Bean drying temperature control method and bean drying device

Abstract

The invention discloses a bean drying temperature control method and a bean drying device, the bean drying temperature control method is suitable for a bean drying machine comprising a heating roller, and the method comprises the following steps: obtaining a plurality of bean drying data, wherein the plurality of pieces of bean drying data correspond to a plurality of continuous time points and each of the plurality of pieces of bean drying data comprises the current bean temperature, the current interlayer temperature of the heating roller and the current heating fire power, and inputting the target temperature curve and the plurality of pieces of bean drying data into a neural network to obtain a target heating fire power, and controlling the bean baking machine to heat according to the target heating fire power.

Description

Technical Field

[0001] This invention relates to a method for controlling the temperature of roasted beans and a bean roasting apparatus. Background Technology

[0002] The time and temperature required to roast coffee beans vary depending on a number of factors. Due to these complexities, the roasting process often requires experienced personnel to monitor it. Therefore, roasting beans is labor-intensive and prone to human error, which can cause the final product's flavor to deviate from the intended flavor.

[0003] Therefore, in order to achieve a consistent bean temperature profile during the heating process and obtain coffee beans with the same flavor, a proportional-integral-derivative (PID) control method is currently used to control the heating process of roasting beans. However, the PID control method still results in temperature differences in the beans during the heating process. Summary of the Invention

[0004] In view of the above, the purpose of the present invention is to provide a method and apparatus for controlling the temperature of roasted beans to solve the above problems.

[0005] A method for controlling roasting temperature according to an embodiment of the present invention is applicable to a roaster including a heating drum, comprising: acquiring multiple roasting data points, wherein the multiple roasting data points correspond to multiple consecutive time points and each of the multiple roasting data points includes the current roasting temperature, the current jacket temperature of the heating drum, and the current heating power; inputting a target temperature curve and the multiple roasting data points into a neural network to obtain a target heating power; and controlling the roaster to heat according to the target heating power.

[0006] A coffee roasting apparatus according to an embodiment of the present invention includes: a coffee roaster, a first temperature sensor, a second temperature sensor, and a control element. The coffee roaster includes a heating drum. The first temperature sensor is disposed in the interlayer of the heating drum for sensing multiple current interlayer temperatures. The second temperature sensor is disposed in the heating drum for sensing multiple current coffee temperatures. The control element is connected to the coffee roaster, the first temperature sensor, and the second temperature sensor, and is configured to perform the following actions: acquiring multiple coffee roasting data points, wherein the multiple coffee roasting data points correspond to multiple consecutive time points, and each of the multiple coffee roasting data points includes a corresponding value among multiple current coffee temperatures, a corresponding value among multiple current interlayer temperatures, and a current heating intensity; inputting a target temperature curve and the multiple coffee roasting data points into a neural network to obtain a target heating intensity; and controlling the coffee roaster to heat according to the target heating intensity.

[0007] In summary, based on the coffee roasting temperature control method and roasting apparatus of one or more embodiments described above, an appropriate heating intensity can be inferred from the current coffee roasting sensing data to achieve accurate coffee roasting temperature control. Accordingly, the coffee bean temperature can be effectively controlled to meet the target temperature. Furthermore, by using the current jacket temperature, which affects the coffee bean temperature, as one of the reference parameters for controlling the roaster's heating, the target heating intensity can be made more accurate.

[0008] The foregoing description of the contents of this disclosure and the following description of the embodiments are intended to demonstrate and explain the spirit and principles of the present invention, and to provide a further explanation of the claims of the present invention. Attached Figure Description

[0009] Figure 1 This is a block diagram of a bean roasting apparatus according to an embodiment of the present invention.

[0010] Figure 2 This is a schematic diagram of a heating drum according to an embodiment of the present invention.

[0011] Figure 3 This is a flowchart illustrating a method for controlling the temperature of roasted beans according to an embodiment of the present invention.

[0012] Figure 4 This is a schematic diagram of a roasting profile according to an embodiment of the present invention.

[0013] Figure 5 This is a schematic diagram of training data according to an embodiment of the present invention.

[0014] The attached figures are labeled as follows:

[0015] 1: Bean roasting device

[0016] 11: Coffee roaster

[0017] 12: First temperature sensor

[0018] 13: Second temperature sensor

[0019] 14: Control elements

[0020] 111: Heated drum

[0021] 111a: Baked Bean Curd

[0022] 111b: Inner Liner

[0023] 111c: Outer shell

[0024] C11~C17, C21~C25: Curves

[0025] T1~T6: Time points

[0026] S101, S103, S105: Steps Detailed Implementation

[0027] The following detailed description of the features and advantages of the present invention in the embodiments is sufficient to enable those skilled in the art to understand the technical content of the present invention and to implement it accordingly. Based on the disclosure, claims, and drawings in this specification, those skilled in the art can easily understand the related objects and advantages of the present invention. The following embodiments further illustrate the viewpoints of the present invention in detail, but are not intended to limit the scope of the present invention in any way.

[0028] Please refer to this as well. Figure 1 and Figure 2 ,in Figure 1 The diagram shows a block diagram of a bean roasting apparatus according to an embodiment of the present invention. Figure 2 This is a schematic diagram of a heating drum according to an embodiment of the present invention. Figure 1 As shown, the bean roasting device 1 includes a bean roaster 11, a first temperature sensor 12, a second temperature sensor 13, and a control element 14.

[0029] The coffee roaster 11 includes a heating drum 111, which may include a roasting drum 111a, an inner chamber 111b, and an outer chamber 111c. The roasting drum 111a is used to hold coffee beans. The coffee roaster 11 may also include one or more heating rods located between the roasting drum 111a and the inner chamber 111b for heating the coffee beans inside the roasting drum 111a.

[0030] A first temperature sensor 12 is disposed in the interlayer of the heating drum 111 to sense multiple current interlayer temperatures. For example, the first temperature sensor 12 may be disposed in the interlayer between the inner chamber 111b and the outer chamber 111c to sense the current interlayer temperature. Furthermore, since heat-absorbing cotton may be provided between the inner chamber 111b and the outer chamber 111c, the interlayer between the inner chamber 111b and the outer chamber 111c has heat storage capacity. Therefore, in addition to the heating rod, the heat storage in the interlayer can be regarded as another heat source for heating coffee beans. In other words, using the current interlayer temperature, which affects the temperature of coffee beans, as one of the reference parameters for controlling the heating of the roaster 11 can make the target heating intensity more accurate.

[0031] The second temperature sensor 13 is disposed in the heating drum 111, for example, in the roasting drum 111a of the heating drum 111, to sense the temperature of the coffee beans as the current bean temperature as described below.

[0032] Control element 14 is communicatively or electrically connected to roaster 11, first temperature sensor 12, and second temperature sensor 13. Additionally, control element 14 can store multiple candidate temperature curves, each a past roasting curve. Control element 14 is used to input roasting data indicating the current roasting state and the target temperature curve from the candidate temperature curves into a neural network to control the heating power of roaster 11 based on the output of the neural network, for example, to control the heating power of the aforementioned heating element. Control element 14 may include one or more processors, such as microcontrollers, microprocessors, central processing units, graphics processors, programmable logic controllers, or other processors with signal processing capabilities.

[0033] The neural network can be a fully connected neural network, and it can include one input layer, two hidden layers, and one output layer. The input layer can include 36 neurons, the two hidden layers can each include 20 neurons, and the output layer can include 1 neuron. The number of layers and neurons in the neural network described above are merely examples.

[0034] It should also be noted that the coffee roaster 11, the first temperature sensor 12, the second temperature sensor 13, and the control element 14 can be housed together in one machine; or, the coffee roaster 11, the first temperature sensor 12, and the second temperature sensor 13 can be housed together in one machine, while the control element 14 can be located outside the machine and connected to the coffee roaster 11, the first temperature sensor 12, and the second temperature sensor 13.

[0035] Please refer to this as well. Figure 1 and Figure 3 ,in Figure 3 This is a flowchart illustrating a method for controlling the temperature of roasted beans according to an embodiment of the present invention. Figure 3 As shown, the method for controlling the temperature of roasted beans includes: step S101: obtaining multiple roasted bean data; step S103: inputting the target temperature curve and multiple roasted bean data into a neural network to obtain the target heating power; and step S105: controlling the roaster to heat according to the target heating power.

[0036] In step S101, the control element 14 acquires multiple roasting data points, each corresponding to a series of consecutive time points. Each data point includes the current bean temperature, the current interlayer temperature of the heating drum 111, and the current heating power. In other words, one data point corresponds to one time point. The control element 14 acquires the current interlayer temperature from the first temperature sensor 12, the current bean temperature from the second temperature sensor 13 of the roaster 11, and the current heating power of the heating element from the roaster 11.

[0037] In addition, each of the multiple roasting data points may include one or more of the following: current vent temperature, current airflow speed, and current rotational speed of the heating drum 111. The roaster 11 may also include another temperature sensor located at the air outlet of the heating drum 111; the current vent temperature may be the temperature of the air outlet of the heating drum 111, and the current airflow speed may be the air velocity at the air outlet of the heating drum 111. Furthermore, when the weight of coffee beans is the same each time the roasting device 1 roasts, the roasting data may not include the weight of the coffee beans; conversely, when the roasting device 1 is used to roast different weights of coffee beans, the roasting data may include the weight of the coffee beans.

[0038] In step S103, the control element 14 inputs the target temperature curve and multiple roasting data points into the trained neural network to obtain the target heating power output by the neural network. The neural network can be a fully connected neural network. The target temperature curve is a curve indicating how the current bean temperature should change over time. Specifically, the target temperature curve input to the neural network is a complete curve (as shown in the following reference). Figure 4 The curve C11 described above, and the roasted bean data input to the neural network, are discrete information from multiple consecutive time points preceding the current time point. These multiple consecutive time points can be six time points, and the duration of each of these six time points can be 30 seconds, but this invention does not impose such a limitation.

[0039] The target temperature profile can be selected by the user from multiple candidate temperature profiles. These candidate temperature profiles may correspond to different roasting conditions, such as light, medium, or dark roasts, from which the user can choose one as the target temperature profile. The different roasting degrees mentioned are merely examples; roasting conditions may also include coffee bean variety, coffee bean flavor, etc., and this invention is not limiting.

[0040] Please refer to this as well. Figures 1 to 4 ,in Figure 4 This is a schematic diagram of a roasting profile according to an embodiment of the present invention. Figure 4This is presented as a roasting curve, showing the target temperature curve and multiple roasting data points of the completed roasting process. Curve C11 represents the target temperature curve; curve C12 represents the current bean temperature; curve C13 represents the current jacket temperature of the heating drum 111; curve C14 represents the current heating power (i.e., the heating power of the heating rod), where the current heating power of the roaster 11 is controlled by the control element 14, so the control element 14 can obtain the current heating power; curve C15 represents the current air vent temperature; curve C16 represents the current airflow, where the current airflow of the roaster 11 is controlled by the control element 14, so the control element 14 can obtain the current airflow; curve C17 represents the rate of rise (ROR) of the current bean temperature per unit time, where the unit time is, for example, 1 minute, but this invention is not limited thereto.

[0041] In step S105, the control element 14 controls the heating force of the heating rod of the coffee roaster 11 according to the target heating force output by the neural network, so as to control the heating of the coffee roaster 11.

[0042] Furthermore, as described above, the roasting data input to the neural network consists of discrete information from multiple consecutive time points preceding the current time point. Therefore, when executing the roasting temperature control method of one or more embodiments described above, the control element 14 can repeatedly execute the following... Figure 3 The steps S101, S103 and S105 shown are used to continuously obtain the latest roasting data and input the roasting data and multiple roasting data from previous consecutive time points into the neural network to obtain the new target heating power and control the roaster 11 to heat.

[0043] According to the bean roasting temperature control method and apparatus of one or more embodiments above, the appropriate heating intensity can be inferred based on the current bean roasting sensing data to achieve accurate bean roasting temperature control. Accordingly, the bean temperature can be effectively controlled to meet the target temperature.

[0044] In one embodiment, the method for controlling the temperature of roasted beans may further include training a neural network using multiple training data points, wherein the multiple training data points correspond to multiple consecutive historical time points and each training data point includes historical bean temperature, historical interlayer temperature of the heating drum, and historical heating power.

[0045] Please refer to this as well. Figure 1 and Figure 5 ,in Figure 5This is a schematic diagram of training data according to an embodiment of the present invention. Curve C21 represents the historical bean temperature; curve C22 represents the historical interlayer temperature of the heating drum 111; curve C23 represents the historical heating power (i.e., the heating power of the heating rod); curve C24 represents the historical airflow; and curve C25 represents the heating rate per unit time of the historical bean temperature, wherein the unit time is, for example, 1 minute, but is not limited thereto by the present invention.

[0046] The training of a neural network may involve multiple iterations, and each iteration may include inputting six training data points corresponding to six consecutive historical time points into the neural network. Figure 5 For example, one of the multiple iterations may include inputting six training data points corresponding to time points T1 to T6 into the neural network, while the next iteration may include inputting six training data points corresponding to time points T2 to T6 and the next time point after T6 into the neural network. The duration of the six consecutive historical time points may be 30 seconds, but this invention is not limited thereto.

[0047] Each of the multiple training data points may also include one or more of the historical vent temperature, historical airflow, and historical rotation speed of the heating drum 111. Specifically, the historical bean temperature, historical interlayer temperature of the heating drum 111, historical heating power, historical vent temperature, historical airflow, and historical rotation speed of the heating drum 111 in the training data can be of the same parameter type and obtained in the same way as the current bean temperature, current interlayer temperature of the heating drum 111, current heating power, current vent temperature, current airflow, and current rotation speed of the heating drum 111 in the roasting data, respectively, and therefore will not be elaborated further here. The difference between training data and roasting data is that training data corresponds to multiple consecutive historical time points, and training data is data from a completed roasting process, while roasting data is data generated in real time during the current roasting process.

[0048] In addition, when the weight of coffee beans is the same each time the roasting device 1 roasts coffee beans, the training data may not include the historical weight of coffee beans; conversely, when the roasting device 1 is used to roast coffee beans of different weights, the training data may include the historical weight of coffee beans.

[0049] It should also be noted that the parameters included in the roasted bean data can be the same as those included in the training data. Furthermore, the steps described above for training the neural network using the training data can be performed by... Figure 1 The control element 14 can execute the training, or it can be executed by another control element, and the neural network is output to the control element 14 after training is completed.

[0050] Through the above-described embodiments of training neural networks, the trained neural network can accurately predict the appropriate heating intensity based on similar experiences found in the training data, and then output the target heating intensity.

[0051] In summary, the roasting temperature control method and roasting apparatus according to one or more of the above embodiments can predict the appropriate heating intensity based on the current roasting sensor data, thereby achieving accurate roasting temperature control. Accordingly, the bean temperature can be effectively controlled to meet the target temperature. Furthermore, by using the current jacket temperature, which affects the coffee bean temperature, as one of the reference parameters for controlling the roaster's heating, the target heating intensity can be made more accurate. Moreover, through the above-described embodiment of training a neural network, the trained neural network can accurately predict the suitable heating intensity based on similar experiences found in the training data, and then output the target heating intensity.

[0052] While the present invention has been disclosed above with reference to the foregoing embodiments, it is not intended to limit the invention. Any modifications and refinements made without departing from the spirit and scope of the invention are within the scope of patent protection of the present invention. For a description of the scope of protection defined in the present invention, please refer to the appended claims.

Claims

1. A method for controlling the temperature of roasted beans, applicable to a roaster including a heating drum, characterized in that, comprising: obtaining a plurality of roasting data, wherein the plurality of roasting data correspond to a plurality of consecutive time points and each of the plurality of roasting data comprises a current bean temperature, a current bed temperature of the heating drum, and a current heating power; inputting a target temperature profile and the plurality of roasting data into a neural network to obtain a target heating power; and controlling the roaster to heat according to the target heating power.

2. The bean temperature control method according to claim 1, wherein further comprising: training the neural network using a plurality of training data, wherein the plurality of training data correspond to a plurality of consecutive historical time points and each of the plurality of training data comprises a historical bean temperature, a historical bed temperature of the heating drum, and a historical heating power.

3. The bean temperature control method according to claim 2, wherein each of the plurality of training data further comprises one or more of a historical tuyere temperature, a historical air flow, and a historical rotation speed of the heating drum.

4. The bean temperature control method according to claim 3, wherein each of the plurality of training data further comprises a historical coffee bean weight.

5. The bean temperature control method according to claim 1, wherein the neural network is a fully connected neural network.

6. A bean roasting apparatus, characterised in that, comprising: a roaster comprising a heating drum; a first temperature sensor disposed at a bed of the heating drum for sensing a plurality of current bed temperatures; a second temperature sensor disposed at the heating drum for sensing a plurality of current bean temperatures; a control element connected to the roaster, the first temperature sensor, and the second temperature sensor, the control element configured to perform: obtaining a plurality of roasting data, wherein the plurality of roasting data correspond to a plurality of consecutive time points and each of the plurality of roasting data comprises a corresponding one of the plurality of current bean temperatures, a corresponding one of the plurality of current bed temperatures, and a current heating power; inputting a target temperature profile and the plurality of roasting data into a neural network to obtain a target heating power; and controlling the roaster to heat according to the target heating power.

7. A bean roasting apparatus as claimed in claim 6, wherein the neural network is generated by training using a plurality of training data, wherein the plurality of training data correspond to a plurality of consecutive historical time points and each of the plurality of training data comprises a historical bean temperature, a historical bed temperature of the heating drum, and a historical heating power.

8. A bean roasting apparatus as claimed in claim 7, wherein each of the plurality of training data further comprises one or more of a historical tuyere temperature, a historical air flow, and a historical rotation speed of the heating drum.

9. A bean roasting apparatus as claimed in claim 8, wherein each of the plurality of training data further comprises a historical coffee bean weight.

10. A bean roasting apparatus as claimed in claim 6, wherein the neural network is a fully connected neural network.

Images