of them mean “including but not limited to”, and they are not intended to (and do not) exclude other moieties, additives, components, integers or steps. Throughout the description and claims of this specification, the singular encompasses the plural unless the context otherwise requires. In particular, where the indefinite article is used, the specification is to be understood as contemplating plurality as well as singularity, unless the context requires otherwise.

[00126] Features, integers, characteristics, compounds, chemical moieties or groups described in conjunction with a particular aspect, embodiment or example of the invention are to be understood to be applicable to any other aspect, embodiment or example described herein unless incompatible therewith. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and/or all of the steps of any method or process so disclosed, may be combined in any combination, except combinations where at least some of such features and/or steps are mutually exclusive. The invention is not restricted to the details of any foregoing embodiments. The invention extends to any novel one, or any novel combination, of the features disclosed in this specification (including any accompanying claims, abstract and drawings), or to any novel one, or any novel combination, of the steps of any method or process so disclosed.

[00127] The reader's attention is directed to all papers and documents which are filed concurrently with or previous to this specification in connection with this application and which are open to public inspection with this specification, and the contents of all such papers and documents are incorporated herein by reference.

Claims

1. A calibration device for performing sensor calibration, wherein the device comprises:

1) one or more analyte sources;

5) a control unit optionally comprising an interface, wherein the control unit is adapted to receive a signal produced by the one or more sensors in response to the analyte concentration and the control unit is adapted to be responsive to said signal.

2. The calibration device of claim 1, wherein the calibration device further comprises a scrubber.

3. The calibration device of claim 1 or claim 2, wherein the analyte source may be: ambient air, bottled gas, UV ozone generator, electrostatic ozone generator, a combustion unit, fabric pads that have been impregnated with a compound that releases gas or particles for analysis, or a combination thereof.

4. The calibration device of any preceding claim, wherein the reference instruments are instruments that are used in statutory measurement which have the appropriate certifications and calibration certificates.

5. The calibration device of any preceding claim, wherein the control unit is configured to carry out the calibration phase and/or the data analysis phase.

6. The calibration device of any one of claims 1 to 4, wherein the control unit is configured to compare the one or more sensor signals to the one or more reference signals to produce one or more calibration parameters for each sensor.

7. The calibration device of claim 6, wherein the control unit is configured to apply the calibration parameter to the one or more sensors.

8. A method of calibrating sensors, wherein the method comprises:

3) carrying out a calibration phase comprising: a. introducing an analyte into the chamber to produce the analyte concentration within the chamber; b. causing the one or more sensors to measure the analyte concentration to produce one or more sensor signals; c. causing the one or more reference instruments to measure the analyte to produce one or more reference signals;

9. The method of claim 8, wherein the step of introducing the analyte comprises exposing the one or more sensors and the one or more reference instruments to ambient air, optionally for a period of from 1 hour to 24 hours.

10. The method of claim 8 or 9, wherein the method comprises one or more calibration phases.

11. The method of any one of claims 8 to 10, wherein the method comprises one or more data analysis phases.

12. The method of any one of claims 8 to 11, wherein the calibration phase comprises an analyte phase, wherein the analyte phase comprises the steps of a. obtaining an analyte range from the controller; b. introducing analyte into the mixing chamber prior to introduction into the chamber; c. introducing an analyte into the chamber to produce the analyte concentration within the chamber; d. reading the analyte levels from the reference instruments; e. checking that the analyte levels are within the target range; f. optionally adjusting the analyte level to within the target range; g. causing the one or more sensors to measure the analyte concentration to produce one or more sensor signals; and h. causing the one or more reference instruments to measure the analyte to produce one or more reference signals.

13. The method of claim 12, wherein the method further comprises the step of querying whether further a analyte phase is required.

14. The method of any one of claims 9 to 13, wherein the method further comprises informing a database that a sensor has entered the chamber.

15. The method of any one of claims 9 to 14, wherein the method further comprises the control unit allocating a unique identification number to each sensor that enters the chamber.