A method for controlling a coffee machine
The coffee machine control method uses an electronic controller with a machine learning algorithm to dynamically adjust rinse cycles, addressing ineffective cleaning and enhancing coffee quality by adapting to user habits.
Patent Information
- Application Number
- PCT/EP2025/055804
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2024-03-29
- Filing Date
- 2025-03-04
- Publication Date
- 2025-10-02
AI Technical Summary
Existing coffee machines have ineffective rinse cycles that fail to adequately clean the hydraulic circuit and maintain optimal organoleptic properties of the coffee, requiring complex modifications or aggressive cleaning agents.
A method for controlling a coffee machine using an electronic controller with a memory and machine learning algorithm to dynamically select and execute customizable startup and shutdown cycles, including water quantity, temperature, and flow rate, to enhance cleaning and coffee quality without structural changes.
The method ensures effective cleaning of the hydraulic circuit and improves coffee quality by adapting rinse cycles based on user habits, reducing the need for aggressive agents and structural modifications.
Smart Images

Figure EP2025055804_02102025_PF_FP_ABST
Abstract
Description
[0001] A METHOD FOR CONTROLLING A COFFEE MACHINE
[0002] DESCRIPTION
[0003] The present invention relates to a method for controlling a coffee machine.
[0004] A commercially available coffee machine of a known type includes a hydraulic circuit comprising, in sequence, a supply pump, a boiler, and a coffee brewer featuring a chamber for housing a coffee charge.
[0005] Currently, upon switching on and / or off — before the coffee charge is placed in its housing chamber or after it has been removed — a coffee machine can perform a hot rinse cycle of the hydraulic circuit.
[0006] This rinse cycle serves to clean the hydraulic circuit, thereby preventing the accumulation of dirt and debris as well as the proliferation of bacteria that could contaminate the beverage.
[0007] Additionally, when associated with the machine’s startup, the rinse cycle preheats the hydraulic circuit, ensuring that the first coffee dispensed into the cup reaches an acceptable temperature. The parameters defining the rinse cycle are generally preset at the factory and cannot be modified. This often results in ineffective rinsing, which fails to ensure adequate cleaning of the hydraulic circuit and optimal organoleptic properties of the coffee in the cup.
[0008] The technical problem addressed by the present invention is to develop a method for controlling a coffee machine that overcomes the technical drawbacks observed in the prior art.
[0009] As part of this technical objective, one aim of the invention is to implement a method for controlling a coffee machine that enables effective cleaning of the hydraulic circuit.
[0010] Another aim of the invention is to provide a method for controlling a coffee machine that ensures efficient cleaning of the hydraulic circuit without requiring complex structural modifications and without the need for particularly aggressive or non-eco-ffiendly cleaning agents. A further aim of the invention is to develop a method for controlling a coffee machine that not only ensures effective cleaning of the hydraulic circuit but also enhances the desired organoleptic characteristics of the coffee dispensed into the cup.
[0011] The technical problem, as well as the aforementioned and other objectives, are achieved through a method for controlling a coffee machine comprising an electronic controller, a user interface, and a hydraulic circuit, wherein the electronic controller includes a memory in which a plurality of selectable dispensing cycles is stored, the user interface comprises a display screen for visualization, the hydraulic circuit consists of, in sequence, a supply pump, a boiler, and a coffee brewer with a chamber for housing a coffee charge, characterised in that in the electronic controller’s memory is stored a plurality of selectable startup cycles that can be selectively activated when the coffee machine is turned on, these cycles including at least a predefined amount of water dispensed into the hydraulic circuit by the supply pump in the absence of a coffee charge in the housing chamber, and a predefined heating temperature of the boiler in the absence of a coffee charge in the housing chamber, and in that said electronic controller records operational data in said memory, including the number and types of dispensing cycles executed, and when these cycles were performed, and using a machine learning algorithm, it learns from these data which startup cycle to select and execute, or suggests a cycle to the user via said interface.
[0012] The memory of the electronic controller may also store shutdown cycles that can be selectively activated when turning off the coffee machine. These cycles include at least a predefined amount of water dispensed into the hydraulic circuit by the supply pump in the absence of a coffee charge in the housing chamber, and a predefined heating temperature of the boiler in the absence of a coffee charge in the housing chamber, the electronic controller learning from these data through the machine learning algorithm which shutdown cycle to select and execute, or suggests its execution to the user via said interface. Therefore, according to the teachings of the invention, the electronic controller dynamically adapts the selection of the startup cycle to be executed or suggested based on the historical operating data of the coffee machine.
[0013] If necessary, the electronic controller can also dynamically adjust the selection of the shutdown cycle, executing or proposing it to the user based on the machine's past usage data.
[0014] As the user’s habits change, the machine learning algorithm can advantageously adjust the selection of the appropriate startup or shutdown cycle, either executing it automatically or suggesting it to the user.
[0015] Other features of the present invention are further defined in the subsequent claims.
[0016] Additional features and advantages of the invention will become more apparent from the description of a preferred but non-exclusive embodiment of the coffee machine control method, which is illustrated for reference and not limitation in the attached drawings, wherein: figure 1 schematically shows the coffee machine implementing the control method.
[0017] Referring to the aforementioned figures, a control method for a coffee machine is illustrated, generally referred to by reference number 1.
[0018] The coffee machine 1 comprises an electronic controller 17, a hydraulic circuit 6 for coffee dispensing, a user interface 2.
[0019] The user interface 2 includes a power button 3 for turning the machine on and off, a display screen 4, selection buttons 5.
[0020] The user interface 2 also features a communication module 19, which is configured to remotely communicate with the user. For example, it may be an RF or wired transmitter that connects via the Internet to a software application installed on an electronic device such as a smartphone.
[0021] The hydraulic circuit 6 includes, in sequence, a supply pump 7, a boiler 8, and a coffee brewer 9 with a chamber 10 for housing a coffee charge 11. The supply pump 7 is connected to a water source, which may be either a water tank 12 or a direct connection to a water supply network.
[0022] A flow meter 16 is positioned between the supply pump 7 and the water tank 12 to monitor water flow.
[0023] The coffee brewer 9 is connected to an external dispenser 13, which is designed and positioned to dispense coffee into a cup 14 placed on a cup holder 15.
[0024] The electronic controller 17 includes a memory, where a plurality of m dispensing cycles Dj (j=l, 2, ...m) is advantageously stored, and selectable.
[0025] The dispensing cycles Dj may include, for example, a DI espresso coffee dispensing cycle, a D2 double coffee dispensing cycle, a D3 long coffee dispensing cycle, a D4 american coffee dispensing cycle, a D5 cold coffee dispensing cycle.
[0026] Additionally, the memory of the electronic controller 17 advantageously stores a plurality of n startup cycles Ci (i=l, 2, ...n), which can be selectively activated when turning on the coffee machine 1.
[0027] The startup of the coffee machine 1 can be requested by the user via the user interface 2, either physically using the power button 3 or remotely, for instance, from a smartphone through the transmission module 19.
[0028] Similarly, the shutdown of the coffee machine 1 can be requested by the user through the user interface 2, either by pressing the power button 3 or remotely, for example, via a smartphone through the transmission module 19.
[0029] Additionally, the electronic controller 17 may include an auto-shutdown circuit that deactivates the machine after a predetermined time has elapsed since it was turned on or since the last dispensing cycle was executed.
[0030] The startup cycles Ci are defined by at least a water quantity value Qi (i=l, 2, ...n), representing the amount of water dispensed into the hydraulic circuit 6 by the supply pump 7, in the absence of the coffee charge 11 in the housing chamber 10 of the brewing unit 9 and a heating temperature value Ti of the boiler 8, in the absence of the coffee charge in the housing chamber 10 of the brewing unit 9.
[0031] Preferably, the startup cycles Ci are also defined by a flow rate value Pi of the supply pump 7. Furthermore, the startup cycles Ci are also defined by an execution time value ti.
[0032] If the pump 7 and the boiler 8 are activated / deactivated at different times, ti is the sum of the time when only one of the two (either the pump 7 or the boiler 8) is active and the time when both are simultaneously active.
[0033] If the pump 7 and boiler 8 are activated and deactivated simultaneously, then ti corresponds to the period during which both are active.
[0034] Advantageously, the electronic controller 17 stores operating data of the coffee machine 1, including which dispensing cycles Dj were executed and when these cycles were executed and using a machine learning algorithm, the controller learns from this data to determine which startup cycle Ci to select and execute or to suggest to the user via the interface 2.
[0035] The coffee machine 1 may also feature a manual selection mode for the startup cycle Ci through the interface 2.
[0036] In manual selection mode, unlike the automatic selection mode described above, the user’s choice is not correlated with the historical operation of the coffee machine 1.
[0037] The interface 2 may include a selection button 5 that allows the user to switch between automatic and manual selection modes for the startup cycles Ci.
[0038] Some startup cycles Ci may involve the simultaneous activation of both the supply pump 7 and the boiler 8 or others may activate them asynchronously.
[0039] In the case of asynchronous activation, the boiler 8 for example may be activated before the supply pump 7 and turned off before the supply pump 7. Certain startup cycles Ci may include the selective activation of either only the supply pump 7 or the boiler 8, for example only the supply pump 7 for a cold water rinse, or only the boiler 8 for heating the hydraulic circuit 6 without a rinse cycle.
[0040] The startup cycles Ci include at least a Cl hot rinse cycle with water quantity QI, temperature Tl, and flow rate Pl, a C2 hot rinse cycle with water quantity Q2 < QI, temperature T2 > Tl, and flow rate P2 < Pl, a C3 hot rinse cycle with water quantity Q3 < QI, temperature T3 < Tl, and flow rate P3 > Pl, a C4 hot rinse cycle with water quantity Q4 > QI, temperature T4 > T2, and flow rate P4 < Pl, a C5 startup cycle with no water rinse (Q5 = 0).
[0041] The above startup cycles Ci, in quantitative terms, may include the following values:
[0042] Cl: 50 ml < Ql< 80 ml; 60 °C <T1< 80 °C; 4 ml / s <P1< 5 ml / s; 45 s <tl< 60 s
[0043] C2: 10 ml < Q2< 40 ml; 80 °C <T2< 100 °C; 0,5 ml / s <P2< 2 ml / s; 20 s <t2< 50 s
[0044] C3: 10 ml < Q3< 30 ml; 40 °C <T3< 60 °C; 5 ml / s <P3< 6 ml / s; 10 s <t3< 20 s
[0045] C4: 80 ml < Q4< 150 ml; 90 °C <T4< 110 °C; 1 ml / s <P4< 2 ml / s; 60 s <t4< 120 s
[0046] C5: Q5=0; t5= 20 s
[0047] The Cl startup cycle is a default hot rinse cycle that broadly adapts to all dispensing cycles Dj.
[0048] The C2 startup cycle is particularly well suited for the espresso coffee dispensing cycle DI, as it features a higher heating temperature and a very low flow rate compared to Cl, thereby increasing the contact time of water with the components of the hydraulic circuit 6.
[0049] The C3 startup cycle, which is of short duration, is particularly well suited for the long coffee dispensing cycle D2.
[0050] The C4 startup cycle is particularly effective for deep cleaning and sanitization of the hydraulic circuit 6.
[0051] The C5 startup cycle is particularly well suited for a remotely selected dispensing cycle Dj, where the cup 14 is positioned under the coffee dispenser 13 before the coffee machine 1 is turned on., and so this prevents any rinse water, which is usually dispensed through the same coffee dispenser 13, from being discharged into the cup 14.
[0052] The startup cycles Ci may include additional cycles, such as a C6 water-saving startup cycle, where 10 ml < Q6 < 15 ml; 50°C < T6 < 80°C; 2 ml / s < P6 < 3 ml / s; t6 = 30 s.
[0053] A typical example of the operation of the coffee machine 1 is as follows:
[0054] During the initial learning phase, which may last for a certain number of startup cycles of the coffee machine 1, the electronic controller 17 automatically executes the default startup cycle Cl at each startup.
[0055] Alternatively, during the initial learning phase, the electronic controller 17 may periodically prompt the user, for example, via a message on the display screen 4, asking whether to modify the default startup cycle Ci and which cycle Ci to use when turning on the coffee machine 1.
[0056] The user, using the selection buttons 5, can confirm the selection of the default startup cycle Cl, or modify the selection.
[0057] The selected or modified startup cycle Ci setting can then be applied to the subsequent startup cycle / s Ci of the coffee machine 1.
[0058] Once the machine learning algorithm has collected a significant amount of data and learned the user’s habits, the coffee machine 1 can automatically and autonomously select the startup cycle Ci to be executed at each startup.
[0059] Consider the following scenario:
[0060] The initial learning phase involves collecting data from 500 machine startups.
[0061] The training of the machine learning algorithm continues even after these 500 startups.
[0062] Over the first six months of use, the coffee machine 1 was turned on 500 times, with a frequency of 2-3 times per day. The daily startup times are statistically distributed in a first time slot between 7:00 AM and 8:00 AM, a second time slot between 1:30 PM and 2:30 PM, a third time slot between 5:00 PM and 6:00 PM.
[0063] In the first time slot, the selection of a DI espresso coffee dispensing cycle via the selection buttons 5 has statistically prevailed, in the second time slot, the remote selection of a DI espresso coffee dispensing cycle has statistically prevailed, in the third time slot, the selection of a D2 long coffee dispensing cycle via the selection buttons 5 has statistically prevailed.
[0064] Once the user's behavior has been learned through the analysis of these data, the machine learning algorithm schedules the execution of the C2 startup cycle in association with the startup of the coffee machine 1 in the first time slot, the execution of the C5 startup cycle in association with the startup of the coffee machine 1 in the second time slot, and the execution of the C3 startup cycle in association with the startup of the coffee machine 1 in the third time slot.
[0065] Additionally, the machine learning algorithm schedules the execution, at a specific startup — particularly during the startup of the coffee machine 1 in the first time slot — of a C4 sanitization and deep cleaning startup cycle in place of a C2 startup cycle. This is particularly convenient because the rinse water temperature set in the C4 startup cycle is also suitable for optimizing the outcome of the DI dispensing cycle, which is typically requested in the first time slot.
[0066] The C4 sanitization and deep cleaning cycle may be executed automatically, for example after a predetermined number of machine startups, but only when a predefined number of specific dispensing cycles Dj has also been executed.
[0067] For example, the machine learning algorithm may schedule the execution of a C4 cycle for deep cleaning and sanitization in the morning slot after X coffee machine 1 startups, if at least Y dispensing cycles Dk have been executed, otherwise at the next startup where the Y dispensing cycles Dk have been completed. The electronic controller 17 can automatically execute the scheduled startup cycle or notify the user via a message on the display screen 4, informing them of the scheduled startup cycle and requesting confirmation before execution.
[0068] In the case of a notification to the user via a message on the display screen 4, informing them that a specific startup cycle has been scheduled, the user's confirmation may apply to the next startup cycle executed in the same time slot.
[0069] In the case of a notification to the user via a message on the display screen 4, informing them that a specific startup cycle has been scheduled, the user's denial of confirmation may apply to the next startup cycle executed in the same time slot and may result in the user manually selecting their preferred startup cycle for the next startup in the same time slot.
[0070] For example, if the display screen 4 notifies that the Cl startup cycle has been scheduled and the user confirms (via a physical confirmation button 5 or a virtual confirmation button on the display screen 4, if it is a touchscreen), the user's confirmation applies to the next startup cycle executed in the same time slot.
[0071] If the user does not confirm (by failing to press the physical or virtual confirmation button within a certain period or by pressing a physical or virtual denial button), they are prompted to manually select the desired startup cycle, for example, C2, for the next startup in the same time slot.
[0072] If the user does not manually select a startup cycle through the interface 2, the same current startup cycle Cl will be executed at the next startup in the same time slot.
[0073] Preferably, within the automatic startup cycle selection mode, the execution mode of the scheduled startup cycle, either automatic or requiring user confirmation, can be set by the user through a selection button 5.
[0074] The same principles described above for the startup cycle also apply to the shutdown cycle of the coffee machine 1. Therefore, the memory of the electronic controller 17 also stores a plurality of n shutdown cycles C’i, which can be selectively activated when turning off the coffee machine 1, including at least a water quantity value Q’i, representing the amount of water dispensed into the hydraulic circuit 6 by the supply pump 7, in the absence of the coffee charge 11 in the housing chamber 10, and a heating temperature value T’i of the boiler 7, in the absence of the coffee charge 11 in the housing chamber 10.
[0075] It is preferable that the shutdown cycles C’i are also defined by a flow rate value P’i of the supply pump 7.
[0076] Additionally, the shutdown cycles C’i are preferably also defined by an execution time value t’i. The execution time t’i, if the pump 7 is activated / deactivated out of phase with the boiler 8, corresponds to the time during which only one between the pump 7 and the boiler 8 is active, plus the time during which both are active simultaneously.
[0077] Obviously, the execution time t’i, if the pump 7 and the boiler 8 are activated and deactivated simultaneously, corresponds to the period during which both are active.
[0078] The shutdown cycles C’i described above, in quantitative terms, may include the following values: C’I: Q’l= 50 ml; T’l= 60 °C; P’l= 5 ml / s; t’l= 10 s
[0079] C’2: Q’2= 40 ml; T’2= 60 °C; P’2 = 4 ml / s; t’2= 10 s
[0080] C’3: Q’3= 60 ml; T’3= 90 °C; P’3= 3 ml / s; t’3= 20 s
[0081] C’4: Q’4= 0; T’4= -; P’4= 0; t’4= 0
[0082] C’5: Q’5= 80 ml; T’5= 90 °C; P’5= 2,7 ml / s; t’5= 30 s
[0083] C’6: Q’6= 20 ml; T’6= 60 °C; P’6= 5 ml / s; t’6= 4 s
[0084] The electronic controller 17, in this case as well, learns from the data through the machine learning algorithm which shutdown cycle C’i to select and execute, or to propose its execution to the user via the interface 2. Preferably, the machine learning algorithm establishes an association between a shutdown cycle C’i and a startup cycle Ci, so that the machine learning algorithm schedules the execution of a startup cycle Ci when the coffee machine is turned on and the corresponding shutdown cycle C’i when it is subsequently turned off.
[0085] Specifically, in the present case, the shutdown cycle C’I is associated with the startup cycle Cl, the shutdown cycle C’2 is associated with the startup cycle C2, the shutdown cycle C’3 is associated with the startup cycle C3, the shutdown cycle C’4 (which in this case does not include either a rinse or heating cycle) is associated with the startup cycle C4, the shutdown cycle C’5 is associated with the startup cycle C5, the shutdown cycle C’6 is associated with the startup cycle C6.
[0086] The coffee machine 1 features a manual selection mode not only for the startup cycle Ci but also for the shutdown cycle C’i, which can be accessed through the interface 2.
[0087] The interface 2 may therefore include a selection button 5, allowing for reversible switching between the automatic selection mode and the manual selection mode for both the startup cycles Ci and the shutdown cycles C’i.
[0088] The coffee machine control method, as conceived, is subject to numerous modifications and variations, all falling within the scope of the inventive concept. Additionally, all details may be replaced with technically equivalent elements.
[0089] In practice, the materials used and the dimensions can be adapted as needed, depending on specific requirements and the state of the art.
Claims
CLAIMS1. A method for controlling a coffee machine (1) comprising an electronic controller (17), a user interface (2), and a hydraulic circuit (6), wherein the electronic controller (17) includes a memory storing a plurality of selectable dispensing cycles (Dj), the user interface (2) comprises a display screen (4), the hydraulic circuit (6) comprises, in sequence, a supply pump (7), a boiler (8), and a coffee brewer (9) featuring a housing chamber (10) for a coffee charge (11), characterized in that the memory of the electronic controller (17) stores a plurality of startup cycles (Ci), which can be selectively activated when the coffee machine (1) is turned on, including at least a quantity value (Qi) of water dispensed into the hydraulic circuit (6) by the supply pump (7) in the absence of the coffee charge (11) in the housing chamber (10) and a heating temperature value (Ti) of the boiler (8) in the absence of the coffee charge (11) in the housing chamber (10), and in that said electronic controller (17) stores operating data in the memory, including the number and type of dispensing cycles (Dj) executed and the time at which they were executed, and using said data learns through a machine learning algorithm which startup cycle (Ci) to select and execute or propose to the user for execution via said user interface (2).
2. Method for Controlling a Coffee Machine (1) according to Claim 1 characterized in that the startup cycles (Ci) include at least an execution time value (ti).
3. Method for Controlling a Coffee Machine (1) according to any of the Previous Claims characterized in that the startup cycles (Ci) include at least a flow rate value (Pi) for the supply pump (7).
4. Method for Controlling a Coffee Machine (1) according to any of the Previous Claims characterized in that the coffee machine (1) features a manual selection mode for the startup cycle (Ci) via the user interface (2).
5. Method for Controlling a Coffee Machine (1) according to any of the Previous Claims characterized in that the startup cycles (Ci) provide for the activation of both the supply pump (7) and the boiler (8).
6. Method for Controlling a Coffee Machine (1) according to any of Claims 1 to 4 characterized in that the startup cycles (Ci) provide for the selective activation of either the supply pump (7) or the boiler (8).
7. Method for Controlling a Coffee Machine (1) according to any of the Previous Claims characterized in that the startup cycles (Ci) include at least a Cl cycle with a rinse using a water quantity QI, at temperature Tl, and flow rate Pl, a C2 cycle with a rinse using a water quantity Q2 < QI, at temperature T2 > Tl, and flow rate P2 < Pl, a C3 cycle with a rinse using a water quantity Q3 < QI, at temperature T3 < Tl, and flow rate P3 > Pl, a C4 cycle with a rinse using a water quantity Q4 > QI, at temperature T4 > T2, and flow rate P4 < Pl, a C5 cycle with no water rinse (Q5 = 0).
8. Method for Controlling a Coffee Machine (1) according to any of the Previous Claims characterized in that the memory of the electronic controller (17) also stores a plurality of shutdown cycles (C’i), which can be selectively activated when turning off the coffee machine (1), including at least a water quantity value (Q’i) dispensed into the hydraulic circuit (6) by the supply pump (7) in the absence of the coffee charge (11) in the housing chamber (10) and a heating temperature value (T’i) of the boiler (8) in the absence of the coffee charge (11) in the housing chamber (10), and in that said electronic controller (17) learns from said data using said machine learning algorithm to determine which shutdown cycle (C’i) to select and execute or propose to the user via the user interface (2).
9. Coffee Machine (1) that implements a control method according to any of the previous Claims.
Citation Information
Patent Citations
Machine to dispense coffee-based beverages, and corresponding dispensing method and program
US20200367689A1
Method and system for encapsulated beverage preparation and brewing
US20220047112A1
AU2015101973A4