METHOD OF CONTROLLING AN ELECTRONIC POWER DEVICE DESIGNED TO TRANSFER CURRENT TO A LOAD TO REGULATE THE TEMPERATURE OF THE LOAD FOLLOWING THE EARLY DETECTION OF MALFUNCTIONING IN A THERMAL PROCESS OF A MATERIAL

Description

Title METHOD OF CONTROLLING AN ELECTRONIC POWER DEVICE SUITABLE FOR TRANSFER CURRENT TO A LOAD TO REGULATE THE LOAD TEMPERATURE FOLLOWING THE EARLY DETECTION OF MALFUNCTION IN A THERMAL PROCESS OF A MATERIAL I0206275-SP Owner: GEFRAN SpA METHOD OF CONTROLLING AN ELECTRONIC POWER DEVICE DESIGNED TO TRANSFER CURRENT TO A LOAD TO REGULATE THE TEMPERATURE OF THE LOAD FOLLOWING THE EARLY DETECTION OF MALFUNCTIONING IN A THERMAL PROCESS OF A MATERIAL DESCRIPTION

[0001] . TECHNOLOGICAL BACKGROUND OF THE INVENTION

[0002] . Scope of application

[0003] . The present invention relates, in general, to the field of power control models for industrial applications in sectors such as, for example, plastics processing, packaging, the food and beverage industry, the pharmaceutical sector, glass, metal and ceramic processing.

[0004] . In particular, the invention refers to an innovative method of controlling a power electronic device, usable in a control system of a thermal process for the processing of a material, capable of transferring current to a load, for example resistive, to regulate the temperature or thermal state of the processing process by detecting in advance any malfunctions of the processing process and adopting adequate resolution measures.

[0005] . Known art

[0006] . As is known, heat treatment processes or thermal processes can be divided into continuous thermal processes and cyclic thermal processes from the point of view of thermal regulation.

[0007] . For these two types of thermal processes, the occurrence of anomalies in the process can cause interruption of the production cycle, losses in efficiency (energy or material consumption), or can lead to a reduced quality of the products manufactured, in the event that such products do not comply with pre-established specifications. Non-compliant and reduced quality products are, for example, incorrectly sealed packages, unsterilized bottles, and incorrectly pasteurized food products.

[0008] . Some of the causes that can generate anomalies in thermal processes are linked to malfunctions of components of the thermal process control system.

[0009] . A first example of malfunction concerns thermal dispersions that can occur between a resistive load and elements containing the material to be heat treated, caused by insulation losses or mechanical problems. Such thermal dispersions can manifest themselves as variations in power supplied to the resistive load, independent of the presence of the material to be thermally processed, or as variations in the temperature of the material to be processed for the same power supplied to the resistive load.

[0010] . A second example of malfunction concerns the deterioration or breakdown of the resistive load, which can manifest itself as: short circuit, that is, the voltage applied to the resistive load tends to zero; total breakdown of the load, that is, the current applied to the resistive load tends to zero; partial breakdown of the load, that is, the current applied to the resistive load is less than pre-established standard process values; or in general loss of efficiency, whereby the power supplied to the material - and therefore the temperature of the material - decreases for the same electrical power transmitted to the load.

[0011] . A third example of malfunction concerns the malfunction of electrical connection elements between the AC power electronics or power controller and the resistive load, resulting in the electrical power transmitted to the resistive load being abnormally lower than the power output from the power controller. This difference results in a decrease in thermal energy delivered by the resistive load and, consequently, a decrease in the temperature of the material to be thermally processed for the same duty cycle on the current signal set by the power controller.

[0012] . A fourth example of malfunction concerns malfunctions of the power electronic device itself. This device is embodied, in particular, in a power electronic converter using thyristors. SCR (Silicon Controlled Rectifier) ​​connected to each other in a bridge.

[0013] . Such malfunctions manifest themselves in the inability of one or more of the thyristor devices themselves to deliver the same power at the same duty cycle received from the process controller and input power from the power supply network. Such malfunctions may be linked to aging phenomena, whereby the SCR devices dissipate more power to operate and, therefore, the output power decreases.

[0014] . Furthermore, phenomena may also occur involving an increase in the resistance of one of the thyristors of the electronic power converter or of the internal electrical connections of this device, resulting in a drop in the output current or even a cancellation of this current in the event of a thyristor breakdown. These phenomena are generally linked to the temperature of the thyristor junction or to the temperature of the substrate (Direct Copper Bonding or DCB) on which this device is mounted.

[0015] . A fifth example of malfunction concerns malfunctions that may involve the thermal process control algorithm, for example caused by a variation in the material to be thermally processed, which cannot be compensated for by an electronic thermal process control unit. Such malfunctions of the control algorithm have been found, in particular, in the case of the use of recycled materials for the production of plastic materials, or in the case of changes to the production system related to aging or other external factors.

[0016] . Methods and systems configured to detect / identify the malfunctions described above are known and already widely implemented. These methods and systems can be classified into three categories.

[0017] . A first category of methods and systems for detecting / locating malfunctions includes alarm signals transmitted by the power electronics device in the event of: absence of input voltage to the device; total or partial load failures; anomalous temperature values ​​detected on the device's electronics.

[0018] . A second category of malfunction detection / identification methods includes alarm signals activated by the electronic process control unit as a result of the system's inability to reach the target process temperature or following a drop in the temperature of the material to be processed, with the same duty cycle transmitted to the power electronics. These alarms derive from algorithms specific to the thermal process, but require being reset with each variation in the production process parameters. Furthermore, since these alarms can derive from multiple causes, they make it difficult to identify the physical phenomenon that determined the malfunction.

[0019] . A third category of malfunction detection / identification methods includes quality controls performed downstream of the production process itself, based for example on manual random checks, visual analysis using optical devices, or other mechanisms specific to the production process. Such quality controls, however, on the one hand, do not allow for the detection of all product anomalies and the production process malfunctions that may have caused them, and on the other, they are difficult to correlate with the original causes of the malfunctions within the production process.

[0020] . One of the main limitations of the known methods and systems described above for detecting malfunctions is linked to the fact that all three methods detect the malfunction in the process or the anomaly in the product only when such malfunction has occurred, therefore it has already caused: time losses (production interruption for the identification of the problem, the resolution of the problem and the recovery of the process), economic impacts (production delays, waste of material, management of defective products) or even reputational damage or legal consequences, in the case in which the defective products are not intercepted before being placed on the market.

[0021] . Furthermore, the second and third methods of detecting / locating malfunctions described are independent of the thermal process and therefore require subsequent analyses to identify the cause of the failure.

[0022] . In light of the above, there is a strong need for a new method for regulating the thermal state of a load in a thermal material processing process following the detection of malfunctions in the processing process.

[0023] . SUMMARY OF THE INVENTION

[0024] . An object of the present invention is to devise and provide a method of controlling a power electronic apparatus, usable in a control system of a thermal process for the processing of a material, capable of transferring current to a heating load, for example resistive, to regulate the temperature or thermal state of the heating load following an early detection of malfunctions in the processing process which allows to overcome, at least partially, the drawbacks complained of above with reference to the prior art.

[0025] . In particular, the aim of the invention is to provide a control method that allows the operation of the power electronic apparatus used in thermal material processing to be modelled in order to prevent, mitigate or correct malfunctions that may affect both the power electronic apparatus and the load powered by it.

[0026] . The method of the present invention is configured to generate corrections to the control parameters of the thermal process controllers in order to detect system malfunctions in advance, maintain the production process within the preset process parameters, at least for a significant and pre-determined time interval, and accelerate the identification of the cause of a malfunction.

[0027] . This object is achieved by implementing a control method of a power electronic device in accordance with claim 1.

[0028] . A particular object of the invention provides for the implementation, on the power electronic apparatus itself or on a further local intelligent element of the manufacturing process control system, of a functional block representing a digital twin converter of a real power converter included in the above-mentioned power electronic apparatus, capable of performing simulations on at least part of the components of the control system in a time interval following the instant of start-up or execution of the process which depends on the characteristics of the process, in accordance with claim 2.

[0029] . In relation to the different types of production process, in the following a distinction will be made between thermal processes that can be performed during time intervals of unpredictable duration after the start of the manufacturing process, or those that can be performed during time intervals of predictable duration after the start of the manufacturing process. In the latter case, such time intervals of predictable duration can be medium / long or short.

[0030] . For example, in the case of an extrusion process for the production of plastic sheets, once the processing temperature of the plastic material to be extruded has been reached, this temperature is maintained for days or even months. For this extrusion process, therefore, one can speak of a processing time interval of predictable duration, in particular of medium / long duration.

[0031] . In the case of processes in the packaging sector, fast processing cycles are typically expected, on the order of milliseconds up to tens of minutes. Therefore, these processes can be considered applications with processing time intervals of predictable duration, particularly of short duration, based on the type of production batch being processed.

[0032] . On the contrary, in the case of a steam production system, once the production temperature has been reached, to ensure that this temperature is maintained when a request from multiple users varies, it may be necessary to modulate the heating thermal power according to unpredictable timescales. Therefore, this type of thermal processing is characterised by processing time intervals of unpredictable duration.

[0033] . Another object of the invention is to provide a method that allows the operating conditions of the thyristors of the power electronic converter that equips the power electronic apparatus to be controlled to be detected in advance by using a digital twin of such thyristor power electronic converter and to use it in the event of malfunctions in a specific production process to modify - directly or through the intervention of an operator - the output parameters of an electronic process control unit and, at the same time, produce a set of information on the state of the system in order to allow the identification of such malfunctions.

[0034] . In particular, the method of the present invention provides for the use of three functional blocks: a digital twin functional block of the thyristor electronic converter (SCR), a feedback functional block for correcting the parameters of the thermal processing, and a functional block for analyzing the outputs of the digital twin block for identifying anomalies or malfunctions.

[0035] . The present invention also relates to a control system for a thermal process for processing a material which implements the proposed method, in accordance with claim 14.

[0036] . Some advantageous embodiments of the method and system are the subject of the dependent claims.

[0037] . BRIEF DESCRIPTION OF THE DRAWINGS

[0038] . Further features and advantages of the control method of a power electronic apparatus, usable in a control system of a thermal process for the processing of a material, to regulate the temperature or thermal state of a load following an advance detection and classification of malfunctions in the processing, will appear from the following description of preferred embodiments thereof, given for illustrative purposes only, with reference to the attached figures in which:

[0039] . - Figure 1 illustrates, by means of a block diagram, an example of a control system for a thermal process for working a material configured to implement the method of the present invention;

[0040] . - figure 2 illustrates, by means of a block diagram, an example of a power electronic apparatus, for example in alternating current, included in the system of figure 1, capable of transferring current to a load, to regulate the temperature (or thermal state) of the manufacturing process on the basis of control signals sent from a process controller to the power electronic apparatus;

[0041] . - Figure 3 illustrates, by means of a functional block diagram, the interaction between a real SCR or thyristor power electronic converter, a digital twin of the real converter, and a feedback algorithm configured to correct, following an early detection of process malfunctions, a duty cycle parameter applied to the real SCR power electronic converter from an output of the digital twin;

[0042] . - Figure 4 illustrates, by means of a functional block diagram, the interaction between the real thyristor or SCR-based power electronic converter, the digital twin electronic converter of the aforementioned real converter and a comparison algorithm configured to generate status information, related to a detected malfunction, based on an analysis of an output of the digital twin electronic converter;

[0043] . - Figure 5 illustrates, by means of a flowchart, an example of a general embodiment of the control method of a power electronic apparatus, usable in a control system of a thermal process for the processing of a material, to regulate the temperature or thermal state of a load following an early detection and classification of malfunctions of the processing by means of the digital twin electronic converter or “digital twin”, of the present invention;

[0044] . - figure 6 illustrates, by means of a flowchart, an example of a particular embodiment of a phase of detection and classification of malfunctions of the control method of figure 5;

[0045] . - figure 7 illustrates, by means of a flowchart, an example of a particular implementation of the control algorithm optimization phases and of the controlled / emergency shutdown phase of the control method of figure 5.

[0046] . In the above figures, equal or similar elements are indicated with the same numerical references.

[0047] . DETAILED DESCRIPTION

[0048] . With reference to figure 1, the numerical reference 100 indicates a control system for a thermal process for processing a material configured to implement the proposed control method 200, in an embodiment example. In the following, the control system for a thermal process 100 will also be referred to as a control system or, more simply, as a system. Furthermore, in the following, the term material will be used to indicate indifferently: metallic materials, plastic materials, glass, food products, such as bread and pasta, and in general all those materials that can be subjected to a thermal processing.

[0049] . The control system 100 comprises a heat transmission element 1 (Physical Product), configured to be in contact with the aforementioned material to be subjected to thermal processing, i.e. to be subjected to heating. Such heat transmission element 1 is, for example, made of metallic material or other heat-resistant materials. In one embodiment, such heat transmission element 1 takes the form of a welding blade.

[0050] . In a different embodiment, the heat transmission element 1 takes the form of a container element 1 suitable for containing the material to be subjected to thermal processing, in particular a box-shaped component of the system 100 closed on all sides.

[0051] . In a further embodiment, the container element 1 can be closed only on some sides, to contain the material to be processed, while still ensuring heat transfer to the material itself. For example, this container element 1 is a tunnel oven comprising a respective oven inlet and an open outlet to allow the introduction of the material to be processed and the unloading of the processed material, respectively.

[0052] . In the following description, reference will be made, by way of example but not limitation, to the heat transmission element 1 represented as the element containing the material to be heated.

[0053] . The system 100 comprises, for example, a temperature probe 2 (Temperature Probe), operatively associated with the container element 1 to detect, for each instant, a current temperature Tmat of the material to be heated contained in the container 1.

[0054] . In this case, the thermal regulation of the process consists in ensuring that the current temperature Tmat of the material to be heated is as close as possible to a desired temperature or target temperature Ttgt. Acceptable discrepancies in the values ​​of the aforementioned temperatures Tmat and Ttgt are part of the thermal process regulation parameters.

[0055] . The control system 100 further comprises one or more heating elements 3 (Physical Heater), for example of the resistive type, such as resistors based on metal alloys, infrared or ultraviolet lamps, silicon carbide elements, etc., suitable for heating the container element 1 containing the material to be heated.

[0056] . In particular, a thermal power WTres delivered by the heating element 3 depends on a first current Ires and a first voltage Vres applied to the heating element 3, according to a technical data sheet released by the manufacturer of the resistive element itself.

[0057] . The control system 100 also comprises a power electronic device 10 (power controller), for example in alternating current, suitable for transferring current to a load, in particular to the heating element 3. This power electronic device 10 includes a power electronic converter using thyristors or real SCR 101 and a suitable control for transforming an alternating current Isup and an alternating voltage Vsup of input power supply into an alternating current lout and an alternating output voltage Vout which are transferred to the resistive heating element 3, through a power transmission module 4 (Power Transmission). This current lout and this output voltage Vout are such that the thermal power WTres supplied by the resistive heating element 3 allows the current temperature value Tmat of the material to be brought as close as possible to the desired temperature value Ttgt.

[0058] . It should be noted that the aforementioned power transmission module 4 of the system 100 comprises electrical connection elements (e.g. cables or bars) suitable for connecting the power electronic device or power device 10 with the resistive load 3. These electrical connection elements may have specific physical design characteristics or be linked to malfunctions, so that the values ​​of the first current Ires and the first voltage Vres transmitted to the resistive load 3 are generally different from the values ​​of the current Iout and voltage Vout of the output of the power electronic device 10. In particular, for some applications, it is envisaged that the system 100 also comprises an electrical transformer connected between the power device 10 and the resistive load 3.

[0059] . Furthermore, the power electronic apparatus 10 is configured to receive as input: a duty-cycle parameter DTout representing a percentage of time during which the power apparatus 10 is capable of enabling the passage of the output current lout; control parameters PFout capable of describing how the aforementioned duty-cycle parameter is to be applied by the power apparatus 10.

[0060] . The control system 100 also comprises a process controller 11 (Process Controller) configured to manage the parameters of the thermal process as a function of the processing to be carried out. In particular, this process controller 11 is able to receive as input both the value of the current temperature Tmat of the material detected by the probe 2 and the desired temperature value Ttgt and is configured to modify the value of the duty-cycle parameter DTout, instant by instant, to be supplied to the electronic apparatus 10 to bring the detected temperature value Tmat closer to the desired temperature value Ttgt necessary for the processing phase in progress.

[0061] . In addition, the process controller 11 is configured to also transmit to the power electronics apparatus 10 the control parameters PFout which define how the duty-cycle parameter DTout is to be applied to the current wave.

[0062] . In one embodiment, the process controller 11 is a physical device, such as a programmable logic controller (PLC), which is independent (stand-alone) with respect to the power electronic apparatus 10 or can be integrated into such power electronic apparatus.

[0063] . In a different embodiment, the process controller 11 takes the form of a logic algorithm implemented on various devices, such as: operator panels or HMI (Human Machine Interface), industrial PCs, Edge Servers.

[0064] . The control system 100 further comprises a supervision unit 12 (supervision unit / cloud) configured to exchange process data D with the process controller 11. The exchange of such process data D allows the supervision unit 12 to verify that the process temperature parameters are respected, to record significant variables (e.g., energy consumption) and to detect any alarms or operating anomalies. The identification of such operating anomalies may determine automatic adjustments of the process control or may provide an operator with the information necessary for the same operator to perform a manual adjustment of the parameters of the process controller 11.

[0065] . In an embodiment, the aforementioned supervision unit 12 is integrated into the process controller 11, managed by a (local) operator panel, or it can be implemented on a remote Cloud platform capable of communicating with the process controller 11 via a telecommunications network (not shown).

[0066] . In a non-limiting embodiment, the control system 100 may also comprise one or more external physical sensors 13 configured to detect the values ​​of the first current Ires and the first voltage Vres transmitted to the resistive load 3. These sensors 13 allow for improved accuracy in thermal control by detecting the correct current and voltage values ​​supplied to the load 3 which may be different from those supplied by the power electronic apparatus 10 in the case of the presence of the power transmission module 4.

[0067] . The control system 100 implementing the present invention comprises, in addition to the real thyristor or SCR power electronic converter block 101, also a digital twin electronic converter functional block 102 or “digital twin” of the aforementioned real converter 101.

[0068] . Furthermore, as will be explained in more detail below, the control system 100 also comprises a functional block representing a feedback algorithm 103 configured to correct the value of the duty cycle parameter DTout fed to the SCR power electronic device 10 starting from an output of the digital twin electronic converter functional block 102 of the aforementioned real converter 101.

[0069] . Furthermore, the control system 100 also includes a functional block representative of a comparison algorithm 103' configured to generate status information related to a detected malfunction based on an analysis of an output of the digital twin electronic converter 102.

[0070] . In an embodiment, the digital twin electronic converter functional block 102 of the real converter 101 can be implemented in one of the aforementioned components of the thermal process control system 100: in the power electronics apparatus 10; in the process controller 11; in the supervision unit (local or cloud) 12. In particular, the digital twin block 102 is a software module loadable into a memory of one of the aforementioned components of the system 100.

[0071] . In an embodiment, the functional blocks representing the feedback algorithm 103 or the comparison algorithm 103' can be implemented in one of the aforementioned components of the thermal process control system 100: in the power electronics device 10; in the process controller 11; in the supervision unit (local or cloud) 12. In particular, the algorithms 103, 103' are software modules that can be loaded into a memory of one of the aforementioned components of the system 100.

[0072] . It should be noted that the thermal process of working a material controlled by the system 100 described above which implements the invention can be of a continuous or cyclic type.

[0073] . As is known, continuous thermal processes require maintaining the detected Tmat temperature value substantially stable for long periods of time (from hours up to months) and are characterised by a rather high and constant thermal inertia, for example of the order of approximately 1000 kJ / K. Generally, the control of continuous thermal processes is divided into three phases.

[0074] . In an initial heating phase, the container element 1 (Physical Product) is brought from ambient temperature to a process temperature, generally in the absence of material to be heated inside it. In the case of complex systems, the heating times are managed so that all the components of the system reach the process temperature simultaneously. In some cases, it may be necessary to provide heating procedures that involve a temperature increase with a ramp trend (steep increasing ramp) to manage transient phenomena in the heating phase such as, for example, the elimination of condensation inside the container element 1 to be heated.

[0075] . In a subsequent continuous thermal modulation phase, once the target temperature Ttgt has been reached to start the processing process, the material to be processed is introduced. The temperature regulation performed by the control system 100 therefore manages both the variation in thermal inertia linked to the presence of the material in the container element 1, and the temperature variations required by the various process phases. Furthermore, the temperature regulation also provides for the compensation of heat flows coming from other sources in the system (for example, generated by mechanical friction between materials and extruder screws). In general, the measured temperature of the material Tmat cannot fall below a threshold temperature value in this phase.

[0076] . In a subsequent shutdown phase, once the processing has finished, the container element 1 is brought back to room temperature. In this case, the temperature decrease ramp (decreasing ramp) is generally gradual, both to allow the cleaning of the container element 1 from material residues and to avoid damaging the system in the event of excessively sudden cooling.

[0077] . Examples of continuous thermal processes are present in the following applications: extrusion, flat glass processing, polymer production.

[0078] . Cyclic thermal processes are characterized by the relatively rapid repetition of production cycles, without modifying the process parameters. Cycle times are generally linked to the thermal inertia associated with the presence of the material in the container element 1 and can range from milliseconds to tens of seconds.

[0079] . For each cycle, the thermal processing is repeated through the following phases.

[0080] . A first phase involves heating the resistive element 3 up to the process temperature, which is generally fixed. This heating can also be of short duration, if the inertia of the heating element is low, for example of the order of approximately 1 kJ / K, with respect to the cycle times.

[0081] . Subsequently, the thermal process is carried out, during which the detected temperature Tmat is generally kept constant, while the power WTres supplied by the heating element 3 can be varied to compensate for the presence of the material to be worked in the container element 1.

[0082] . Subsequently, the heating element 3 is deactivated when the material to be processed has passed to the next stage of the process. Generally, this stage does not involve a return to room temperature and can be very short in the case of rapid processes.

[0083] . Examples of very rapid cyclic processes are related to packaging applications, where an electrically heated blade is used to seal the product packaging, often up to hundreds of times per minute. Other examples of cyclic processes involve processing steps used in the food or pharmaceutical industries.

[0084] . An example of the power electronic apparatus 10, for example in alternating current, suitable for transferring current to the heating element 3, through the power transmission module 4, is described with reference to figure 2.

[0085] . This power apparatus 10 comprises the above-mentioned electronic power converter 101 with semiconductors, generally “Silicon Controlled Rectifier” or SCR thyristors in phase opposition. This converter 101 is configured to enable or disable the current transmitted to the resistive load 3. In particular, the SCR converter 101 is configured to transform the current Isup and the input supply voltage Vsup into an output current Iout and voltage Vout supplied to the resistive load 3. The SCR converter 101 is suitable to be driven punctually by means of an input duty cycle signal DTin.

[0086] . The power electronic apparatus 10 also comprises current, voltage or power sensors, collectively indicated by the reference 104, suitable for detecting the electrical variables transmitted to the resistive load 3.

[0087] . In particular, when such sensors 104 are internal to the apparatus 10 and measure the electrical quantities output from the power electronic apparatus 10, the output current and voltage measured by the sensors 104 are Iout and Vout.

[0088] . When, however, such sensors 104 take the form of sensors external to the apparatus 10, for example they comprise current transformers or voltage shunts, applied to the resistive load 3, the measured current and voltage values ​​are respectively the first current Ires and the first voltage Vres, mentioned above.

[0089] . The power electronics apparatus 10 further comprises a control block 105, configured to receive the DTout and PFout parameters generated by the process controller 11 through one or more communication interfaces 106.

[0090] . The control block 105 is configured to generate, on the basis of the above parameters DTout and PFout, the above input duty cycle signal DTin which directly drives the SCR power converter 101, turning it on or off.

[0091] . For this purpose, in an embodiment, the control block 105 can also use measured current and voltage values, Vmes and Imes, detected by the internal or external electrical sensors 104, when available.

[0092] . Preferably, the power electronic apparatus 10 comprises a cooling block 107, comprising all the equipment, both passive and active, which ensures the air or water cooling of the power electronic apparatus 10.

[0093] . In an embodiment, the digital twin electronic converter functional block 102 or “digital twin” of the real converter 101, or the functional block representing the feedback algorithm 103 or comparison algorithm 103', or both, can be implemented in the control block 105 of the power electronic apparatus 10.

[0094] . Note that, depending on the complexity of the power electronic device 10, measured current, voltage and power values ​​and any alarms related to the electrical connections or the status of the device can be made available as output to the process controller 11 and the supervision unit 12. The alarms related to the electrical connections are normally related to the lack of input voltage to the device 10 or to the total or partial breakdown of the resistive load.

[0095] . In accordance with the present invention, an interaction between the real SCR or thyristor power electronic converter 101, the digital twin (SCR Digital Twin) of said real converter 102, and the feedback algorithm 103 configured to correct the input duty cycle signal DTin applied to the SCR power electronic converter 101, following an early detection of process malfunctions, from an output of the digital twin electronic converter 102 is described with reference to FIG. 3.

[0096] . In particular, the real SCR power electronic converter 101 is configured to receive as input, from the feedback algorithm 103, a value of the input duty cycle signal DTin relating to a current time instant t. This converter 101 is able to provide, as output, a peak value of the measured current Imes at that current time instant t, starting from that current duty cycle value.

[0097] . The digital twin electronic converter 102 is configured to receive, as input, in addition to the value of the peak current Imes measured by the physical SCR converter 101 at time t, control parameters PFout generated by the process controller 11 representing preset values ​​of frequency and supply voltage.

[0098] . The digital twin converter 102 is configured to calculate a plurality of sets of simulated temperature values ​​T1(t+dt, ..., t+f), T2(t+dt, ..., t+f), ..., Tm(t+dt, ..., t+f) at a plurality of specific points, e.g. m points, within the physical SCR converter 101 or the power apparatus 10. The simulated temperature value at each internal point is related to a time instant of a plurality of time instants t+dt, t+2dt, ..., t+f following the current instant t, within a prediction time interval having duration f, where such prediction time interval f is process-dependent. In particular, the time interval dt, e.g. measured in seconds, represents a predetermined sampling time interval. The aforementioned specific internal points of the actual converter 101 or the power apparatus 10 are those where malfunctions may occur, in particular due to overtemperature phenomena. Such specific points are, for example: the junction points of each SCR of the physical SCR converter 101; points of the power electronic apparatus 10 affected by dissipation phenomena; points of the control electronics of the power apparatus 10. In the following discussion, the term “specific internal points of the real power converter 101 of the power electronic apparatus 10” will be used to indicate the set of said specific internal points both of the real converter 101 and of the power electronic apparatus 10.

[0099] . Similarly, the digital twin converter 102 is configured to calculate a plurality of simulated values ​​of power P(t+dt, t+2dt, ..., t+f) dissipated by the physical SCR converter 101, each relating to a time instant t+dt, t+2dt, ..., t+f following the current time instant t within the prediction time interval having duration f.

[00100] . The digital twin electronic converter 102 is configured to make available such plurality of sets of simulated temperature values ​​Ti(t+dt, ..., t+f), with i = 1, 2, ..., m, and the simulated power values ​​P(t+dt, ..., t+f) both to the supervisory unit 12 of the system 100 and to the feedback algorithm 103.

[00101] . Note that the digital twin converter 102 of the invention is obtained from the physical SCR power converter 101. The starting point for the implementation of the digital twin 102 is a detailed electronic knowledge of the physical SCR 101, a knowledge of the possible failure modes of the physical device and the mechanical structure of the power electronic apparatus 10.

[00102] . An embodiment of the digital twin converter 102 that can be used in the present invention is described in the paper: F. Toso, R. Torchio, A. Favato, PG Carlet, S. Bolognani and P. Alotto, Digital Twins as Electric Motor SoftSensors in the Automotive Industry, 2021 IEEE International Workshop on Metrology for Automotive (MetroAutomotive), Bologna, Italy, 2021, pp. 13-18, doi: 10.1109 / MetroAutomotive50197.2021.9502885

[00103] . The proposed solution envisages using the potential of the digital twin converter 102 for real-time simulation or prediction of some temperatures of interest virtually measured at some specific points of the physical SCR 101, defined during the development phase of the digital twin 102. In other words, the digital twin 102 operates as if virtual temperature sensors were positioned at these points. The mathematical structure of the digital twin converter 102 is configured to relate the supply frequency and voltage, the measured peak current value and the estimated duty cycle at each of the time instants t+dt, t+2dt, ..., t+f following the current instant t, with the estimate of the power dissipated by the SCR at the same instants t+dt, t+2dt, ..., t+f. This estimate of dissipated power is used for the calculation of the temperatures Ti(t+dt, , t+f), with i= 1, 2, , m, measured by virtual sensors.

[00104] . In the case of real-time simulation, the digital twin converter 102 is configured to use a real temperature value measured at a point inside the power electronics apparatus 10, where a real sensor is positioned. Note that the digital twin converter 102 provides internal feedback, i.e., such converter is configured to detect differences between the aforementioned temperature value measured by the real sensor and the simulated value at the same point. Furthermore, the digital twin converter 102 is configured to use such detected differences to correct a mathematical structure of the digital twin converter 102.

[00105] . Still referring to Figure 3, the feedback algorithm 103 is configured to receive as input the plurality of sets of simulated temperature values ​​Ti(t+dt, ..., t+f), with i=1, 2, m, and the power values ​​P(t+dt, ..., t+f) generated by the digital twin converter 102 at the plurality of time instants of the prediction interval f.

[00106] . Furthermore, the feedback algorithm 103 is configured to receive as input target duty cycle values ​​Tdc(t+dt, t+2dt, ..., t+f) predicted by the thermal process at the same instants t+dt, t+2dt, ..., t+f of the prediction time interval fe made available by the process controller 11. Additionally, the feedback algorithm 103 is configured to receive additional process parameters as input.

[00107] . Note that the prediction time interval f has a duration that depends on the number of target duty cycle values ​​Tdc(t+dt, ..., t+f) made available by the process controller 11 to the feedback algorithm 103. In particular, the number of target duty cycle values ​​known to the process controller 11 defines how far into the future the process conditions can be maintained, taking into account the type of process being controlled and the phase being executed. For example, such a prediction time interval f is between a few seconds and a few hours.

[00108] . The feedback algorithm 103 is configured to evaluate whether the sets of simulated temperature values ​​Ti(t+dt, ..., t+f), with i=1, 2, ..., m, and the power values ​​P(t+dt, ..., t+f) provided by the digital twin electronic converter 102 are acceptable in relation to the process parameters.

[00109] . In case such sets of simulated temperature values ​​Ti(t+dt, ..., t+f), with i=1, 2, ..., m, and the power values ​​P(t+dt, ..., t+f) provided by the digital twin electronic converter 102 are not acceptable, in relation to the process parameters, the feedback algorithm 103 is able to calculate a plurality of new duty cycle values, Duty Cycle (t+dt,...t+f), each relating to a time instant t+dt, t+2dt, ..., t+f of the prediction time interval f considered on the basis of the process parameters.

[00110] . The feedback algorithm 103 is adapted to provide this plurality of new duty cycle values ​​both to the digital twin converter 102 and, optionally, to the physical SCR power converter 101 as a new value of the input duty cycle signal DTin relating to the aforementioned time instants t+dt, t+2dt, ..., t+f, where dt is the sampling time interval.

[00111] . Furthermore, the feedback algorithm 103 is adapted to provide status and alarm information to the supervisory unit 12 of the system 100.

[00112] . Note that a peculiar aspect of the feedback algorithm 103 consists in the evaluation of the acceptability of the values ​​predicted by the digital twin converter 102 and in the possible recalculation of the control parameters.

[00113] . To perform this step, the feedback algorithm uses information on limit values ​​provided by the process controller 11. These limit values ​​can be physical limits, such as maximum temperatures of the internal components of the SCR that ensure its correct operation, or they can be process-related specifications, such as maximum acceptable cycle times that ensure product quality.

[00114] . If the analysis of the sets of simulated temperature values ​​Ti(t+dt, ..., t+f), with i=1, 2, ..., m, and the power values ​​P(t+dt, ..., t+f) provided by the digital twin electronic converter 102 signals a possible exceeding of the aforementioned limits, the feedback algorithm 103 is configured to implement a strategy for recalculating the control values ​​(duty cycle). This feedback algorithm 103 is configured to implement various possible strategies within itself, such as slowing down the ignition ramp (if the anomaly occurs during the ignition phase), extending the process cycle time (during the working phase for acyclic processes), or controlled shutdown, if the algorithm 103 assesses that it is not possible to modify the control while maintaining the integrity of the process.

[00115] . In accordance with the present invention, an interaction between the real thyristor or SCR power electronic converter 101, the digital twin electronic converter 102 of said real converter, and the above-mentioned comparison algorithm 103' (Comparison algorithm), configured to generate status information related to a detected malfunction based on an analysis of an output of the digital twin electronic converter, is described with reference to FIG. 4.

[00116] . In particular, the real SCR power electronic converter 101 is configured to receive as input, from the process controller 11, a value of the input duty cycle signal DTin relating to a current time instant t. This converter 101 is capable of providing, as output, both to the digital twin electronic converter 102 of the aforementioned real converter and to the comparison algorithm 103' a peak value of the measured current Imes at that current time instant t, starting from that current duty cycle value.

[00117] . The digital twin electronic converter 102, entirely analogous to the one described in reference to figure 3, is configured to receive, at input, in addition to the value of the peak current Imes measured by the physical SCR converter 101 at time t, control parameters PFout generated by the process controller 11 representing preset values ​​of frequency and supply voltage.

[00118] . Furthermore, the digital twin converter 102 is configured to receive, as input, target duty cycle values ​​Tdc(t+dt, t+2dt, ..., t+f) predicted by the thermal process at instants t+dt, t+2dt, ..., t+f of the prediction time interval f and made available by the process controller 11. In particular, dt is a fixed sampling time interval.

[00119] . The digital twin converter 102 is configured to calculate a plurality of sets of simulated temperature values ​​Ti(t+dt, ..., t+f), with i=1, 2, ..., m, at a plurality of specific points within the physical SCR converter 101 or the power apparatus 10. Each temperature value of such plurality is related to a time instant t+dt, t+2dt, ..., t+f following the current time instant t, within a prediction time interval having duration f, where such prediction time interval f is process dependent.

[00120] . Similarly, the digital twin converter 102 is configured to calculate a plurality of simulated values ​​of power P(t+dt, t+2dt, ..., t+f) dissipated by the physical SCR converter 101, each relating to a time instant t+dt, t+2dt, ..., t+f following the current time instant t within the prediction time interval f.

[00121] . The digital twin electronic converter 102 is configured to make such sets of simulated temperature values ​​Ti(t+dt, ..., t+f), with i=1, 2, ..., m, and power values ​​P(t+dt, ..., t+f) available both to the supervisory unit 12 of the system 100 and to the comparison algorithm 103'.

[00122] . The comparison algorithm 103' is configured to receive as input the sets of simulated temperature values ​​Ti(t+dt, ..., t+f), with i=1, 2, ..., m, and the power values ​​P(t+dt, ..., t+f) generated by the digital twin converter 102 at the time instants t+dt, t+2dt, ..., t+f.

[00123] . Furthermore, the comparison algorithm 103' is configured to receive as input the target duty cycle values ​​Tdc(t+dt, ..., t+f) expected by the thermal process at the same instants t+dt, ..., t+f and made available by the process controller 11. The comparison algorithm 103' is configured to also receive the value of the peak current measured Imes by the physical SCR converter 101 at instant t.

[00124] . The comparison algorithm 103' is configured to compare the sets of simulated temperature values ​​Ti(t+dt, ..., t+f), with i=1, 2, ..., m, the power values ​​P(t+dt, ..., t+f) and the target duty cycle values ​​Tdc(t+dt, ..., t+f) predicted by the thermal process, with the measured physical values, i.e. the measured peak current value Imes, and to evaluate any anomalous conditions.

[00125] . Furthermore, the comparison algorithm 103' is designed to provide information on any detected anomalies and alarms to the supervision unit 12 of the system 100.

[00126] . The central aspect of the comparison algorithm 103' consists in the analysis of the trends of the temperature and power signals provided by the digital twin converter 102, and their correlation with any anomalies.

[00127] . In particular, the virtually measured temperature is compared with the temperature actually measured on the physical object 101 and any deviations between the two values ​​are evaluated. An anomaly or malfunction occurs when the behavior of the real temperature deviates from the simulated one. If this condition occurs, some relevant information is extracted from the temperature curve, such as any significant peaks, both positive and negative temperature peaks, and the temporal duration of such peaks.

[00128] . The comparison algorithm 103' is configured to compare this information with known constant values, stored in table form together with the algorithm, which relate the extent of the variation to the corresponding anomaly. These known values ​​are characterized previously, during the development phase of the power electronics product 10, or during the testing phase of the system or machinery during the setup of the system 100. In the first case, the values ​​relating to possible internal anomalies of the power electronics product 10 are characterized, while in the second case, the values ​​relating to possible anomalies due to malfunctions or disturbances external to the power electronics product 10 are characterized.In particular, the comparison algorithm 103' is configured to acquire from the digital twin 102 the time trend of the difference between the calculated and the actual measured temperature at a specific point inside the physical controller 101. This difference value deviates from zero when the behavior of the physical SCR converter 101 deviates from its nominal behavior, i.e., that described by the digital twin converter 102 in the absence of operating disturbances. The nature of the disturbance can therefore be deduced from the trend of the difference (delta) between the simulated and measured temperatures, together with the set of temperature setpoints Ti(t+dt, ..., t+f) and power values. P(t+dt, ..., t+f) and in the context of the process parameters provided by the process controller 11. For example, a negative temperature delta measured on the controller's electronic board, combined with a forecast power increase P and a request for a higher duty cycle by the power apparatus 10, may indicate an upcoming bonding failure of the physical SCR converter 101. In fact, the increase in resistance on the physical SCR converter 101 determines a higher power, but the local temperature increase is not matched by an equally rapid temperature increase on the electronic board and, at the same time, the lower current transferred to the heating load 3 leads the process controller 11 to request a higher duty cycle Tdc to maintain the temperature.

[00129] . Examples of anomalies detectable by the 103' comparison algorithm are:

[00130] . ​​- anomaly or malfunction of the cooling block 107 of the power electronic system 10;

[00131] . - fault or malfunction of the real SCR power converter 101;

[00132] . - fault or malfunction of the cooling system of the power electronic control panel 10.

[00133] . With reference to Figure 5, the reference numeral 200 indicates, as a whole, a general example of the control method of a power electronic apparatus 10, usable in a control system 100 of a thermal process for the processing of a material, to regulate the temperature or thermal state of a load 3 following an early detection of malfunctions in the processing by the digital twin electronic converter 102, according to the invention.

[00134] . The method in Figure 5 begins with a symbolic start phase “STR” and ends with a symbolic end phase “ED”.

[00135] . The control method 200 initially provides for a step 201 of early detection and classification of malfunctions in the thermal process of processing a material.

[00136] . As mentioned above, the control method 200 is applied to a power electronic apparatus 10 of the control system 100. The control system 100 comprises: - a heat transmission element 1 configured to be in contact with the material to be subjected to the thermal processing; - a heating load 3; - the aforementioned power electronic apparatus 10, suitable for transferring electric current Iout, Ires to the heating load 3 to modify the thermal state of the heat transmission element 1; the power electronic apparatus 10 includes a real power converter 101 driven by an input duty cycle signal DTin generated by the power electronic apparatus 10 itself to enable / disable the transfer of electric current to the heating load 3; - a process controller 11 configured to compare a detected current temperature value Tmat of the material subjected to the thermal processing with a reference temperature value Ttgt; this process controller 11 is configured to control the power electronics apparatus 10 by modifying the input duty cycle signal DTin applied to the real power converter 101 to bring the detected current temperature value Tmat closer to the reference temperature value Ttgt; - a functional block representing a digital twin converter 102 of the real power converter 101; - a functional block representing a comparison algorithm 103' operatively associated with the digital twin converter 102.

[00137] . In this initial phase, the method 200 comprises the advance detection and classification 201, by the comparison algorithm functional block 103', of at least one malfunction in the thermal process of processing the material.

[00138] . Furthermore, the method 200 includes a step of evaluating 202 a change in the control applied, by the process controller 11, to the power electronic apparatus 10 following the early detection of such at least one malfunction.

[00139] . Method 200 comprises the further step of:

[00140] . - apply 203 such modified control to the power electronic apparatus 10 to prevent the at least one malfunction detected in advance in the event that such modified control ensures the integrity of the machining process performed; or

[00141] . - to shut down 204, in a controlled manner, the control system 100 in the event that such modified control compromises the integrity of the manufacturing process being performed.

[00142] . With reference to the embodiment example of Figure 6, the general steps of the algorithm described with reference to Figure 5 are described in more detail to be applicable to the diagnosis of at least one malfunction or anomaly in the context of a thermal material processing process.

[00143] . In an exemplary embodiment, the above-mentioned step of detecting in advance and classifying 201 at least one malfunction in the thermal material processing process comprises the steps of:

[00144] . - start 2011 the digital twin converter function block 102 together with the control system 100;

[00145] . - simulating 2012, by the digital twin converter functional block 102, a plurality of sets of temperature values ​​Ti(t+dt, ..., t+f), with i=1, 2, ..., m, at a plurality of specific internal points of the real power converter 101 of the power electronic apparatus 10; each of the simulated temperature values ​​of each set is associated with a time instant of a plurality of time instants t+dt, t+2dt, ..., t+f of a prediction time interval f following a current time instant t of the manufacturing process, where dt is the above-mentioned predetermined sampling time interval;

[00146] . - compare 2012' each of the sets of simulated temperature values ​​Ti(t+dt, ..., t+f) at the specific points with a respective threshold temperature value TSi, with i=1, 2, ..., m, of a plurality of threshold values.

[00147] . In case at least one of the simulated temperature values ​​of one of the sets Ti (t+dt, ..., t+f) is greater than the respective threshold temperature value TSi, the method includes the steps of:

[00148] . - classify 2013, for each set of temperature values, by the comparison algorithm functional block 103', the at least one malfunction that caused the at least one simulated temperature value to be greater than the respective threshold temperature value TSi;

[00149] . - activate 2014 a feedback algorithm 103, operationally associated with the digital twin converter functional block 102, configured to evaluate whether said plurality of sets of simulated temperature values ​​Ti (t + dt, ..., t+f) provided by the digital twin converter 102 are acceptable in relation to the process parameters.

[00150] . In a further embodiment, with reference to critical trends in malfunction detection, the aforementioned step of early detection and classification 201 of at least one malfunction in the thermal material processing process of the method 200 comprises the steps of:

[00151] . - start 2011 the digital twin converter function block 102 together with the control system 100;

[00152] . - simulating 2012, by the digital twin converter functional block 102, a plurality of sets of temperature values ​​Ti(t+dt, ..., t+f), with i=1, 2, ..., m, at a plurality of specific internal points of the real power converter 101 of the power electronic apparatus 10; each of the simulated temperature values ​​of each set is associated with a time instant of a plurality of time instants t+dt, t+2dt, ..., t+f of a prediction time interval f following a current time instant t of the manufacturing process, where dt is the above-mentioned predetermined sampling time interval;

[00153] . - calculate 2012a, for each specific point i, with i=1, 2, ..., m, internal to the real power converter 101, values ​​of average temperature variation VTi(t+dt, ..., t+f), VTi(t+2dt, ..., t+f),.., VTi(t+f-dt, ..., t+f); each of these values ​​of average temperature variation VTi at the specific point i is calculated in a number of time instants equal to or less than the plurality of time instants t+dt, t+2dt, ., t+f of a prediction time interval f following a current time instant t of the process; each of these values ​​of mean temperature variation VTi at point i is relative to a plurality of temperature values ​​Ti(t+dt,., t + f), Ti(t+2dt,..., t+f),., Ti(t+f-dt,., t+f) simulated at that point;

[00154] . - compare 2012a', each of the values ​​of mean temperature variation VTi (t+dt, ., t+f), VTi(t+2dt, ., t+f),., VTi(t+f-dt, ., t+f) associated with the internal point i of the converter, with a threshold mean temperature variation value VT*i specific for that point ie determined on a process basis.

[00155] . In the case in which at least one of the values ​​of average temperature variation VTi(t+dt, ., t+f), VTi(t+2dt, ., t+f),., VTi(t+f-dt, ., t+f) is greater than the respective value of average threshold temperature variation VT*i, the method includes the steps of:

[00156] . - classifying 2013a, by the comparison algorithm functional block 103', the at least one malfunction that has caused at least one of the values ​​of average temperature variation VTi(t+dt, ., t+f), VTi(t+2dt, ., t+f),., VTi(t+f-dt, ., t+f) to be greater than the respective value of average threshold temperature variation VT*i;

[00157] . - activate 2014 a feedback algorithm 103, 4 operationally associated with the digital twin converter 102, configured to evaluate whether said plurality of sets of simulated temperature values ​​Ti(t+dt, ..., t+f) provided by the digital twin converter 102 are acceptable in relation to the process parameters.

[00158] . Still with reference to critical trends in the detection of malfunctions, in a further alternative embodiment to the previous one, the phase of detecting in advance and classifying 201 at least one malfunction in the thermal process of material processing, the method 200 comprises the phases of:

[00159] . - start 2011 the digital twin converter function block 102 together with the control system 100;

[00160] . - simulating 2012, by the digital twin converter functional block 102, a plurality of sets of temperature values ​​Ti(t+dt, ..., t+f), with i=1, 2, ..., m, at a plurality of specific internal points of the real power converter 101 of the power electronic apparatus 10; each of the simulated temperature values ​​of each set is associated with a time instant of a plurality of time instants t+dt, t+2dt, ..., t+f of a prediction time interval f following a current time instant t of the manufacturing process, where dt is the above-mentioned predetermined sampling time interval;

[00161] . - associate 2012b, to each point inside the converter i, a set of binary parameters representing the presence / absence of oscillatory trends AOi(t+dt,...,t+f), AOi(t+2dt,..,t+f), ..., AOi(t+(f-dt),...,t+f) of temperature; each of these parameters is related to a sub-interval of temperatures Ti(t+dt,...,t+f), Ti(t+2dt,...,t+f),..., Ti(t+(fdt),...,t+f) of these simulated temperature values ​​Ti(t+dt, ..., t+f);

[00162] . - detect 2012b', in at least one of said temperature sub-intervals, the presence of at least one oscillatory trend AOi(t+dt,...,t+f), AOi(t+2dt,...,t+f), ..., AOi(t+(fdt),...,t+f), relating to the point inside the converter, not compliant with the process when said sub-interval includes at least one first local minimum temperature value, one first local maximum temperature value and a further second consecutive local minimum temperature value or at least one first local maximum temperature value, one first local minimum temperature value and a further second consecutive local maximum temperature value;

[00163] . - classify 2013b, by the comparison algorithm functional block 103', the at least one malfunction that has determined the presence of such at least one oscillatory trend AOi(t+dt,.,t+f), AOi(t+2dt,.,t+f), ..., AOi(t+(fdt),...,t+f) of temperature;

[00164] . - activate 2014 a feedback algorithm 103, operatively associated with the digital twin converter 102, configured to evaluate whether said plurality of sets of simulated temperature values ​​Ti(t+dt, ..., t+f) provided by the digital twin converter 102 are acceptable in relation to the process parameters.

[00165] . Note that the aforementioned first local temperature maximum is present in the temperature sub-range when three reference temperature values ​​of the first maximum are detected, in particular a first value T0_max of the first maximum, a second value T1_max of the first maximum and a third temperature value T2_max of the first maximum such that: T0_max <= T1_max - Trefi and T1_max >= T2_max + Trefi, where Trefi is a tolerance temperature value associated with the i-th point inside the converter, dependent on the process and defined by the process parameters.

[00166] . Similarly, the aforementioned second local temperature maximum is present in the temperature sub-range when three reference temperature values ​​of the second maximum are detected, in particular a first value T0'_max of the second maximum, a second value T1'_max of the second maximum and a third temperature value T2'_max of the second maximum such that: T0'_max <= T1'_max - Trefi and T1'_max >= T2'_max + Trefi.

[00167] . Similarly, the first local temperature minimum is present in the temperature sub-range when three further reference temperature values ​​of the first minimum are detected, a further first temperature value T0_min of the first minimum, a further second value T1_min of the first minimum and a further third value T2_min of the first minimum such that: T0_min >= T1_min + Trefi and T1_min <= T2_min-Trefi.

[00168] . Similarly, the aforementioned second local minimum temperature is present in the temperature sub-range when three further reference temperature values ​​of the second minimum are detected, a further first T0'_min value of the second minimum, a further second T1'_min value of the second minimum and a further third T2'_min temperature value of the second minimum such that: T0'_min >= T1'_min + Trefi and T1'_min <= T2'_min-Trefi.

[00169] . In a first embodiment, a methodology for detecting the aforementioned temperature oscillatory trend operates in the following way.

[00170] . The method involves associating the first temperature value of the set Ti(t+dt) with an initial reference temperature value T0.

[00171] . Each temperature value Ta following the first reference temperature value T0 is compared with that temperature value T0 by testing two first conditions A) and B), which are alternative to each other and do not occur simultaneously, i.e.: A) Ta>= T0+Trefi B) Ta<= T0-Trefi.

[00172] . The method involves testing these first conditions iteratively until either at least one of the two conditions is verified or all possible temperature values ​​in the set have been tested, but neither of the first two conditions has ever occurred. In this second case, the method ends with a negative result, i.e. no oscillatory trend has been identified in the sub-range of temperatures examined.

[00173] . If condition A) is verified, the method of the invention provides for assigning the initial reference temperature value T0 to the first reference temperature value of the first maximum T0_max. The method of the invention therefore provides for activating a search phase for a local maximum based on the following phases: - associate the temperature value Ta that verified condition A) with the second reference temperature value of the first maximum T1_max; - compare each temperature value Tb following the second reference temperature value of the first maximum T1_max with said second reference temperature value by testing two second conditions, A1) and A2), which are alternatives to each other and do not occur simultaneously: A1) Tb<=T1_max-Trefi A2) Tb>=T1_max+Trefi.

[00174] . These second conditions A1) and A2) are tested iteratively until either at least one of the two is verified or all possible temperature values ​​of the set have been tested, but neither of the two has ever occurred. In this second case, the search for a local maximum has a negative outcome.

[00175] . If condition A1) is verified, the method associates the temperature value Tb, which satisfies condition A1), with the third reference temperature value of the first maximum T2_max. In this case, the set of three temperature values ​​(T0_max, T1_max, T2_max) identified defines the aforementioned first local maximum. The method for finding the local maximum therefore concludes with a positive result.

[00176] . Instead, if condition A2) is verified, the method provides for reassigning the reference temperature values ​​so that: - the second reference temperature value of the first maximum T1_max is used as the first reference temperature value of the first maximum T0_max, and - the temperature value Tb that verified condition A2) is used as the second reference temperature value of the first maximum T1_max.

[00177] . The method for finding the local maximum involves subsequently repeating the four steps mentioned above in an iterative manner.

[00178] . If the method for finding a local maximum has been successful and the set of values ​​(T0_max, T1_max, T2_max) that defines the local maximum has been identified, the method involves finding a first local minimum according to the following steps.

[00179] . It is expected to associate the second reference temperature value of the first maximum T1_max of the triplet of the first local maximum to the further first temperature value T0_min of the triplet of the first local minimum sought.

[00180] . Subsequently, the method involves associating the third temperature value of the triplet of the first local maximum T2_max with the further second reference temperature value of the first minimum T1_min.

[00181] . In this case, each temperature value Td following the aforementioned second temperature value of the first minimum T1_min is compared with T1_min by testing two further third conditions A1') and A2'), which are alternatives to each other and are not verified simultaneously: A1') Td>=T1_min+Trefi A2') Td<=T1_min-Trefi.

[00182] . These third conditions A1') and A2') are tested iteratively until either at least one of the two is verified or all possible temperature values ​​of the set have been tested, but neither of the two has ever occurred. In this second case, the search for a local minimum fails.

[00183] . If condition A1') is verified, the method requires that the temperature value Td being verified be associated with the further third reference temperature value of the first minimum T2_min, such that the triplet (T0_min, T1_min, T2_min) defines the first local minimum identified. The local minimum search method therefore concludes with a positive outcome.

[00184] . Instead, if condition A2') is verified, the method provides for reassigning the reference temperature values ​​so that: - the further second T1_min reference temperature value of the first minimum is associated with the further first reference temperature value of the first minimum T0_min, and - the temperature value Td that occurred (A2') is used as the further second reference temperature value of the first minimum T1_min.

[00185] . The method for finding the local minimum involves subsequently repeating the four steps mentioned above in an iterative manner.

[00186] . If the local minimum search method is successful and the set of three (T0_min, T1_min, T2_min) that defines the local minimum has been identified, then the method provides for the search for a second local maximum according to the following steps.

[00187] . It is expected to associate the second reference temperature value of the first minimum T1_min of the first local minimum set of three to the first temperature value T0'_max of the second local maximum set of three sought.

[00188] . Subsequently, the method involves associating the third temperature value of the triplet of the first local minimum T2_min to the second reference temperature value of the second maximum T1'_max.

[00189] . The method therefore involves activating the search method for a second local maximum in a manner similar to that described previously for the first local maximum.

[00190] . If the local maximum search method is successful and the triplet (T0'_max, T1'_max, T2'_max) that defines the second local maximum has been identified, then the method concludes with a positive outcome and an oscillatory trend has been identified.

[00191] . In the case in which condition A) is not verified but condition B) of the first conditions A) and B) is verified, the method provides for assigning the initial reference temperature value T0 to the further first reference temperature value T0_min of the first local minimum set of three sought.

[00192] . The temperature value Ta which verified condition B) is used as a further second reference temperature value of the first minimum T1_min.

[00193] . The method for finding a first local minimum works as described previously.

[00194] . If the local minimum search methodology has been successful and the further set of three (T0_min, T1_min, T2_min) that defines the first local minimum has been identified, then it is expected to: - associate the further second reference temperature value of the first minimum T1_min of the set of three local minimums to the first value T0_max of the set of three local maximums sought; - associate the further third reference temperature value of the first minimum T2_min of the first local minimum triplet to the second reference temperature value of the first maximum T1_max.

[00195] . The method therefore involves activating the method for finding a first local maximum as described previously.

[00196] . If the local maximum search method was successful and the triplet (T0_max, T1_max, T2_max) that defines the first local maximum was identified, then it is expected to: - associate the second reference temperature value of the first maximum T1_max of the set of three local maximums to the further first value T0'_min of the set of three local minimums sought; - associate the third reference temperature value of the first maximum T2_max of the triplet of the first local maximum to the further second reference temperature value of the second minimum T1'_min.

[00197] . The method therefore involves activating the method for finding a second local minimum in a manner similar to that described previously.

[00198] . If the local minimum search method has been successful and the further triplet (T0'_min, T1'_min, T2'_min) that defines the second local minimum has been identified, then the method concludes with a positive outcome as an oscillatory trend has been identified.

[00199] . In an exemplary embodiment of the control method 200 of Figures 5 or 6, the aforementioned shutdown step 204 comprises a step of performing a controlled shutdown of the control system 100, by the process controller 11, after a shutdown time interval s(n) from said current time instant t, wherein the shutdown time interval s(n) has a duration shorter than the aforementioned prediction time interval f.

[00200] . In particular, this turn-off time interval s(n) is calculated by the expression: s(n) = f-(n*dt) (1) where f is the prediction time interval, dt is the above-mentioned sampling time interval and n is an integer.

[00201] . Note that s(n) represents a controlled shutdown time interval. In particular, the controlled shutdown time interval s(n) is a time datum, whose value is an integer multiple of the sampling time interval dt, which determines the granularity with which the controlled shutdown analysis takes place. This sampling time dt is given by the process parameters and depends on how fast the controlled process is. In other words, based on equation (1), the controlled shutdown time interval s(n) takes on a value that decreases at each iteration. The last value that the time interval s(n) takes on is determined by the prediction time interval f and the last value of the integer n such that (n*dt)>f.

[00202] . Still referring to the example of figure 5, in a further embodiment of the control method 200, the steps of evaluating 202 a modification of the applied control and applying 203 the modified control to the power electronic apparatus 10 to prevent the at least one malfunction detected in advance comprise the steps of:

[00203] . - execute a first control optimization algorithm 2031 applied to the power electronic apparatus 10;

[00204] . - evaluate 2032 whether the at least one early detected malfunction is prevented based on the execution of such first control optimization algorithm 2031.

[00205] . In the event that the at least one malfunction detected in advance persists, the method 200 comprises the further steps of:

[00206] . - execute a second control optimization algorithm 2033 applied to process controller 11;

[00207] . - evaluate 2034 whether the at least one pre-detected malfunction is prevented based on the execution of such second control optimization algorithm 2033.

[00208] . In the event that the at least one malfunction detected in advance persists, the method comprises the controlled shutdown step 204 of the control system 100.

[00209] . In a particular embodiment of the control method 200, such a step of executing the first control optimization algorithm 2031 comprises a step of iteratively executing such first control optimization algorithm 2031.

[00210] . In a particular embodiment of the control method 200, such a step of executing the second control optimization algorithm 2033 comprises a step of iteratively executing such second control optimization algorithm 2031.

[00211] . With reference to the embodiment example of Figure 7, the first control optimization algorithm 2031 comprises a step 2031a of evaluating whether a reduction in the input duty cycle signal DTin applied to the real power electronic converter 101 maintains the integrity of the machining process being performed.

[00212] . In the case where the integrity of the performed machining process is maintained, the method 200 comprises the steps of:

[00213] . - calculating 2031', by the feedback algorithm 103, a first plurality of duty cycle values ​​DC1(t+dt, ..., t+f) each associated with a time instant t+dt, ..., t+f of the plurality of instants in the prediction time interval f, following a current time instant t of the machining process, where dt is the predetermined sampling time interval;

[00214] . - make the first plurality of duty cycle values ​​D-C1(t+dt, ..., t+f) available to the digital twin converter function block 102;

[00215] . - simulating 2031'', by the digital twin converter functional block 102, a plurality of sets of first temperature values ​​T1i(t+dt, ..., t+f), with i=1, 2, ..., m, at a plurality of specific internal points of the real power converter 101 of the power electronic apparatus 10; each of such sets of simulated first temperature values ​​being associated with such first plurality of duty cycle values ​​D-C1(t+dt, ..., t+f) at a time instant t+dt, ..., t+f;

[00216] . - compare 2031b each such set of simulated first temperature values ​​T1i(t+dt, ..., t+f) with a respective threshold temperature value TSi, with i=1, 2, ..., m, of a plurality of threshold values.

[00217] . In case each of the first temperature values ​​T1i(t+dt, ..., t+f) simulated at each specific point is lower than the respective threshold value TSi, i.e. no anomaly has been identified, the method comprises the steps of:

[00218] . - apply 2031c to the real power converter 101 an input duty cycle signal DTin including such first plurality of duty cycle values ​​D-C1(t+dt, ..., t+f) to enable / disable the transfer of electric current to the heating load 3;

[00219] . - signaling, by the feedback algorithm functional block 103, an error condition to a supervisory unit of the control system 11.

[00220] . Instead, in the case in which at least one of the first temperature values ​​T1i(t+dt, ..., t+f) simulated at each specific point is greater than the respective threshold value TSi, the step 2031 of the method 200 provides for the repetition of such step of evaluating 2031a whether a further reduction of the input duty cycle signal DTin applied to the real power electronic converter 101 maintains the integrity of the performed machining process, and the step of calculating 2031', by the feedback algorithm 103, a further first plurality of duty cycle values ​​DC1'(t+dt, ..., t+f) each associated with a time instant t+dt, ..., t+f of the plurality of instants in the prediction time interval f.

[00221] . At this point, the repetition of the same phases 2031'' and 2031b mentioned above is expected.

[00222] . In other words, in the presence of an overtemperature, an optimization strategy implemented by algorithm 2031 provides for reducing the duty cycle by an amount related to the type of thermal process and made available by the process controller 11.

[00223] . For example, dc_x indicates the duty cycle value that allows the controlled process to maintain the desired target temperature and dc_y indicates the duty cycle value provided by the process controller 11. Note that dc_y could be different from dc_x because it represents the duty cycle value optimized by the process controller 11 that allows obtaining the desired regulation performance in terms of robustness / speed. In case the digital twin converter 102 signals an overtemperature for a duty cycle equal to dc_y, the first optimization algorithm 2031 is configured to perform a progressive reduction of the duty cycle from the limit value dc_y up to the value dc_x. The variation of this duty cycle can be performed in a maximum number of steps indicated by an integer parameter N. This integer parameter N is, for example, fixed to a reference value of 10. The value of the parameter N is provided by the process parameters and can also be changed arbitrarily by the user. In consideration of this, with the algorithm 2031, the duty cycle is varied at each step by an amount equal to: (dc_y-dc_x) / N (2) Note that the dc_x value is also known because it depends on the type of machining process and can be changed by the user.

[00224] . In a further embodiment, in the case of critical trends, the first control optimization algorithm 2031 provides for replacing the phase 2031b so as to include the phases of calculating 2012a values ​​of average temperature variation VTi(t+dt, ..., t+f), VTi(t+2dt, ..., t+f),..., VTi(t+f-dt, ..., t+f) and comparing 2012a' each of the values ​​of average temperature variation VTi(t+dt, ..., t+f), VTi(t+2dt, ..., t+f),., VTi(t+f-dt, ., t+f) with a threshold average temperature variation value VT*i, in a manner similar to that described with reference to figure 6.

[00225] . In a different embodiment, in the case of critical trends, the first control optimization algorithm 2031 provides for replacing the phase 2031b so as to include the phases of associating 2012b, to each point inside the converter i, a set of binary parameters representing the presence / absence of oscillatory trends AOi(t+dt,...,t+f), AOi(t+2dt,..,t+f), ..., AOi(t+(f-dt),...,t+f) of temperature and of detecting 2012b' the presence of at least one oscillatory trend AOi(t+dt,.,t+f), AOi(t+2dt,.,t+f), ..., AOi(t+(f-dt),...,t+f), relating to the point inside the converter, not compliant with the process in a manner similar to that described with reference to figure 6.

[00226] . With reference to the example of implementation of the method 200 of figure 7, the second control optimization algorithm 2033 includes an evaluation phase 2033a relating to the extension of the cycle time of the system 100 which preserves the integrity of the manufacturing process. More generally, the algorithm 2033 includes the evaluation of control strategies which alter the progressive behavior of the system and which, therefore, require the intervention of the process controller 11.

[00227] . In the case where the integrity of the performed machining process is maintained, the method 200 comprises the steps of:

[00228] . - calculating 2033', by the feedback algorithm 103, a second plurality of duty cycle values ​​DC2(t+dt, ..., t+f) each associated with a time instant t+dt, ..., t+f of such plurality of time instants in the prediction time interval f, following a current time instant t of the machining process, where dt is the predetermined sampling time interval;

[00229] . - make this second plurality of duty cycle values ​​D-C2(t+dt, ..., t+f) available to the digital twin converter functional block 102;

[00230] . - simulating 2033'', by the digital twin converter functional block 102, a plurality of sets of second temperature values ​​T2i(t+dt, ..., t+f), with i=1, 2,..., m, at a plurality of specific internal points of the real power converter 101 of the power electronic apparatus 10, each of such sets of simulated second temperature values ​​being associated with such second plurality of duty cycle values ​​D-C2(t+dt, ..., t+f) at a time instant t+dt, ..., t+f of such prediction time interval f;

[00231] . - compare 2033b each of said sets of simulated second temperature values ​​T2i(t+dt, ..., t+f) with a respective threshold temperature value TSi, with i=1, 2, ..., m, of a plurality of threshold values.

[00232] . In case each of the second temperature values ​​T2i(t+dt, ..., t+f) simulated at each specific point of the converter 101 is lower than the respective threshold value TSi, the method includes the steps of:

[00233] . - apply 2033c to the real power converter 101 an input duty cycle signal DTin including such second plurality of duty cycle values ​​D-C2(t+dt, ..., t+f) to enable / disable the transfer of electric current to the heating load 3;

[00234] . - reporting, by the feedback algorithm functional block 103, an error condition to a supervisory unit of the control system 11.

[00235] . Instead, in the case in which at least one of the second temperature values ​​T2i(t+dt, ..., t+f) simulated at each specific point of the converter 101 is greater than the respective threshold value TSi, the step 2033 of the method 200 provides for the repetition of said evaluation step 2033a of the cycle time of the system 100 which preserves the integrity of the manufacturing process and the step of calculating 2033', by the feedback algorithm 103, a further second plurality of duty cycle values ​​D-C2'(t+dt, ..., t+f) each associated with a time instant t+dt, ..., t+f of the plurality of instants in the prediction time interval f.

[00236] . At this point, the repetition of the same phases 2033'' and 2033b mentioned above is expected.

[00237] . The aforementioned extension 2033a of the cycle time of the system 100 is realized in an incremental extension in time of a switch-on ramp of the system 100, i.e. the second algorithm 2033 provides for extending the time duration of a switch-on ramp by an amount relative to the type of thermal process and made available by the process controller 11. In this case, for example, the procedure is similar to that used for stepping the duty cycle, where instead of two limit duty cycle values, we have two limit ramp time values. In this case, Ty is used to indicate the maximum allowable time for the process, which depends on the manufacturing process (the ramp cannot be excessively long, otherwise it will negatively impact the process). Ty is a known value, provided by the process parameters and can be set by the user. Tx, on the other hand, indicates the optimal ramp time value, which will be less than Ty. In this case, until the digital twin converter 102 signals an overtemperature, the ramp time is increased by a value equal to: (Ty-Tx) / N (3) up to a maximum value of Ty. Similarly to the above, the integer parameter N is, for example, fixed to a reference value of 10. The value of the parameter N is provided by the process parameters and can also be changed arbitrarily by the user.

[00238] . In a further embodiment, in the case of critical trends, the second control optimization algorithm 2033 provides for replacing the step 2033b so as to include the steps of calculating 2012a the mean temperature variation values ​​VTi(t+dt, ..., t+f), VTi(t+2dt, ..., t+f),..., VTi(t+f-dt, ..., t+f) and comparing 2012a' each of the mean temperature variation values ​​VTi(t+dt, ., t+f), VTi(t+2dt, ., t+f),., VTi(t+f-dt, ., t+f) with a threshold mean temperature variation value VT*i, in a manner similar to that described with reference to figure 6.

[00239] . In a different embodiment, in the case of critical trends, the second control optimization algorithm 2033 provides for replacing the phase 2033b in order to include the phases 2012b to associate, at each point inside the converter i, a set of binary parameters representing the presence / absence of oscillatory trends AOi(t+dt,...,t+f), AOi(t+2dt,..,t+f), ..., AOi(t+(f-dt),...,t+f) of temperature and to detect 2012b' the presence of at least one oscillatory trend AOi(t+dt,..,t+f), AOi(t+2dt,..,t+f), ..., AOi(t+(f-dt),...,t+f), relating to the point inside the converter, not compliant with the process in a similar way to what is described in reference to figure 6.

[00240] . Still referring to the embodiment example of figure 7, the aforementioned step of switching off 204 the control system 100 of the method 200 comprises a step of evaluating 2041 whether it is possible to maintain the integrity of the manufacturing process performed for a controlled shutdown time interval of limited duration s(n) following a current time instant t of the manufacturing process, and of anticipating the shutdown of the system 100 to time t+s(n). In particular, the control method 200 provides the possibility of obtaining the shutdown time value s(n) by decreasing a prediction time interval f on the basis of the equation: s(n) = f-(n*dt), with n= 1, 2, 3, etc. Based on this notation, the turn-off time s(n) can take on the discrete values ​​f-dt, f-2dt, f-3dt, etc.

[00241] . In the event that the process specifications allow the integrity of the manufacturing process to be maintained for a limited time, the method 200 comprises a step of repeating, for each of the discrete values ​​of such controlled shutdown time interval s(n) determined by the values ​​of the integer n for which (n*dt) <f, le fasi di:

[00242] . - calculate 2042, by the feedback algorithm 103, a third plurality of duty cycle values ​​DC3(t+dt, ..., t+s(n)) each associated with a time instant t+dt, ..., t+s(n) of the plurality of time instants of the controlled switch-off time interval s(n);

[00243] . - make the third plurality of duty cycle values ​​D-C3(t+dt, ..., t+s(n)) available to the digital twin converter function block 102;

[00244] . - simulating 2043, by the digital twin converter functional block 102, a plurality of sets of third temperature values ​​T3i(t+dt, ..., t+s(n)), with i=1, 2, ..., m, at a plurality of specific internal points of the real power converter 101 of the power electronic apparatus 10; each of such sets of simulated third temperature values ​​is associated with such third plurality of duty cycle values ​​DC3(t+dt, ..., t+s(n)) at a time instant t+dt, ..., t+s(n) of the controlled turn-off time interval s(n);

[00245] . - compare 2044 each of such sets of simulated third temperature values ​​T3i(t+dt, ..., t+s(n)) with respective sets of threshold temperature values ​​TSi, with i=1, 2, .., m.

[00246] . This repeat step is performed in case at least one of the third temperature values ​​T3i(t+dt, ..., t+s(n) simulated at each specific point of the real converter 101 is greater than the respective threshold value TSi.

[00247] . In an embodiment, in the case where each of the third temperature values ​​T3i(t+dt, ..., t+s(n) simulated at a specific point of the converter 101 is less than the respective threshold value TSi, the step of switching off 204 the control system 100 comprises, the steps of:

[00248] . - apply 2045 to the real power converter 101 an input duty cycle signal DTin which includes such third plurality of duty cycle values ​​D-C3(t+dt, ..., t+s(n)) to enable / disable the transfer of electric current to the heating load 3;

[00249] . - signaling, by the feedback algorithm functional block 103, an error condition to a supervisory unit 12 of the control system 100.

[00250] . In other words, assuming the prediction time interval f equal to 8, dt=1 and n initially equal to 0, based on equation (1) s(n)=8, i.e. the first value of s(n) is equal to f. If the temperature predicted by the digital twin converter 102 assumes values ​​higher than the threshold value TS at time t+8, the shutdown step 204 of the method provides for increasing the value of n from 0 to 1, setting the value of the controlled shutdown time interval s(n)=7. At this point, step 204 of the method provides for the recalculation of the Duty Cycle value (step 2042) which is made available to the digital twin converter 102 (step 2043). If the digital twin converter 102 reports a simulated temperature even higher than the threshold temperature TS at time t+7, the method involves incrementing the value of n from 1 to 2, setting the value of the controlled turn-off time interval s(n)=6. At this point, step 204 of the method involves recalculating the Duty Cycle value which is made available again to the digital twin converter 102. If the digital twin converter 102 does not show overtemperatures at time t+6, the duty cycle values ​​calculated up to time t+6 are made available to the power electronics device 10, i.e. the duty cycle values ​​calculated from t to t+s(n). Therefore, in relation to the example provided, the duty cycle values ​​calculated at time instants t+6 to t+8 are set to 0.

[00251] . In a further embodiment, in the case of critical trends, the aforementioned switching off phase 204 of the control system 100 provides for replacing the phase 2044 so as to include the phases of calculating 2012a values ​​of average temperature variation VTi(t+dt, ..., t+f), VTi(t+2dt, ..., t+f),..., VTi(t+f-dt, ..., t+f) and comparing 2012a' each of the values ​​of average temperature variation VTi(t+dt, ., t+f), VTi(t+2dt, ., t+f),., VTi(t+f-dt, ., t+f) with a threshold average temperature variation value VT*i, in a manner similar to that described with reference to figure 6.

[00252] . In a different embodiment, in the case of critical trends, the aforementioned phase of switching off 204 the control system 100 provides for replacing the phase 2044 so as to include the phases of associating 2012b, to each point inside the converter i, a set of binary parameters representing the presence / absence of oscillatory trends AOi(t+dt,...,t+f), AOi(t+2dt,..,t+f), ..., AOi(t+(f-dt),...,t+f) of temperature and of detecting 2012b' the presence of at least one oscillatory trend AOi(t+dt,..,t+f), AOi(t+2dt,..,t+f), ..., AOi(t+(f-dt),...,t+f ), relating to the point inside the converter, not compliant with the process in a manner similar to that described with reference to figure 6.

[00253] . In a further embodiment of the method 200 of the invention, the step 204 of shutting down the control system 100 further comprises a step 2046 of initiating and signaling by the process controller 11 of a scheduled shutdown of the control system 100 after such controlled shutdown time interval s(n) from the current time instant t.

[00254] . In a further embodiment of the method 200 of the invention, the step 204 of shutting down the control system 100 further comprises a step 2047 of initiating and signaling by the process controller 11 of an emergency shutdown of the control system 100 if, based on the process specifications, at the current time instant t it is not possible to maintain the integrity of the process for a controlled shutdown time interval s(n).

[00255] . Note that the first control optimization algorithm 2031 is performed at the power electronics level 10 and it is not necessary to modify the control of the rest of the system 100.

[00256] . The optimization of the control performed by the second control algorithm 2033 requires, instead, a modification of the overall behavior of the system 100, which must be implemented at the level of the process controller 11 implemented as a PLC.

[00257] . Furthermore, the processing performed by the method steps 201, 203, 204 can be performed both in the case of continuous processes and in the case of cyclic processes, adapting the control strategies to the specificity of the application.

[00258] . Some examples of optimization strategies implemented at the power electronics level 10 are:

[00259] . - rounding of ignition profiles (e.g. chamfering sawtooth profiles”);

[00260] . - redistribution of power to adjacent zones, when controlled by the same power apparatus.

[00261] . These strategies can be implemented autonomously by the power device 10, which is however capable of signaling their implementation to the process controller 11. The device 10 is, in fact, generally equipped with a feedback loop that allows for the detection and correction of process anomalies.

[00262] . It is also possible that the change in control at the power device level 10 could be interpreted as an anomaly. In this sense, several approaches are possible: - the power apparatus 10 signals the application of a process change to the PLC controller 11, which then does not activate the correction of the anomaly; - optimization at the power unit level 10 is implemented in a way that is transparent to the PLC controller 11 (e.g., sawtooth smoothing); - optimization at the power unit level 10 is designed to work in coordination with the anomaly correction at the PLC controller level 11 (e.g., over-reduction of the duty cycle of all heating zones to “pre-compensate” for the subsequent increase by the PLC).

[00263] . These approaches may be provided in the logic of the power apparatus 10 and enabled depending on the application and operating conditions.

[00264] . Some examples of process optimization strategies at the control system or machine level 100, particularly at the process controller level 11, are: - slowing down of the process controller 11 power-on ramps, resulting in longer start-up times of the system 100; - slowing down the operating cycle of system 100, so as to keep the process active even in the face of lower productivity. Both of these strategies require an adjustment of the control of the system 100: the optimization algorithm is therefore configured to communicate to the PLC controller 11 the request that they be implemented.

[00265] . This assumes that the control logic is distributed between the power apparatus 10 and the PLC controller 11, as both have a role in applying the optimization strategy.

[00266] . An even more general case involves the presence of two or more power devices 10 to be controlled in the same system 100, each equipped with a respective digital twin 102 and optimization algorithm.

[00267] . In this case, the optimization of the control of the system 100 can be carried out by a PLC controller 11 that manages all the devices 10, or by the regulators themselves, based on distributed logic.

[00268] . In the event that the new control conditions developed by algorithms 2031 and 2033 are not sufficient to maintain the production process of system 100 stable, phase 204 is activated to verify whether it is still possible to keep system 100 active for a limited time, before its definitive shutdown.

[00269] . This function can prevent the system from suddenly stopping, a condition that could cause serious damage. This situation is particularly critical in the case of production lines where several systems are integrated 100.

[00270] . In this case, ensuring a synchronised shutdown between the various sections of a plant ensures the least possible impact on production and a faster plant restart.

[00271] . In the case of a system in which a single power electronics device 10 predicts operating anomalies, it is this same device 10 that determines the remaining production time based on the optimal control strategy. Device 10 is configured to pass this information on the remaining operating time to the process controller or PLC 11, which prepares the rest of the system 100 for shutdown and - if applicable - also informs other machines upstream and / or downstream in the production line.

[00272] . In the case of systems 100 with multiple power devices 10 that foresee operating anomalies, it is the task of the PLC process controller 11 to coordinate the shutdown based on the residual times of each device 10 and their position in the production process.

[00273] . To the embodiments of the method and system described above, a person skilled in the art, in order to meet contingent needs, may make modifications, adaptations and replacements of elements with functionally equivalent ones, without departing from the scope of the following claims. Each of the features described as belonging to a possible embodiment can be implemented independently of the 5 other embodiments described.

Claims

1. A method of controlling (200) a power electronic apparatus (10), employable in a control system (100) of a thermal processing process for processing a material, for regulating the temperature of a heating load (3) following the advance detection of malfunctions in the processing process, wherein the control system (100) comprises: - a heat transfer element (1) configured to be in contact with said material to be subjected to the thermal processing process; - a heating load (3);- said power electronic apparatus (10) capable of transferring electric current (Iout, Ires) to the heating load (3) to modify the thermal state of the heat transmission element (1), said power electronic apparatus (10) including a real power converter (101) driven by an input duty cycle signal (DTin) generated by the power electronic apparatus (10) to enable / disable the transfer of electric current to the heating load (3);- a process controller (11) configured to compare a detected current temperature value (Tmat) of the material subjected to the thermal processing with a reference temperature value (Ttgt), said process controller (11) being configured to control the power electronics apparatus (10) by modifying the input duty cycle signal (DTin) applied to the real power converter (101) to bring the detected current temperature value (Tmat) closer to the reference temperature value (Ttgt); - a functional block representing a digital twin converter (102) of said real power converter (101);- a functional block representing a comparison algorithm (103') operatively associated with the digital twin converter (102), the method (200) comprising the steps of: - detecting in advance and classifying (201), by the comparison algorithm functional block (103'), at least one malfunction in the thermal process of processing the material;- evaluating (202) a modification of the control applied, by the process controller (11), to the power electronic apparatus (10) following the early detection of said at least one malfunction, the method (200) comprising the further step of: - applying (203) said modified control to the power electronic apparatus (10) to prevent the at least one malfunction detected in advance in case such modified control ensures the integrity of the performed machining process, or - shutting down (204) in a controlled manner the control system (100) in case such modified control compromises the integrity of the performed machining process.; 2. A control method (200) according to claim 1, wherein said step of detecting in advance and classifying (201) at least one malfunction in the thermal material processing process comprises the steps of: - starting (2011) the digital twin converter functional block (102) together with the control system (100); - simulating (2012), by the digital twin converter functional block (102), a plurality of sets of temperature values ​​(Ti(t+dt, ..., t+f)) at a plurality of specific points inside the real power converter (101) of the power electronic apparatus (10), each of the simulated temperature values ​​of each set being associated with a time instant of a plurality of time instants (t+dt, ..., t+f) of a prediction time interval (f) following a current time instant (t) of the manufacturing process, where dt is a predetermined sampling time interval; - compare (2012') each of the simulated temperature values ​​of each set (Ti(t+dt, ..., t+f)) with a respective threshold temperature value (TSi) of a plurality of threshold values; in the case where at least one of the simulated temperature values ​​of one of the sets (Ti(t+dt, ..., t+f)) is greater than the respective threshold temperature value (TSi), the method includes the following phases: - classifying (2013), for each set of temperature values, by the comparison algorithm functional block (103'), the at least one malfunction that has caused said at least one simulated temperature value to be greater than the respective threshold temperature value (TSi); - activating (2014) a feedback algorithm (103) operationally associated with the digital twin converter (102), configured to evaluate whether said plurality of sets of simulated temperature values ​​(Ti(t+dt, ..., t+f)) provided by the digital twin converter (102) are acceptable in relation to the thermal process parameters.

3. A control method (200) according to claim 1 or 2, wherein said step of detecting in advance and classifying (201) at least one malfunction in the thermal material processing process comprises the steps of: - starting (2011) the digital twin converter functional block (102) together with the control system (100); - simulating (2012), by the digital twin converter functional block (102), a plurality of sets of temperature values ​​(Ti(t+dt, ..., t+f)) at a plurality of specific points inside the real power converter (101) of the power electronic apparatus (10), each of the simulated temperature values ​​of each set being associated with a time instant of a plurality of time instants (t+dt, t+2dt, ..., t+f) of a prediction time interval (f) following a current time instant (t) of the manufacturing process, where dt is a fixed sampling time interval; - calculate (2012a), for each specific point inside the real power converter (101), mean temperature variation values ​​(VTi(t+dt, ..., t+f), VTi(t+2dt, ..., t+f),..., VTi(t+f-dt, ..., t+f)), each of said mean temperature variation values ​​(VTi) at the specific internal point being calculated in a number of time instants equal to or less than the plurality of time instants (t+dt, t+2dt, ., t+f) of the prediction time interval (f) following a current time instant (t) of the process, each of said mean temperature variation values ​​(VTi) being related to a plurality of temperature values ​​(Ti(t+dt,., t+f), Ti(t+2dt,., t+f),., Ti(t+f-dt,., t+f)) simulated at that point; - compare (2012a'), each of the values ​​of mean temperature variation (VTi(t+dt, ., t+f), VTi(t+2dt, ., t+f),., VTi(t+f-dt, ., t+f)) associated with the specific internal point, with a value of threshold mean temperature variation (VT*i) specific for that point and determined on the basis of the thermal process; in the case where at least one of the mean temperature variation values ​​(VTi(t+dt, ..., t+f), VTi(t+2dt, ..., t+f),..., VTi(t+fdt, ..., t+f)) is greater than such respective threshold mean temperature variation value (VT*i), the method comprises the phases of: - classifying (2013a), by the comparison algorithm functional block (103'), the at least one malfunction that has determined at least one of the mean temperature variation values ​​(VTi(t+dt, ., t+f), VTi(t+2dt, ., t+f),., VTi(t+f-dt, ., t+f)) greater than the respective threshold mean temperature variation value (VT*i); - activate (2014) a feedback algorithm (103) operationally associated with the digital twin converter (102), configured to evaluate whether said plurality of sets of simulated temperature values ​​(Ti(t+dt, ., t+f)) provided by the digital twin converter (102) are acceptable in relation to the thermal process parameters.

4. A control method (200) according to claim 1 or 2, wherein said step of detecting in advance and classifying (201) at least one malfunction in the thermal material processing process comprises the steps of: - starting (2011) the digital twin converter functional block (102) together with the control system (100); - simulating (2012), by the digital twin converter functional block (102), a plurality of sets of temperature values ​​(Ti(t+dt, , t+f)) at a plurality of specific points inside the real power converter (101) of the power electronic apparatus (10); each of the simulated temperature values ​​of each set being associated with a time instant of a plurality of time instants (t+dt, t + 2dt, ..., t+f) of a prediction time interval (f) following a current time instant (t) of the manufacturing process, where dt is a predetermined sampling time interval; - associate (2012b), to each specific internal point, a set of binary parameters representing the presence / absence of oscillatory trends (AOi(t+dt,...,t+f), AOi(t+2dt,...,t+f), ..., AOi(t+(f-dt),...,t+f)) of temperature, each of said parameters being related to a sub-interval of temperatures (Ti(t+dt,...,t+f), Ti(t+2dt,...,t+f),..., Ti(t+(f-dt),...,t+f)) of said simulated temperature values ​​Ti(t+dt, ., t+f); - detect (2012b'), in at least one of said temperature sub-intervals, the presence of at least one oscillatory trend (AOi(t+dt,.,t+f), AOi(t+2dt,.,t+f), ..., AOi(t+(f-dt),...,t+f)) of the temperature, relative to the specific internal point, not compliant with the process when said sub-interval comprises at least one value of the first local minimum of temperature, one value of the first local maximum of temperature and a further value of the second consecutive local minimum of temperature or at least one value of the first local maximum of temperature, one value of the first local minimum of temperature and a further value of the second consecutive local maximum of temperature; - classify (2013b), by the comparison algorithm functional block (103'), the at least one malfunction that has determined the presence of said at least one oscillatory trend (AOi(t+dt,..,t + f), AOi(t+2dt,...,t+f), ..., AOi(t+(f-dt),...,t+f)) of temperature; - activate (2014) a feedback algorithm (103) operationally associated with the digital twin converter (102), configured to evaluate whether said plurality of sets of temperature values ​​(Ti(t+dt, ..., t+f)) simulated provided by the digital twin converter (102) are acceptable in relation to the thermal process parameters.

5. A control method (200) according to any preceding claim, wherein said shutdown step (204) comprises a step of performing a controlled shutdown of the control system (100), by the process controller (11), after a shutdown time interval (s(n)) from said current time instant (t), wherein said shutdown time interval (s(n)) has duration shorter than the prediction time interval (f).

6. Control method (200) according to claim 5, wherein said turn-off time interval (s(n)) is calculated by the expression: s(n) = f- (n*dt) where f is said prediction time interval, dt is the sampling time interval and n is an integer.

7. A control method (200) according to any of claims 1-6, wherein said steps of evaluating (202) a modification of the applied control and applying (203) said modified control to the power electronic apparatus (10) to prevent the at least one pre-detected malfunction comprises the steps of: - executing a first control optimization algorithm (2031) applied to said power electronic apparatus (10); - evaluating (2032) whether the at least one pre-detected malfunction is prevented based on the execution of said first control optimization algorithm (2031); in the event that the at least one pre-detected malfunction persists, the method comprises the further steps of: - executing a second control optimization algorithm (2033) applied to said process controller (11);- evaluating (2034) whether the at least one pre-detected malfunction is prevented based on the execution of said second control optimization algorithm (2033); in case the at least one pre-detected malfunction persists, the method comprises said step of turning off (204) the control system (100).; 8. Control method (200) according to claim 7, wherein said first control optimization algorithm (2031) comprises a step of evaluating (2031a) whether a reduction of the input duty cycle signal (DTin) applied to the real power electronic converter (101) maintains the integrity of the performed machining process; in the case where the integrity of the performed machining process is maintained, the method comprises the steps of: - calculating (2031'), by said feedback algorithm (103), a first plurality of duty cycle values ​​(D-C1(t+dt, ..., t+f)) each associated with a time instant (t+1, ..., t+f) of said plurality of instants in said prediction time interval (f), subsequent to a current time instant (t) of the machining process; - making available said first plurality of duty cycle values ​​(D-C1(t+dt, ..., t+f), t+f)) to the digital twin converter functional block (102); - simulating (2031''), by the digital twin converter functional block (102), a plurality of sets of first temperature values ​​(T1i(t+dt, ..., t+f)) at a plurality of specific internal points of the real power converter (101) of the power electronic apparatus (10), each of said sets of simulated first temperature values ​​being associated with said first plurality of duty cycle values ​​(D-C1(t+dt, ..., t+f)) at a time instant (t+dt, ..., t+f); - comparing (2031b) each of said sets of simulated first temperature values ​​(T1i(t+dt, ..., t+f)) with a respective threshold temperature value (TSi) of a plurality of threshold values; in case each of the first temperature values ​​(T1i(t+dt, ..., t+f)) simulated at each specific internal point is less than the respective threshold value (TSi), the method includes the steps of: - applying (2031c) to the real power converter (101) an input duty cycle signal (DTin) including said first plurality of duty cycle values ​​(D-C1(t+dt, ..., t+f)) to enable / disable the transfer of electric current to the heating load (3); - signalling, by the feedback algorithm functional block (103), an error condition to a supervision unit of the control system (11); in the case in which at least one of the first temperature values ​​(T1i(t+dt, ..., t+f)) simulated at each specific internal point is greater than the respective threshold value (TSi), the method comprises the steps of: - repeating said step of evaluating (2031a) whether a further reduction of the input duty cycle signal (DTin) applied to the real power electronic converter (101) maintains the integrity of the performed machining process; and in the case in which the integrity of the performed machining process is maintained, the step of: - calculating (2031'), by said feedback algorithm (103), a further first plurality of duty cycle values ​​(D-C1'(t+dt, ..., t+f)) each associated with a time instant (t+1, ..., t+f) of said plurality of instants in said prediction time interval (f).

9. A control method (200) according to claim 7 or 8, wherein said second control optimization algorithm (2033) comprises a step of evaluating (2033a) whether an increment of the input duty cycle signal (DTin) applied to the real power electronic converter (101) maintains the integrity of the performed machining process; in the case where the integrity of the performed machining process is maintained, the method comprises the steps of: - calculating (2033'), by said feedback algorithm (103), a second plurality of duty cycle values ​​(D-C2(t+dt, ..., t+f)) each associated with a time instant (t+dt, ..., t+f) of said plurality of time instants in said prediction time interval (f), subsequent to a current time instant (t) of the machining process; - make available said second plurality of duty cycle values ​​(D-C2(t+dt, ..., t+f)) to the digital twin converter functional block (102); - simulating (2033''), by the digital twin converter functional block (102), a plurality of sets of second temperature values ​​(T2i(t+dt, ..., t+f)) at a plurality of specific internal points of the real power converter (101) of the power electronic apparatus (10), each of said sets of simulated second temperature values ​​being associated with said second plurality of duty cycle values ​​(D-C2(t+dt, ..., t+f)) at a time instant (t+dt, ..., t+f) of said prediction time interval (f); - comparing (2033b) each of said sets of simulated second temperature values ​​(T2i(t+1, ..., t+f)) with a respective threshold temperature value (TSi) of a plurality of threshold values; in case each of the second temperature values ​​(T2i(t+dt, ..., t+f)) simulated at each specific internal point is less than the respective threshold value (TSi), the method includes the steps of: - applying (2033c) to the real power converter (101) an input duty cycle signal (DTin) including said second plurality of duty cycle values ​​(D-C2(t+dt, ..., t+f)) to enable / disable the transfer of electric current to the heating load (3); - signalling, by the feedback algorithm functional block (103), an error condition to a supervision unit of the control system (11); in the case in which at least one of the second temperature values ​​(T2i(t+dt, ..., t+f)) simulated at each specific internal point is greater than the respective threshold value (TSi), the method comprises the steps of: - repeating said step of evaluating (2033a) whether a further increase of the input duty cycle signal (DTin) applied to the real power electronic converter (101) maintains the integrity of the performed machining process; and in the case in which the integrity of the performed machining process is maintained, the step of: - calculating (2033'), by said feedback algorithm (103), a further second plurality of duty cycle values ​​(D-C2'(t+dt, ..., t+f)) each associated with a time instant (t+dt, ..., t+f) of said plurality of instants in said prediction time interval (f).

10. Control method (200) according to claim 5 or 6, wherein said step of switching off (204) the control system (100) comprises a step of evaluating (2041) whether the integrity of the performed machining process is maintained for a controlled shutdown time interval of limited duration (s(n)) which takes on discrete values ​​given by s(n)=f-(n*dt), of anticipating the shutdown of the system at time t+s(n); in the case in which the integrity of the performed machining process is maintained for said controlled shutdown time interval (s(n)) of limited duration, the method comprises a step of repeating, for each of the discrete values ​​of such controlled shutdown time interval for which (n*dt) <f, le fasi di: - calcolare (2042), da parte di detto algoritmo di retroazione (103), una terza pluralità di valori di duty cycle (D-C3(t+dt, ..., t+s(n))) ciascuno associato ad un istante di tempo (t+dt, ..., t+s(n)) of said plurality of time instants of the controlled turn-off time interval (s(n)); - making said third plurality of duty cycle values ​​(D-C3(t+dt, ..., t+s(n))) available to the digital twin converter functional block (102); - simulating (2043), by the digital twin converter functional block (102), a plurality of sets of third temperature values ​​(T3i(t+dt, ..., t+s(n))) at a plurality of specific internal points of the real power converter (101) of the power electronic apparatus (10), each of said sets of simulated third temperature values ​​being associated with said third plurality of duty cycle values ​​(D-C3(t+dt, ..., t+s(n))) at a time instant (t+dt, ..., t+s(n)) of said controlled turn-off time interval (s(n)); - compare (2044) each of said sets of third temperature values ​​(T3i(t+dt, ..., t+s(n))) simulated with respective sets of threshold temperature values ​​(TSi); said repeat step being performed in case at least one of the third temperature values ​​(T3i(t+1, ..., t+s(n)) simulated at each specific internal point is greater than the respective threshold value (TSi).

11. Control method (200) according to the preceding claim, wherein said step of switching off (204) the control system (100) comprises, in the case where each of the third temperature values ​​(T3i(t+dt, ..., t+s(n)) simulated at a specific internal point is lower than the respective threshold value (TSi), the steps of: - applying (2045) to the real power converter (101) an input duty cycle signal (DTin) including said third plurality of duty cycle values ​​(D-C3(t+dt, ..., t+s(n))) to enable / disable the transfer of electric current to the heating load (3); - signalling, by the feedback algorithm functional block (103), an error condition to a supervision unit (12) of the control system (100).

12. Control method (200) according to any of claims 5-6 or 10-11, wherein said step of switching off (204) the control system (100) further comprises a step of initiating and signaling (2046), by the process controller (11), a programmed shutdown of the control system (100) after said controlled shutdown time interval (s(n)) from the current time instant (t).

13. Control method (200) according to claims 10-11, wherein said step of switching off (204) the control system (100) further comprises a step of initiating and signalling (2047), by the process controller (11), an emergency shutdown of the control system (100) if at the current time instant (t) it is not possible to maintain the integrity of the process for a controlled shutdown time interval (s(n)).

14. A control system (100) for a thermal processing of a material, for regulating the temperature of a heating load (3) following the early detection of malfunctions in the processing, the control system (100) comprising: - a heat transfer element (1) configured to be in contact with said material to be subjected to the thermal processing; - a heating load (3); - a power electronic apparatus (10) adapted to transfer electric current (Iout, Ires) to the heating load (3) to modify the thermal state of the container element (1), said power electronic apparatus (10) including a real power converter (101) driven by an input duty cycle signal (DTin) generated by the power electronic apparatus (10) to enable / disable the transfer of said electric current to the heating load (3);- a process controller (11) configured to compare a detected current temperature value (Tmat) of the material subjected to the thermal processing with a reference temperature value (Ttgt), said process controller (11) being configured to control the power electronics apparatus (10) by modifying the input duty cycle signal (DTin) applied to the real power converter (101) to bring the detected current temperature value (Tmat) closer to the reference temperature value (Ttgt); - a functional block representing a digital twin converter (102) of said real power converter (101);- functional blocks representative of a comparison algorithm (103') and a feedback algorithm (103), respectively, operatively associated with the digital twin converter (102), said control system (100) being configured to perform the control method (200) in accordance with any one of claims 1-13.; 15. Control system (100) of a thermal process for processing a material according to claim 14, wherein said process controller (11) is a programmable logic controller device (Programmable Logic Controller), of an independent type (stand-alone) with respect to the power electronic apparatus (10) or integrated into such power electronic apparatus.

16. Control system (100) of a thermal process for the processing of a material according to claim 14, wherein said process controller (11) is a logic algorithm implemented on devices selected from the group consisting of: operator panels or HMIs, industrial PCs, Edge Servers.

17. Control system (100) of a thermal material processing process according to any of claims 14-16, further comprising a supervisory unit (12) configured to exchange process data (D) with said process controller (11), said supervisory unit (12) being integrated into the process controller (11), or being implemented on a remote cloud platform capable of communicating with the process controller (11) via a telecommunications network.

18. Control system (100) of a thermal process for processing a material according to claim 17, wherein the digital twin electronic converter functional block (102) of the real converter (101) is implemented in one of the components of the control system (100) of the thermal process chosen from the group consisting of: power electronic apparatus (10), process controller (11), supervision unit (12).

19. Control system (100) of a thermal process for processing a material according to claim 17, wherein one of the components of the control system (100) of the thermal process chosen from the group consisting of: power electronic apparatus (10), process controller (11), supervision unit (12), is configured to implement the functional blocks representing a feedback algorithm (103) and a comparison algorithm (103').