Improved sensing of temperature in a concrete delivery vehicle

Abstract

A computer implemented method for monitoring the temperature of concrete within the interior of a mixing drum 104 of a mixing vehicle 100, where temperature measurement signals are received from at least one indirect temperature sensor 102 mounted to detect the temperature on an external surface drum / vehicle and these signals are processed using a thermodynamic model / algorithm to determine the temperature of the concrete within the drum. Processing of the temperature measurement signals may utilise a Bayesian filter, possibly in the form of a Kalman filter. Indirect temperature sensor 102, in the form of a thermocouple or resistance temperature detector (RTD), is connected to wireless communications module 101 by connector 107. A second wireless module 105 connects to module 101, as does a possible third wireless module 106, to maintain communication with control module 103. A method for calculating an improved temperature estimate from an indirect temperature in combination with a direct temperature sensor is also disclosed (see Fig 6).

Description

FIELD OF THE INVENTION The present invention generaiiy relates to concrete delivery vehicles which mix and deliver concrete. Specifically, the present invention relates to an improved estimation of the temperature of the transported concrete using an indirect temperature sensor. BACKGROUND OF INVENTION The setting of concrete represents a gradual stiffening process where concrete ceases to behave as a liquid and transforms to a rigid material. The setting behaviour is of practical importance in determining the time available for transporting, placing and compacting concrete at the jobsite. This transient behaviour of fresh concrete is primarily attributed to the hydration process which commences as soon as water comes in contact with cement. The hydration process is composed of various reactions taking place simultaneously at the molecular level with attendant effect on the engineering properties on the macrostructural level. For instance, the hydration process is primarily responsible for the loss of workability i.e., slump loss. This hydration process is exothermic and sensitive to temperature. Warmer ambient temperature conditions result in higher concrete temperatures which leads to faster reaction rates. Increased temperature initially accelerates the dissolution of cement phases leading to a faster formation of hydration compounds. This implies that the transition from fresh to hardened state would take place sooner when the concrete temperature is higher, and this can significantly shorten the time available to transport, place, consolidate and finish concrete. Conversely, colder temperatures result in slower reaction rates which retard microstructural development and strength gain. Slower hydration rates also imply extended stiffening times and lower rates of concrete strength development, which can lead to delays in finishing as well as subsequent construction operations. In addition, if the temperature of concrete at the time of placement is below freezing, water molecules may freeze leading to expansion which can compromise overall hydration process and strength development. Because of the significant influence of concrete temperature on the properties of freshly mixed concrete, specifications place limits on the temperature at the time of pour. Installing temperature measurement systems on delivery trucks provides the opportunity to monitor concrete temperature whilst in transit to the jobsite. A number of systems designed to measure on-truck concrete temperature have been described in the literature. Generally, on-truck temperature measurement sensors and systems may be categorised as direct or indirect. Direct measurement sensors and systems make contact with the concrete in the drum, either by placement of sensors within the drum or by mounting through the drum such that a probe makes direct contact with the concrete. SUMMARY OF THE INVENTION According to one aspect of the present invention, there is provided a computer implemented method for monitoring the temperature of concrete within the interior of a mixing drum of a concrete delivery vehicle comprising: receiving temperature measurement signals from at least one indirect temperature sensor mounted to detect the temperature of an external surface of the delivery vehicle; and processing the temperature measurement signals using a thermodynamic model for the concrete to determine the temperature of the concrete within the mixing drum. The processing of the indirect temperature sensor measurement using a thermodynamic model for the concrete improves the fidelity of the indirect temperature estimate. In other words, the present invention allows a more complex and expensive temperature sensor to be replaced with a simpler and less expensive indirect temperature sensor combined with a computer-implemented method for producing an improved temperature estimate from the indirect sensor. The at least one indirect temperature sensor may be mounted by attaching it to the exterior surface of the delivery vehicle (for example the mixing drum), such that the sensor measures the temperature of the exterior surface with which it is in contact. Alternatively, the sensor may be mounted by attaching it to another part of the delivery vehicle, positioned such that the sensor remotely detects the temperature of an exterior surface of another part of the delivery vehicle. For example, the indirect temperature sensor could be mounted on the chassis of the delivery vehicle, and positioned to detect the temperature of an exterior surface of the mixing drum by means of an infrared sensor pointed at the drum. These are indirect temperature measurements because the sensor does not directly measure or monitor the temperature of the concrete within the mixing drum. In order to directly measure the concrete temperature, a direct temperature sensor must have access to the interior of the mixing drum. The sensor must therefore be mounted inside the mixing drum, and exposed to high temperatures and unset concrete. Or alternatively, an access point must be provided through the mixing drum for a direct temperature sensing probe, for example by drilling a hole through the mixing drum. This makes the sensor more expensive and time-consuming to deploy onto or remove from a delivery vehicle. The presence of a direct sensor or sensor probe within the mixing drum may also disturb or interfere with the mixing of the concrete, for example by disrupting the flow of concrete in the drum. Preferably, the thermodynamic model for the concrete is obtained by a process comprising: mixing a sample of concrete; taking experimental measurements of the concrete sample over a period of time; recording the experimental measurement data; and selecting a value for one or more coefficients so as to fit the thermodynamic model to the measurement data. Preferably, one of the taken experimental measurements comprises a temperature measurement taken using a direct temperature sensor inserted into the concrete sample. A direct measurement of the concrete sample provides high accuracy and reliability, which improves the fidelity of the coefficients selected for the thermodynamic model and hence the fidelity of the thermodynamic model itself. The more accurate the thermodynamic model, the greater the fidelity of the indirect temperature estimate. Preferably, the processing of the temperature measurement signals comprises using a Bayesian filter. Bayesian filters allow a variable to be calculated by combining a current measurement with estimates of the variable at a previous point in time. They are advantageous for the present invention because the estimate of a Bayesian filter can be made to converge towards the real value of a variable with time. This ensures that the indirect temperature estimate becomes t accurate and reliable as the indirect sensor takes measurements over time. Preferably, the Bayesian filter is a Kalman filter. A Kalman filter can advantageously be efficiently computer-implemented, and can be configured to receive estimates from the indirect temperature sensors, store previous indirect temperature estimates, and output a current improved indirect temperature estimate. Preferably, the Kalman filter is defined by the equations: -                        1 °" indirect^-) —+(± y^Lirect^) + Ryindlrect(k ~ 1) R°* indirect^-) R + ^Indirect W Alternatively, the Kalman filter is defined by the equations: ~          _ \yindirect U&\ yindirecty *         - t \ WindirecM yindirect— Ayindirect (k ~ 1?^ ^■indireclW ~ ‘^indirect~ 1)-^’ + Q yindirect^ = (yW^indirect+ Ryindtrecti^ “ ^(StLiireccW + ®) ri ri Preferably, A ~ where c is a coefficient characterising at least one thermodynamic property of the concrete. Preferably, temperature measurement signals are also received from an ambient temperature sensor mounted to detect the temperature of the air surrounding the vehicle. By measuring and accounting for the ambient temperature of the air, and variations thereof, the fidelity of the indirect temperature estimate may be further improved. Alternatively, the Kalman filter is defined by the equations: Z2              / V”12                 \-1 (k-XyHr + R] indirect                      Indirect                  ' y. ,. ,(k)-y (k -- 1) + Kk(yQi) - Hy. ,. ,(fc-l)) •'indirect'- z z indirectv      7                     J indirect'- J) v”’2                      X”’2 > (fc) - ( / -- ^ / / ) ) (k - 1) indirect                          Indirect Preferably,         $), where a and b are coefficients characterising at least one property of the indirect temperature sensor or the ambient temperature sensor. Alternatively, temperature measurement signals are also received from a second indirect temperature sensor mounted to detect the temperature of an external surface of the delivery vehicle. A sensor measurement is always subject to noise or disturbances. By combining measurements from two indirect temperature sensors, the indirect temperature estimate becomes less susceptible to these fluctuations, and thereby more accurate. Preferably, the determination of the temperature of the concrete further comprises using at least one temperature estimate determined at a previous point in time. The previously determined temperature estimate can be stored and used when calculating the later indirect temperature estimate. For example, by comparing the current temperature measurement to previously calculated temperature estimates, anomalous and unphysical "spikes" of statistical noise can be identified and removed. Preferably, the temperature measurements are transmitted by means of a wireless communications device to a processor. This removes the need for cabling to be run over or within the delivery vehicle between the sensors and the processor. Advantageously, this further reduces the time and cost for installing or removing the components on a delivery vehicle. Preferably, the processing is performed at least partially at a remote location from the delivery vehicle, and preferably as part of a Cloud-base service. This advantageously reduces the complexity and cost of the processing unit on the concrete delivery vehicle. This invention relates to methods of processing the indirect on-truck measurement such that the resulting temperature estimate is closer to the concrete temperature recorded by a direct on-truck method. The monitoring system takes as inputs temperature readings from the temperature sensors and performs calculations based on a model of the concrete to improve the overall temperature reading fidelity. In embodiments of the present invention, the monitoring system comprises one or more temperature sensors attached to the truck in contact with the drum containing the concrete. An embodiment of the monitoring system may record the temperature readings from the temperature sensors and transmit this information using wireless methods. Another embodiment of the monitoring system may record the temperature readings and store them to some medium on the truck, such as a memory card, for subsequent processing. BRIEF DESCRIPTION OF THE DRAWINGS Examples of the present invention will now be described in detail with reference to the accompanying drawings, in which: Figure 1 shows a concrete mixing truck equipped with a simplified example of an indirect temperature monitoring system; Figure 2 shows a testbed concrete mixing truck with a system for measuring the temperature of concrete inside the drum both directly and indirectly, with the drum of the mixing truck shown in cross-section; Figure 3 shows a simplified representations of a configuration with two indirect temperature sensors, wherein both sensors are directly connected to a control module; Figure 4 shows a simplified representation of a configuration with two indirect temperature sensors, wherein both sensors are directly connected to a first wireless module, which communicates with a second wireless module directly connected to the control module; Figure 5 shows a simplified representations of a configuration with two indirect temperature sensors, wherein both sensors are directly connected to a respective wireless module, the modules communicating with a remote server Figure 6 shows a process flow diagram of a method for calculating an improved temperature estimate using an indirect temperature sensor in combination with a direct temperature sensor; Figure 7 shows a process flow diagram of a method for calculating an improved temperature estimate using an indirect temperature sensor only, without readings from a direct temperature sensor; Figure 8 is a process flow diagram showing a method of operation of a concrete mixing truck 100 equipped with an indirect temperature monitoring system, and; Figure 9 shows a graph of the temperature estimate against time for an indirect temperature sensor and a direct temperature sensor, before and after application of a processing method. DETAILED DESCRIPTION Indirect measurement sensors and systems infer concrete temperature by measuring the temperature of secondary objects or systems such as drum temperature or ambient temperature. Such indirect sensors and systems are not as accurate at measuring concrete temperature as direct methods. Improvement in the manner of indirect temperature measurement would lead to monitoring of concrete in a more accurate way whilst providing benefits in deployment time and cost. Figure 1 is a simplified example of a monitoring system design for determining the temperature of a concrete load within the drum 104 of a concrete mixing truck 100. The system includes one or more indirect temperature sensors 102 affixed to the drum 104 of the concrete mixing truck 100 which is connected to a first wireless module 101 by means of a connector 107. The indirect temperature sensor 102 may comprise a Resistance Temperature Detector (RTD) sensor, or alternatively a thermocouple or a thermistor could be used. The indirect temperature readings are wirelessly transmitted to a second wireless module 105 connected to a control module 103 on the truck, where the temperature readings can be processed to determine the temperature of a concrete load within the drum. In an alternative design, the monitoring system allows for some or ail functionalities of the control module 103 to be operated in a Cloud service over a wireless communications link with the truck. The wireless communications link between the control module 103 may be maintained through the second wireless module 105 connected to the control module, or through a third wireless communications module 106. The third wireless communications module may be configured for communication over a mobile cellular network, for example. Figures 3, 4, and 5 are simplified representations of configurations of an alternative embodiment, wherein rather than a single indirect temperature sensor 102, there comprise two or more indirect temperature sensors 102a and 102b. The indirect sensors 102a and 102b are mounted on the drum 104 of the concrete mixing truck 100. Figure 3 shows the indirect sensors 102a and 102b both connected by connectors 107 to the control module 103 onboard the truck 100, with the sensors, connectors and control module affixed to the drum 104. All processing takes place onboard the control module 103. Figure 4 shows the indirect sensors 102a and 102b connected through a connector 107 to the first wireless module 101, with the sensors, connector and first wireless module affixed to the drum 104. The data from the sensors is then transmitted wirelessly between the first wireless module 101 and the second wireless module 105 mounted on the truck 100. The second wireless module is connected to the control module 103, where the processing is carried out. Figure 5 shows each indirect sensor 102a and 102b connected to its own wireless module 101a and 101b. For ease of manufacture and affixing to the drum 104, each pair of indirect sensor and wireless module may be contained in a single device or housing. The wireless modules 101a and 101b transmit to a remote processor 108, which for example could be a Cloud-based server. Processing is then carried out onboard the processor 108. In addition to the one or more indirect temperature sensors 102, one or more ambient temperature sensors (not shown in the Figures) may also be mounted to the concrete mixing truck 100. Each ambient temperature sensor is mounted on the exterior of the concrete mixing truck, configured so that it can detect the temperature of the air surrounding the vehicle. Like the indirect temperature sensors 102, an RTD, thermocouple, or thermistor sensor could be used. As per the indirect temperature sensors 102 in Figures 3-5, the ambient temperature sensor may be directly connected to an onboard processor 103, or connected to a wireless module 101. A system and experimental procedure is described below, which gathers data used to calibrate the indirect temperature measurement method. A method for calculating an improved temperature estimate using a combination of a direct and indirect temperature estimate is also described. Figure 2 shows an exemplary testbed concrete mixing truck 200 used to gather experimental data for calibrating an indirect temperature measurement method. Temperature readings are taken from an indirect temperature sensor 203 affixed to the drum 201 of the testbed concrete mixing truck 200. Temperature readings are also taken from a direct temperature sensor 202, which comprises a probe 202a which protrudes though the skin of the drum 201. The probe 202a obtains a direct temperature measurement of the concrete load by physical contact between the probe and the concrete 204 within the drum 201. Figure 6 shows a method for calculating an improved temperature estimate from an indirect temperature sensor in combination with a direct temperature sensor. Temperature values are received over a time period T from a direct temperature sensor 202 and an indirect temperature sensor 202a, as shown in step 61. The average and standard deviation over n temperature readings are calculated 62 for each set of sensor values. The correlation between the indirect and direct temperature values is also calculated 63. Steps 22 and 23 could be performed in either order. One example demonstrating the processing 64 to improve the indirect temperature values is determined as follows: ~        —   0 direct ,         _ _       x _ y indirect P         ^indirect y indirect) ^direct ® indirect Where the improved indirect estimate is the correlation between the indirect and direct temperature values is, p, the standard deviation of the direct temperature value is Oab-ea, the standard deviation of the indirect temperature value is ect, the indirect temperature value is y;nc / jrgrt, the average of the indirect temperature values is Vindirect, and the average of the direct temperature values is xdirec[. The processing method described in Figure 6 is commonly referred to as Minimum Mean Squared Error (MMSE) filtering. In a first embodiment of the present invention, a Kalman filtering method is used to improve the fidelity of indirect temperature readings, as shown below and described with reference to Figure 7: In this example, a Bayesian filter is employed to improve indirect temperature sensor readings as follows. The temperature values from an indirect temperature sensor are received 71 and processed to determine their statistical information. The uncorrected variance of the indirect temperature values, ^hdirectOi) at the current timestep k can be calculated 72 for n indirect temperature readings according to the equation: indirect W =        --HQ indirect^ with otndtrect(j) being the corrected variance of the indirect temperature values from some previous timestep j, and Q is an empirically determined value representing a thermodynamic model of the concrete. Mathematically, Q represents the process noise of the system. The improved indirect temperature estimate yincUrect(JO at timestep k can be calculated 73 as follows: yindirect ('") yQc)<J*2 . ,(k) + ^y indirect   ~ 1) x . /                                itiuif trt,v x          / indirect(Rj T R where the variance of the indirect temperature measurements is o^rtirsct(fc), the measurement of the indirect temperature value is y(fc), the improved indirect temperature estimate from the previous timestep is y-ndtrectCk -1), and R is an empirically determined value. Mathematically, R represents the measurement noise of the system. The corrected variance of the indirect temperature measurements, ff-ndirec£(fc), is calculated 74 by combining the empirically determined value R with the uncorrected variance of the indirect temperature measurements, <J*lndirect(JO- _2 n.x ^indirect W rt ,   ,2 r? \ « +(^ indirect^ The improved indirect temperature estimate ytndirectQ^ and the corrected variance of the indirect temperature measurements ff^direc£(fc) are then stored 75 in memory, for use during the next timestep. Figure 8 is a process flow diagram showing a method of operation of a concrete mixing truck 100 equipped with an indirect temperature monitoring system. In order to calibrate the indirect temperature measuring method, Q and R must be found. It has been found that the optimal value of Q and R is specific to a given type of concrete. Furthermore, it has been found to be difficult to calculate or extrapolate the optimal value of Q and R for a new concrete type purely theoretically, even with knowledge of Q and R values for other types of concrete. Therefore, before indirect temperature monitoring can be performed with a new and previously uncalibrated type of concrete, Q and R must be found by means of an experimental calibration procedure 42. An exemplary experimental calibration procedure is as follows. Firstly, a sample of concrete is taken, for example from a mixing truck loaded with the new concrete type. Experimental measurements of the concrete sample are taken over a period of time. For example, a direct temperature sensor inserted into concrete sample could be used to intermittently measure the temperature. The concrete may be continuously mixed during the measurement time period, or in an alternative embodiment, mixed prior to the measurement time period and then allowed to settle during the measurements. The experimental measurements are recorded and plotted against time. A regression curve may be fitted to the data. Q and R are then set so as to maximise the fidelity of the indirect temperature estimate against the reading from the direct temperature sensor. One way to maximise fidelity would be to set Q and R so as to minimise the mean absolute error difference between the indirect temperature estimate and the direct temperature sensor reading. The concrete is then loaded 43 into the mixing truck 100. Readings are taken using the indirect temperature sensor 102 and sent to the control module 103 for processing. The improved indirect temperature estimate may then be calculated 45 according to the method of Figure 7. The improved temperature estimate, along with any other data or telemetry, is transmitted from the mixing truck to a remote server or directly to a user, in order to monitor the temperature of the concrete as the truck is driven to its destination. Based on the temperature history of the concrete, the user or server can command the truck to deliver the concrete load to an alternate destination. In a second embodiment of the present invention, we incorporate a-priori knowledge of concrete load type and the corresponding behaviour on temperature change into the Kalman filter equations of the first embodiment. A vector           is defined, which represents the state of the indirect temperature estimate at timestep k. yiniirect(k') can be written as: yindirect W            , . indirect^ Where ^ndtrectW is the indirect temperature estimate and y,ndire..t(fc) is the first derivative with respect to time of the indirect temperature estimate. The indirect temperature state estimate is then determined as foiiows: .Vindirect (k) — ^y indirect ” ijA ^ndirect(k) = ^'indirect (k - 1MT + Q yindirect^) = (j(W^tdirect (k) + ^mdirect^ ~ 1))(indirect GO + «) where A Is a state transition matrix relating the previous state estimate of y^dtrect at timestep k -1 to the current ymdirect at time k. The variable indirect is an error covariance matrix, with Q and R being empirically determined matrices. A can be expressed by the equation: where c is a coefficient representing the thermodynamic properties of a particular type of concrete. An exemplary value of c would be c = 6.4 x io~3. Figure 9 shows a graph of the temperature estimate against time for an indirect temperature sensor, before and after application of the processing method shown in Figure 6, and compares this to the direct temperature measurement. The post- processed estimate has a significantly improved correlation with the direct temperature measurement. In the case where there are multiple indirect temperature sensors, the Kalman filter approach set out in the above embodiments may easily be adapted for the purposes of sensor fusion, as will be appreciated by one skilled in the art. The one-dimensional value from the indirect temperature sensor y(k}, becomes a vector of length n, where n is the number indirect temperature sensors being inputted. The other coefficients 10 and statistical parameters are similarly transformed. In a third embodiment, the Kalman filter approach is adapted for the case where there comprise one or more ambient temperature sensors mounted to detect the temperature of the surrounding environment. Incorporating the ambient temperature sensor measurements into the equations of the first or second embodiment allows the UI ■“ filter to account for temperature changes in the surrounding environment. This further improves the fidelity of the indirect temperature estimate. The equations of the Kalman filter become: 20 '■sr'",                                      /           2                                          \ (k — I) Hr \ H y      (k — 1) Hr + Rj yindlrectW = Vlndirecl+ Kk      ~            ~ 1)) v”’2                      X"’2 >       (fc) - (I -- KhH) )       (fc - 1) 25 Where H is a matrix, which for an indirect sensor and an ambient sensor combination has the form: Where a and b are empirically determined values, for example a - b - 1. 30

Claims

1. A computer implemented method for monitoring the temperature of concrete within the interior of a mixing drum of a concrete delivery vehicle, comprising:receiving temperature measurement signals from at least one indirect temperature sensor mounted to detect the temperature of an external surface of the delivery vehicle; and,processing the temperature measurement signals using a thermodynamic model for the concrete to determine the temperature of the concrete within the mixing drum.

2. The method of claim 1, wherein the thermodynamic model for the concrete is obtained by a process comprising:mixing a sample of concrete;taking experimental measurements of the concrete sample over a period of time;recording the experimental measurement data; andselecting a value for one or more coefficients so as to fit the thermodynamic model to the measurement data.

3. The process of claim 2, wherein one of the taken experimental measurements comprises a temperature measurement taken using a direct temperature sensor-inserted into the concrete sample.

4. The method of any preceding claim, wherein the processing of the temperature measurement signals comprises using a Bayesian filter.

5. The method of claim 4, wherein the Bayesian filter is a Kalman filter.

6. The method of claim 5, wherein the Kalman filter is defined by the equations:° indirect(.^J ~     ~1 ,7indirect^,, .                     + Rytnd:rSct(k “ 1)y.^Ck)^--“------17 indirect^-)indirect (A)^indirect'-K' ~ *2 I Z? 777 R + O indirect^"7. The method of claim 5, wherein the Kalman filter is defined by the equations:lyindirectCWyindirect^) “   indirectZLiereM = A^dlrect(k - 1)4^ + Qy indirect (^ ) = (yCWLtirectCK) + indirect^ ” ^(^ndirectOQ + «)ri d8. The method of claim 7, wherein A -      , where c is a coefficient characterisingLU 1-1at least one thermodynamic property of the concrete.

9. The method of claim 5, wherein temperature measurement signals are also received from an ambient temperature sensor mounted to detect the temperature of the air surrounding the vehicle.

10. The method of claim 9, wherein the Kalman filter is defined by the equations:Kk --- 2 ft-liTI#) (k — i) Ht + r\ Indirect                      Indirect                  'y. ., .(k) = y. ..                  (y(k) — Hy. .. .(k — 1)^7 indirecty ' 7 indirectv             K v 7      7 indirectK      ')Z2                 2(A) - ( / - / <. / / ) > (A - 1) indirect                         indirect11. The method of claim 9, wherein 2), where a and b are coefficients \b 0 / characterising at least one property of the indirect temperature sensor or the ambient temperature sensor.

12. The method of claim 5, wherein temperature measurement signals are also received from a second indirect temperature sensor mounted to detect the temperature of an external surface of the delivery vehicle.

13. The method of any preceding claim, wherein the determination of the temperature of the concrete further comprises using at least one temperature estimate determined at a previous point in time.

14. The method of any preceding claim, wherein the temperature measurements are transmitted by means of a wireless communications device to a processor.

15. The method of any preceding claim, wherein the processing is performed at least partially at a remote location from the delivery vehicle, and preferably as part of a Cloud-base service.

Images