A real-time compensation control method for thermal disc temperature deviation

Abstract

The present application relates to the technical field of industrial control system, specifically to a real-time compensation control method for temperature deviation of hot plate, comprising the following steps: reading actual temperature and set power of multi-temperature-zone hot plate, positioning static two-dimensional discrete lookup table, and matching to generate adjacent grid point coordinate set; respectively calculating difference value of the actual temperature and adjacent grid point temperature value, difference value of the set power and adjacent grid point power value, and generating interpolation weight coefficient. In the present application, each partition heating action can be made closer to the real thermal response process, the temperature consistency between partitions, the followability of control response and the stability of overall operation of the hot plate are enhanced, and under the working condition of multi-partition linkage, obvious thermal inertia and complex temperature difference transmission, the cumulative risk of over-compensation and under-compensation can be reduced, and the oscillation phenomenon in the target temperature approximation process can be improved.

Description

Technical Field

[0001] This invention relates to the field of industrial control system technology, and in particular to a real-time compensation control method for hot plate temperature deviation. Background Technology

[0002] Industrial control system technology involves real-time monitoring, data acquisition, process regulation, and execution control of the operating status of industrial production equipment. Through the coordinated operation of sensors, controllers, and actuators, it enables the regulation of key process parameters such as temperature, pressure, flow rate, and power.

[0003] Current technologies focus on real-time monitoring, data acquisition, process adjustment, and execution control of industrial production equipment. Their operation is based on the conventional collaborative relationships between sensors, controllers, and actuators. While capable of routinely controlling process parameters such as temperature, pressure, flow rate, and power, in multi-temperature zone heat exchangers and similar objects with thermal coupling and inertia, current operating methods tend to directly respond to single data acquisition results. When faced with the nonlinear relationship between temperature and power changes, this approach tends to compress multiple operating states into a single adjustment action. This results in a near-consistent control rhythm even when local temperature differences are at different stages of formation. In actual operation, this can lead to situations where one zone is already close to the target temperature while residual heat is still being transferred to neighboring zones, yet the control unit only issues adjustment commands based on the current data acquisition, causing the temperature to continue to rise. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the existing technology and propose a real-time compensation control method for hot plate temperature deviation.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a real-time compensation control method for hot plate temperature deviation, comprising the following steps:

[0006] Read the actual temperature and set power of the multi-temperature zone heat plate, locate the static two-dimensional discrete lookup table, and generate a set of coordinates of adjacent grid points; calculate the difference between the actual temperature and the temperature value of the adjacent grid points, and the difference between the set power and the power value of the adjacent grid points, respectively, and generate interpolation weighting coefficients.

[0007] The interpolation weight coefficients are associated with the basic duty cycle of the adjacent grid point coordinate set to generate a nonlinear electrothermal compensation duty cycle; the heating power values ​​of each zone in the previous 10 consecutive control cycles are extracted to generate a previous power time series tensor.

[0008] Obtain the spatial response kernel function matrix and the temporal response kernel function matrix. Calculate the preceding power time series tensor with the spatial response kernel function matrix and the temporal response kernel function matrix to generate a discrete two-dimensional spatiotemporal convolution term. Accumulate each element within the discrete two-dimensional spatiotemporal convolution term to generate a heat conduction feedforward cancellation amount.

[0009] The heat conduction feedforward cancellation amount is converted into a pulse width modulation duty cycle reduction amount; the nonlinear electrothermal compensation duty cycle is subtracted from the pulse width modulation duty cycle reduction amount to generate a collaborative compensation control amount; the collaborative compensation control amount is converted into an electrical signal pulse width duty cycle format to generate a hot plate drive output signal.

[0010] Preferably, the step of obtaining the interpolation weight coefficients is as follows:

[0011] Read the actual temperature and set power of the multi-temperature zone hot plate, pair and register them in the order of actual temperature value first and set power value second, check whether the temperature field and power field of each group of values ​​are complete, remove the pairing records with missing fields, and form the actual temperature and set power pairing values.

[0012] Locate the temperature and power coordinate axes in the static two-dimensional discrete lookup table, retrieve the temperature and power intervals where the actual temperature and power pairing values ​​are located, lock the boundary coordinate positions corresponding to the intersection intervals, and generate a set of coordinates for adjacent grid points.

[0013] Based on the coordinate set of adjacent grid points, the temperature value of the grid point corresponding to the actual temperature and the power value of the grid point corresponding to the set power are extracted one by one. The temperature difference and power difference are calculated respectively. The offset distance is converted according to the grid boundary span. Then, the difference ratio in each direction is allocated and calculated to obtain the interpolation weight coefficient.

[0014] Preferably, the step of obtaining the nonlinear electrothermal compensation duty cycle is as follows:

[0015] Based on the interpolation weight coefficient and the coordinate set of adjacent grid points, the corresponding basic duty cycle is retrieved point by point according to the coordinate position of each grid point in the coordinate set of adjacent grid points. The interpolation weight coefficient corresponding to each grid point coordinate position is multiplied into the corresponding basic duty cycle in turn. Then, the cumulative sum is calculated according to the arrangement order of the coordinate positions of adjacent grid points. The correction increment under the combined effect of the coordinate positions of each grid point is extracted to obtain the deviation correction amount.

[0016] Retrieve the initial duty cycle record of the current control cycle for the partition corresponding to the deviation correction amount, write the deviation correction amount into the duty cycle correction field of the initial duty cycle record, complete the duty cycle adjustment according to the correspondence between the duty cycle correction field and the initial duty cycle value, output the duty cycle correction result of the partition corresponding to the current control cycle, and generate the nonlinear electrothermal compensation duty cycle.

[0017] Preferably, the step of obtaining the preceding power time series tensor is as follows:

[0018] The heating power values ​​of each partition corresponding to the nonlinear electrothermal compensation duty cycle are locked within the first 10 consecutive control cycles. The heating power values ​​of each partition are written into the queue storage unit one by one according to the order of the control cycles. After each writing, the original queue elements are transferred sequentially along the time direction. Then, the elements are tensorized and arranged according to the partition position and the control cycle position to generate the first power time series tensor.

[0019] Preferably, the step of obtaining the discrete two-dimensional spatiotemporal convolution term is as follows:

[0020] Based on the Green's function analytic terms, the partition location identifier, partition spacing parameter, time step parameter, and thermal response decay parameter are read item by item. The spatial response weights are filled in according to the spatial arrangement order corresponding to the partition location identifiers, and the time response weights are filled in according to the time arrangement order corresponding to the time step parameters, thus forming the spatial response kernel function matrix and the time response kernel function matrix.

[0021] The heating power values ​​of each partition in the preceding power time series tensor are retrieved, and the power components of adjacent partitions are extracted layer by layer according to the target partition position. The power components of the historical period are extracted in the order of the control cycle. The power components of adjacent partitions are accumulated with the corresponding position values ​​of the spatial response kernel function matrix, and the power components of the historical period are accumulated with the corresponding position values ​​of the time response kernel function matrix to form a discrete two-dimensional spatiotemporal convolution term.

[0022] Preferably, the step of obtaining the heat conduction feedforward cancellation amount is as follows:

[0023] Read the convolution values ​​of each element in the discrete two-dimensional spatiotemporal convolution term one by one, perform element aggregation within the same partition according to the target partition position, perform cross-period element aggregation according to the control cycle order, extract the heat flow transfer response delay value, write the heat flow transfer response delay value into the target partition power deduction field, and convert it into the target partition power deduction amount according to the correspondence of the target partition power deduction field to generate the heat conduction feedforward cancellation amount.

[0024] Preferably, the step of obtaining the collaborative compensation control quantity is as follows:

[0025] Read the target partition power deduction value corresponding to the heat conduction feedforward cancellation amount, retrieve the duty cycle conversion reference, duty cycle resolution identifier, and duty cycle upper and lower limit identifier of the current control cycle of the target partition, map the target partition power deduction value to the duty cycle change value item by item according to the duty cycle conversion reference, perform gear alignment processing according to the duty cycle resolution identifier, and filter out the excessive change segment according to the duty cycle upper and lower limit identifier to obtain the pulse width modulation duty cycle reduction.

[0026] Read the target partition duty cycle value in the nonlinear electrothermal compensation duty cycle, and map the pulse width modulation duty cycle reduction to the same target partition duty cycle value according to the target partition identifier. Extract the difference between the target partition duty cycle value and the pulse width modulation duty cycle reduction, register the remaining duty cycle value of each target partition, extract the heat flow balance correction amount, and generate the collaborative compensation control amount.

[0027] Preferably, the step of acquiring the hot plate drive output signal is as follows:

[0028] Read the duty cycle values ​​of each target partition in the collaborative compensation control quantity, convert the duty cycle values ​​of each target partition into the conduction period identifier, off period identifier, and cycle length identifier corresponding to the electrical signal pulse width duty cycle format, and write the conduction period identifier, off period identifier, and cycle length identifier into the output register position of the corresponding relay driver terminal according to the partition channel number to generate the hot plate drive output signal.

[0029] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0030] In this invention, the actual temperature and set power of the multi-temperature zone heating plate are read synchronously, and the actual temperature and set power are directly mapped to the coordinates of adjacent grid points in a static two-dimensional discrete lookup table. This transforms the continuously changing temperature and power states into a locatable, comparable, and subdividable discrete correlation, allowing compensation to move beyond a coarse correction of a single parameter deviation to a refined adjustment involving both temperature and power dimensions. Interpolation weighting coefficients are formed using the differences between the actual temperature and the temperature values ​​of adjacent grid points, and the differences between the set power and the power values ​​of adjacent grid points. These coefficients are then used to calculate the base duty cycle, enabling continuous adjustment of the output duty cycle as the operating state changes, mitigating compensation jumps that easily occur during temperature zone switching. Simultaneously, the heating power values ​​of each zone within the pre-control cycle are continuously extracted and organized into a pre-control power time series tensor, ensuring that compensation judgments not only correspond to the instantaneous state at the current moment but also consider the continuous time range of each zone. The power evolution process within the system is incorporated into a unified processing link. By performing spatiotemporal correlation calculations on the preceding power time series tensor in conjunction with the spatial response kernel function matrix and the time response kernel function matrix, the temperature impact caused by heat diffusion between adjacent intervals and the delayed transmission of historical heating processes can be factored into the control quantity correction. This allows for the early identification of future heat flow trends before significant temperature deviations occur. The discrete two-dimensional spatiotemporal convolution terms are accumulated to generate a heat conduction feedforward offset, which is then differentially superimposed with the nonlinear electrothermal compensation duty cycle to form a collaborative compensation control quantity that takes into account both the current deviation state and the future heat conduction impact. Finally, after being converted into a heat plate drive output signal, it can make the heating actions of each zone more closely resemble the actual thermal response process, enhance the temperature consistency between zones, the follow-up of the control response, and the overall stability of the heat plate operation. In operating conditions with multi-zone linkage, significant thermal inertia, and complex temperature difference transmission, it can reduce the cumulative risk of overcompensation and undercompensation and improve the oscillation phenomenon during the target temperature approach process. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the steps of the present invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0033] Please see Figure 1 This invention provides a technical solution: a real-time compensation control method for hot plate temperature deviation, comprising the following steps:

[0034] Read the actual temperature and set power of the multi-temperature zone heat plate, locate the static two-dimensional discrete lookup table, and generate the coordinate set of adjacent grid points; calculate the difference between the actual temperature and the temperature value of adjacent grid points, and the difference between the set power and the power value of adjacent grid points, respectively, and generate interpolation weight coefficients.

[0035] The interpolation weighting coefficients are associated with the basic duty cycle of the adjacent grid point coordinate set to generate the nonlinear electrothermal compensation duty cycle; the heating power values ​​of each zone in the previous 10 consecutive control cycles are extracted to generate the previous power time series tensor.

[0036] Obtain the spatial response kernel function matrix and the temporal response kernel function matrix. Calculate the preceding power time series tensor with the spatial response kernel function matrix and the temporal response kernel function matrix to generate a discrete two-dimensional spatiotemporal convolution term. Accumulate each element within the discrete two-dimensional spatiotemporal convolution term to generate the heat conduction feedforward cancellation amount.

[0037] The heat conduction feedforward cancellation amount is converted into a pulse width modulation duty cycle reduction amount; the nonlinear electrothermal compensation duty cycle is subtracted from the pulse width modulation duty cycle reduction amount to generate a collaborative compensation control amount; the collaborative compensation control amount is converted into an electrical signal pulse width duty cycle format to generate a hot plate drive output signal.

[0038] The steps to obtain the interpolation weight coefficients are as follows:

[0039] Read the actual temperature and set power of the multi-temperature zone hot plate, pair and register them in the order of actual temperature value first and set power value second, check whether the temperature field and power field of each group of values ​​are complete, remove the pairing records with missing fields, and form the actual temperature and set power pairing values.

[0040] Locate the temperature and power coordinate axes in the static two-dimensional discrete lookup table, retrieve the temperature and power intervals where the actual temperature and power pairing values ​​are located, lock the boundary coordinate positions corresponding to the intersection intervals, and generate a set of coordinates for adjacent grid points.

[0041] Based on the coordinate set of adjacent grid points, extract the temperature value of the grid point corresponding to the actual temperature and the power value of the grid point corresponding to the set power one by one, calculate the temperature difference and power difference respectively, convert the offset distance according to the grid boundary span, and then distribute the difference ratio in each direction to obtain the interpolation weight coefficient.

[0042] Specifically, the system reads the actual temperature values ​​collected by the thermocouple sensors of each zone on the multi-temperature zone hot plate and the set power percentage for each zone issued by the main controller. For example, for a 5-zone hot plate, the temperature array collected within one control cycle is... and power setting array The actual temperature and set power of each zone are combined into data pairs in the order of temperature first and power second, forming an initial set of paired records. Next, data integrity and validity checks are performed on each pair of records in the set. First, the temperature and power fields are checked for null values ​​or non-numeric (NaN) entries. Then, each value is compared to a preset valid data range, which is set according to the physical characteristics and process requirements of the heat plate. For example, the valid temperature range is set to 25℃ to 450℃, and the valid power range is set to 0% to 100%. If a pair of records contains a temperature value of -99℃ or a power value of 105%, it is considered invalid data. All pair of records containing invalid fields or values ​​outside the valid range are removed from the set. For example, if... In If the value is 500℃, which exceeds the upper limit of 450℃, the pairing record is discarded and not processed further. All the pairing records that pass the verification are summarized to form a two-dimensional array consisting of valid data, namely the actual temperature set power pairing value.

[0043] Based on the generated static two-dimensional discrete lookup table and the actual temperature set power pairing values ​​obtained in the previous step, for each pairing value... The lookup table is a pre-built matrix where row axes represent discrete temperature points and column axes represent discrete power points. The axes are constructed based on historical equipment operating data and thermal model simulations. For example, the temperature axis... Set as ℃, power coordinate axis Set as %, firstly, on the temperature axis The actual temperature value is retrieved using a binary search method. Determine the lower boundary temperature of the interval in which it is located. and upper boundary temperature , making For example, when At 165℃, on the temperature coordinate axis Positioning to the interval ,therefore ℃, ℃, similarly, on the power axis Search for the set power value Determine the lower boundary power of the interval in which it is located. and upper boundary power , making For example, when When it is 55%, on the power axis Positioning to the interval ,therefore %, %, lock the rectangular grid formed by the intersection of these two intervals, and the coordinates of the four vertices of this grid. , , , These are the boundary coordinates we are looking for. We combine these four coordinates into a set to generate the coordinate set of adjacent grid points.

[0044] Based on the set of coordinates of adjacent grid points generated in the previous step And the actual temperature setting power matching value input. To calculate the interpolation weighting coefficients, firstly, the temperature values ​​of the grid points are extracted from the coordinate set. and and power value and Then calculate the actual points. The relative positions within the grid are determined by calculating and normalizing the temperature and power differences. This normalization process involves converting the offset distance, which is calculated as follows:

[0045] ;

[0046] ;

[0047] in, It is the normalized bias distance in the temperature direction. It is the normalized bias distance in the power direction. and It is the actual temperature and the set power. , , , The boundary temperature and power values ​​of adjacent grid points are used. Then, based on these two offset distances, the interpolation weights of the four adjacent grid points are calculated. The weights are inversely proportional to the distance from the actual point to the grid point. The specific allocation calculation is as follows:

[0048] ;

[0049] ;

[0050] ;

[0051] ;

[0052] in, Corresponding coordinates The weight, Corresponding coordinates The weight, Corresponding coordinates The weight, Corresponding coordinates The weights, the set of these four weight values. This is the final interpolation weight coefficient.

[0053] The steps for obtaining the duty cycle of nonlinear electrothermal compensation are as follows:

[0054] Based on the interpolation weight coefficients and the coordinate set of adjacent grid points, the corresponding basic duty cycle is retrieved point by point according to the coordinate position of each grid point in the coordinate set of adjacent grid points. The interpolation weight coefficient corresponding to each grid point coordinate position is multiplied into the corresponding basic duty cycle in turn. Then, the cumulative sum is calculated according to the arrangement order of the coordinate positions of adjacent grid points. The correction increment under the combined effect of the coordinate positions of each grid point is extracted to obtain the deviation correction amount.

[0055] Retrieve the initial duty cycle record of the current control cycle for the partition corresponding to the deviation correction amount, write the deviation correction amount into the duty cycle correction field of the initial duty cycle record, complete the duty cycle adjustment according to the correspondence between the duty cycle correction field and the initial duty cycle value, output the duty cycle correction result of the partition corresponding to the current control cycle, and generate the nonlinear electrothermal compensation duty cycle.

[0056] Specifically, based on the interpolation weight coefficients obtained in the previous step... and the coordinate set of adjacent grid points First, a multidimensional lookup table is needed to store the base duty cycle increment values. This table is established through extensive experiments or thermal model simulations, with discrete temperature and power values ​​as its index, and storing the duty cycle correction increment under the corresponding operating conditions. For each coordinate point in the coordinate set of adjacent grid points, for example Retrieve the corresponding base duty cycle increment value from the multidimensional lookup table. For example, the values ​​obtained from the query are respectively , , , Next, the interpolation weight coefficient corresponding to the coordinate position of each grid point is multiplied by its corresponding base duty cycle increment value. For example, the weight... and Multiplying these values ​​yields the contribution of each grid point to the total increment. Finally, using bilinear interpolation, the contributions calculated for the four grid points are summed to obtain the final interpolation result. This result represents the correction increment under the current operating condition, and its calculation formula is as follows:

[0057] ;

[0058] in, It is the calculated interpolation duty cycle increment, which is also the deviation correction amount. , , , These are the four interpolation weight coefficients calculated earlier. , , , These are retrieved from the base duty cycle increment lookup table based on coordinates. , , , The retrieved base duty cycle increment value, for example, if the calculated weight is... The queried base increment value is The deviation correction amount is The deviation correction amount is obtained.

[0059] Retrieve the deviation correction amount calculated in the previous step And the initial duty cycle record of the corresponding partition in the current control cycle, the value of the initial duty cycle record. The duty cycle is typically set to the final output duty cycle of the previous control cycle, which serves as the baseline value for this adjustment. For example, if the final duty cycle of the previous cycle was 40%, then the initial duty cycle record value for the current cycle is 40%. Next, the deviation correction amount is... As a correction term, it is algebraically added to the initial duty cycle value to complete the initial adjustment of the duty cycle. For example, if the initial duty cycle is 40%, and the calculated deviation correction is +3.575%, then the initially adjusted duty cycle is... After completing the adjustment calculation, the results are checked for boundary conditions. Since the physical meaning of the duty cycle restricts its value to between 0% and 100%, the adjusted results need to be limited. If the calculated result is less than 0%, it is forcibly set to 0%; if it is greater than 100%, it is forcibly set to 100%. This process can be represented as follows: ,in The duty cycle correction result is after the amplitude limiting process. This result is then written into the duty cycle record of the corresponding partition of the current control cycle, overwriting the original initial value. The final output duty cycle correction result obtained after nonlinear table lookup, interpolation calculation, incremental compensation and boundary limiting is the nonlinear electrothermal compensation duty cycle.

[0060] The steps to obtain the preceding power time series tensor are as follows:

[0061] Lock the heating power values ​​of each partition corresponding to the nonlinear electrothermal compensation duty cycle within the preceding 10 consecutive control cycles. Write the heating power values ​​of each partition into the queue storage unit one by one according to the order of the control cycles. After each write, transfer the original queue elements sequentially along the time direction. Then, tensile the arrangement according to the partition position and the control cycle position to generate the preceding power time series tensor.

[0062] Specifically, based on the nonlinear electrothermal compensation duty cycle generated for all partitions on the hot plate in the previous step, this duty cycle value is first converted into a corresponding heating power value. In an ideal linear system, the heating power is proportional to the applied duty cycle, so a maximum power value can be set. Then the current control cycle Lower partition heating power value It can be calculated as ,in It is a partition In the cycle The nonlinear electrothermal compensation duty cycle is used to normalize the power, that is, the percentage value of the duty cycle is directly used as the heating power value. Next, a queue storage unit with a fixed length of 10 is maintained for each partition. This queue is used to store the heating power values ​​of the most recent 10 control cycles. In each new control cycle... At the beginning, set the latest heating power value. Push into the corresponding partition At the tail of the queue, and at the same time, the oldest power value at the head of the queue, i.e. The periodic power values ​​are removed from the queue, completing one data update. After all partition queues have been updated, these data are tensorized and arranged to create a two-dimensional matrix as the preceding power time series tensor. The size of this tensor is ,in This represents the total number of partitions on the hot disk, and the row index of the tensor. Corresponding partition number, column index (From 0 to 9) Corresponding to the chronological order of time, for example, It stores partitions exist The heating power value at each moment is used to generate the power time series tensor. Each row of the tensor represents the historical power sequence of a partition, while each column represents a power snapshot of all partitions at the same historical moment, ultimately generating the previous power time series tensor.

[0063] The steps for obtaining discrete two-dimensional spatiotemporal convolution terms are as follows:

[0064] Based on the Green's function analytic terms, the partition location identifier, partition spacing parameter, time step parameter, and thermal response decay parameter are read item by item. The spatial response weights are filled in according to the spatial arrangement order corresponding to the partition location identifiers, and the time response weights are filled in according to the time arrangement order corresponding to the time step parameters, thus forming the spatial response kernel function matrix and the time response kernel function matrix.

[0065] The heating power values ​​of each partition in the preceding power time series tensor are retrieved. The power components of adjacent partitions are extracted layer by layer according to the target partition position. The power components of the historical period are extracted in the order of the control cycle. The power components of adjacent partitions are accumulated with the corresponding position values ​​of the spatial response kernel function matrix, and the power components of the historical period are accumulated with the corresponding position values ​​of the time response kernel function matrix to form a discrete two-dimensional spatiotemporal convolution term.

[0066] Specifically, based on the analytical terms of the Green's function obtained in advance from the heat conduction equation or finite element simulation, spatial response kernel function matrices and time response kernel function matrices are constructed. First, key parameters are extracted from the discretized form of the Green's function. The partition location identifier is the unique index of each heating unit in the hot plate coordinate system, for example, from 1 to 25 for a 5x5 array. The partition spacing parameter is the physical distance between the center points of any two partitions. The time step parameter is the time slice length of the control system, for example, 0.5 seconds. The thermal response decay parameter is a constant related to the thermal diffusivity of the material; for example, for an aluminum hot plate, this value is 97× m² / s, calculate the physical adjacency between any two partitions according to the partition location identifier. and Spatial response weights between This weight typically decays exponentially with increasing distance, and its calculation formula is as follows:

[0067] ;

[0068] in, It is a partition and Distance from the center point This is the space decay constant, whose value is determined based on the size of the heat plate and the material properties. For example, it can be set to twice the minimum partition spacing to fill a space. The matrix, where To determine the total number of partitions, a spatial response kernel function matrix is ​​formed. Then, according to the historical time sequence defined by the time step parameter, the past [number]th ... The time response weight of the impact of each control cycle on the current moment This weight decays over time, and its calculation formula is as follows:

[0069] ;

[0070] in, It is an index of historical periods (from 1 to 10). It controls the duration of the cycle. It is the time decay constant. For example, it is set to 3 seconds according to experimental data. The 10 calculated weight values ​​are arranged into a 1x10 vector in time order to form the time response kernel function matrix.

[0071] Retrieve the data generated in the previous step Dimensional pre-power time series tensor ,in Representing partitions In the past The heating power value for each control cycle is retrieved simultaneously. Dimensional spatial response kernel matrix and Dimensional time response kernel matrix For each target partition The spatiotemporal coupling effects are calculated, which is accomplished by performing discrete two-dimensional spatiotemporal convolution operations. First, for the target partition... and a specific historical control cycle Calculate all partitions in this period The total spatial coupling effect produced by the power input is achieved by the first power input through the spatial response kernel matrix. The first row of the preceding power time series tensor Columns (all partitions in) The power at each time point is multiplied element-wise and then summed to obtain the power. This result represents the historical cycle At that time, the power applied to all partitions is applied to the current target partition. The resulting instantaneous thermal effect aggregation value is then compared with the corresponding value in the time response kernel function matrix. element Multiplying these yields the final contribution of the power input at that historical moment to the current moment. This calculation process is then applied to all 10 historical control cycles. Repeat the process and sum the 10 impact contributions to form a value representing the overall thermal crosstalk effect. This calculation is performed on each target partition, ultimately resulting in a partition containing... A vector of values, which is the discrete two-dimensional spatiotemporal convolution term.

[0072] The steps for obtaining the heat conduction feedforward cancellation amount are as follows:

[0073] Read the convolution values ​​of each element in the discrete two-dimensional spatiotemporal convolution term one by one, perform element aggregation within the same partition according to the target partition position, perform cross-period element aggregation according to the control cycle order, extract the heat flow transfer response delay value, write the heat flow transfer response delay value into the target partition power deduction field, and convert it into the target partition power deduction amount according to the correspondence of the target partition power deduction field to generate the heat conduction feedforward cancellation amount.

[0074] Specifically, the convolution values ​​corresponding to each target partition in the discrete two-dimensional spatiotemporal convolution term generated in the previous step are read one by one. These values ​​physically represent the predictive temperature drift that will appear in the current control cycle, caused by the power input of adjacent partitions and its own historical power input. For each target partition... The corresponding convolution value is the reference value that needs to be fed forward compensation, and this value is defined as the heat transfer response delay value. To convert this physically meaningful value into a quantity that can be directly used for power control, it is first written into a logical target partition power deduction field. Then, according to a preset power conversion relationship, the value in this field is converted into the actual power deduction amount. This conversion relationship is usually a scaling factor. The strength of the feedforward compensation is determined through system identification or offline optimization experiments. For example, by applying a step power perturbation to the system in steady state, observing and fitting the temperature response curve, the proportional relationship between theoretical heat conduction and actual temperature change can be calculated, thereby setting the appropriate parameters. The value, for example, is set to 0.9, which represents the target partition power deduction. The calculation is as follows ,in This is the feedforward gain coefficient, which ranges from 0 to 1. A higher value indicates stronger compensation but may increase the risk of system oscillation, while a lower value provides more conservative compensation. It is calculated for all partitions. Collected together, forming a collection A vector of power deduction values, which is the heat conduction feedforward cancellation amount.

[0075] The steps for obtaining the collaborative compensation control quantity are as follows:

[0076] Read the target partition power deduction value corresponding to the heat conduction feedforward cancellation amount, retrieve the duty cycle conversion reference, duty cycle resolution identifier, and duty cycle upper and lower limit identifier of the current control cycle of the target partition, map the target partition power deduction value to the duty cycle change value item by item according to the duty cycle conversion reference, perform gear alignment processing according to the duty cycle resolution identifier, and filter out the excessive change segment according to the duty cycle upper and lower limit identifier to obtain the pulse width modulation duty cycle reduction.

[0077] Read the target partition duty cycle value in the nonlinear electrothermal compensation duty cycle, match the pulse width modulation duty cycle reduction to the same target partition duty cycle value according to the target partition identifier, extract the difference between the target partition duty cycle value and the pulse width modulation duty cycle reduction, register the remaining duty cycle value of each target partition, extract the heat flow balance correction amount, and generate the collaborative compensation control amount.

[0078] Specifically, read the power deduction values ​​corresponding to each target partition in the heat conduction feedforward offset generated in the previous step, such as the target partition. The power deduction is Simultaneously, the duty cycle conversion reference, duty cycle resolution identifier, and upper and lower duty cycle limits for that zone are retrieved from the system configuration parameters. The duty cycle conversion reference is the rated maximum power of the heater in that zone. For example, 100W, this value is determined by the heater hardware specifications. The duty cycle resolution indicator is the minimum duty cycle step supported by the PWM controller, for example, 0.5%, determined by the hardware drive capability. The duty cycle upper and lower limits are the maximum allowable variation per cycle set for system stability, for example, ±10%, obtained through empirical debugging or system model simulation. First, according to the duty cycle conversion reference, the power deduction amount is... Linear mapping to duty cycle change values The calculation process is as follows: For example, if It is 8.3W. If it is 100W, then The percentage is 8.3%. Then, based on a 0.5% duty cycle resolution, the values ​​are... Perform gear alignment processing, which involves rounding 8.3% to the nearest multiple of 0.5%, and calculating it as follows: The value of the duty cycle change after alignment is obtained. Finally, based on the ±10% duty cycle upper and lower limits, the following is applied: The calculation process involves filtering out values. Since the calculated 8.5% is within the allowable range of -10% to +10%, the value is accepted. If the calculated result is 12%, it will be clamped to 10%. The final value, after conversion, alignment, and amplitude limiting, is the pulse width modulation duty cycle reduction.

[0079] Read the duty cycle values ​​of each target partition from the nonlinear electrothermal compensation duty cycle generated in the previous step, for example, the target partition. The duty cycle is Simultaneously, retrieve the pulse width modulation duty cycle reduction calculated in the previous step. According to the target partition identifier, these two values ​​are matched one-to-one. For example, the nonlinear heating compensation duty cycle of partition 5 is paired with the pulse width modulation duty cycle reduction of partition 5. Then, for each target partition, a duty cycle subtraction operation is performed, that is, the value of the nonlinear heating compensation duty cycle is subtracted from the corresponding pulse width modulation duty cycle reduction. The result calculated by this difference extraction process is regarded as an intermediate quantity and used for subsequent final duty cycle determination. For example, if partition Nonlinear electrothermal compensation duty cycle It is 55.0%, while the calculated pulse width modulation duty cycle reduction is... If it is 8.5%, then the difference obtained after performing the subtraction operation is 8.5%. The calculated remaining duty cycle value is registered. Since the duty cycle cannot be negative, a lower limit check is required before registration. If the calculation result is less than 0%, it is forcibly set to 0%. This final difference result is the heat flow balance correction amount. The heat flow balance correction amounts calculated for all partitions are collected to form a vector containing the final duty cycle of all partitions, generating the collaborative compensation control amount.

[0080] The steps for obtaining the output signal of the hot plate drive are as follows:

[0081] Read the duty cycle values ​​of each target partition in the collaborative compensation control quantity, convert the duty cycle values ​​of each target partition into the on-time identifier, off-time identifier, and cycle length identifier corresponding to the electrical signal pulse width duty cycle format, and write the on-time identifier, off-time identifier, and cycle length identifier into the output register position of the corresponding relay driver terminal according to the partition channel number to generate the hot plate drive output signal.

[0082] Specifically, it reads the duty cycle values ​​of each target partition contained in the collaborative compensation control quantity generated in the previous step, such as the target partition. The final duty cycle is This percentage value is then converted into the specific timing parameters required to drive the hardware. This conversion process requires a preset PWM cycle length identifier, which is set based on the switching characteristics of the heating relay and the response time of the thermal system, for example, set to 1000 milliseconds. Based on this cycle length, the on-time and off-time periods are calculated. The on-time identifier is calculated by multiplying the cycle length by the final duty cycle value, i.e. The shutdown period identifier is the cycle length minus the conduction period, i.e. For example, if the target partition If the coordinated compensation control amount is 46.5%, then its conduction period is identified as follows: The shutdown period is marked as After calculating the on and off periods of all partitions, the timing parameters are written into the PWM generator or timer comparison register of the corresponding channel in the microcontroller according to the pre-assigned hardware channel number for each partition. For example, the count value corresponding to the 465ms on period of partition 1 is written into the output comparison register of timer channel 1. When the timer count reaches this value, the level of the drive pin flips, thereby controlling the on and off of the solid-state relay connected to the channel. After completing this series of writing operations, the hardware begins to generate the corresponding pulse width modulation square wave according to the set duty cycle, generating the hot plate drive output signal.

[0083] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A real-time compensation control method for hot plate temperature deviation, characterized in that, Includes the following steps: Read the actual temperature and set power of the multi-temperature zone heat plate, locate the static two-dimensional discrete lookup table, and generate a set of coordinates of adjacent grid points; calculate the difference between the actual temperature and the temperature value of the adjacent grid points, and the difference between the set power and the power value of the adjacent grid points, respectively, and generate interpolation weighting coefficients. The interpolation weight coefficients are associated with the basic duty cycle of the adjacent grid point coordinate set to generate a nonlinear electrothermal compensation duty cycle; the heating power values ​​of each zone in the previous 10 consecutive control cycles are extracted to generate a previous power time series tensor. Obtain the spatial response kernel function matrix and the temporal response kernel function matrix. Calculate the preceding power time series tensor with the spatial response kernel function matrix and the temporal response kernel function matrix to generate a discrete two-dimensional spatiotemporal convolution term. Accumulate each element within the discrete two-dimensional spatiotemporal convolution term to generate a heat conduction feedforward cancellation amount. The heat conduction feedforward cancellation amount is converted into a pulse width modulation duty cycle reduction amount; the nonlinear electrothermal compensation duty cycle is subtracted from the pulse width modulation duty cycle reduction amount to generate a collaborative compensation control amount; the collaborative compensation control amount is converted into an electrical signal pulse width duty cycle format to generate a hot plate drive output signal.

2. The real-time compensation control method for hot plate temperature deviation according to claim 1, characterized in that, The steps for obtaining the interpolation weight coefficients are as follows: Read the actual temperature and set power of the multi-temperature zone hot plate, pair and register them in the order of actual temperature value first and set power value second, check whether the temperature field and power field of each group of values ​​are complete, remove the pairing records with missing fields, and form the actual temperature and set power pairing values. Locate the temperature and power coordinate axes in the static two-dimensional discrete lookup table, retrieve the temperature and power intervals where the actual temperature and power pairing values ​​are located, lock the boundary coordinate positions corresponding to the intersection intervals, and generate a set of coordinates for adjacent grid points. Based on the coordinate set of adjacent grid points, the temperature value of the grid point corresponding to the actual temperature and the power value of the grid point corresponding to the set power are extracted one by one. The temperature difference and power difference are calculated respectively. The offset distance is converted according to the grid boundary span. Then, the difference ratio in each direction is allocated and calculated to obtain the interpolation weight coefficient.

3. The real-time compensation control method for hot plate temperature deviation according to claim 1, characterized in that, The steps for obtaining the duty cycle of the nonlinear electrothermal compensation are as follows: Based on the interpolation weight coefficient and the coordinate set of adjacent grid points, the corresponding basic duty cycle is retrieved point by point according to the coordinate position of each grid point in the coordinate set of adjacent grid points. The interpolation weight coefficient corresponding to each grid point coordinate position is multiplied into the corresponding basic duty cycle in turn. Then, the cumulative sum is calculated according to the arrangement order of the coordinate positions of adjacent grid points. The correction increment under the combined effect of the coordinate positions of each grid point is extracted to obtain the deviation correction amount. Retrieve the initial duty cycle record of the current control cycle for the partition corresponding to the deviation correction amount, write the deviation correction amount into the duty cycle correction field of the initial duty cycle record, complete the duty cycle adjustment according to the correspondence between the duty cycle correction field and the initial duty cycle value, output the duty cycle correction result of the partition corresponding to the current control cycle, and generate the nonlinear electrothermal compensation duty cycle.

4. The real-time compensation control method for hot plate temperature deviation according to claim 1, characterized in that, The steps for obtaining the preceding power time series tensor are as follows: The heating power values ​​of each partition corresponding to the nonlinear electrothermal compensation duty cycle are locked within the first 10 consecutive control cycles. The heating power values ​​of each partition are written into the queue storage unit one by one according to the order of the control cycles. After each writing, the original queue elements are transferred sequentially along the time direction. Then, the elements are tensorized and arranged according to the partition position and the control cycle position to generate the first power time series tensor.

5. The real-time compensation control method for hot plate temperature deviation according to claim 1, characterized in that, The steps for obtaining the discrete two-dimensional spatiotemporal convolution term are as follows: Based on the Green's function analytic terms, the partition location identifier, partition spacing parameter, time step parameter, and thermal response decay parameter are read item by item. The spatial response weights are filled in according to the spatial arrangement order corresponding to the partition location identifiers, and the time response weights are filled in according to the time arrangement order corresponding to the time step parameters, thus forming the spatial response kernel function matrix and the time response kernel function matrix. The heating power values ​​of each partition in the preceding power time series tensor are retrieved, and the power components of adjacent partitions are extracted layer by layer according to the target partition position. The power components of the historical period are extracted in the order of the control cycle. The power components of adjacent partitions are accumulated with the corresponding position values ​​of the spatial response kernel function matrix, and the power components of the historical period are accumulated with the corresponding position values ​​of the time response kernel function matrix to form a discrete two-dimensional spatiotemporal convolution term.

6. The real-time compensation control method for hot plate temperature deviation according to claim 1, characterized in that, The steps for obtaining the heat conduction feedforward offset are as follows: Read the convolution values ​​of each element in the discrete two-dimensional spatiotemporal convolution term one by one, perform element aggregation within the same partition according to the target partition position, perform cross-period element aggregation according to the control cycle order, extract the heat flow transfer response delay value, write the heat flow transfer response delay value into the target partition power deduction field, and convert it into the target partition power deduction amount according to the correspondence of the target partition power deduction field to generate the heat conduction feedforward cancellation amount.

7. The real-time compensation control method for hot plate temperature deviation according to claim 1, characterized in that, The steps for obtaining the coordinated compensation control quantity are as follows: Read the target partition power deduction value corresponding to the heat conduction feedforward cancellation amount, retrieve the duty cycle conversion reference, duty cycle resolution identifier, and duty cycle upper and lower limit identifier of the current control cycle of the target partition, map the target partition power deduction value to the duty cycle change value item by item according to the duty cycle conversion reference, perform gear alignment processing according to the duty cycle resolution identifier, and filter out the excessive change segment according to the duty cycle upper and lower limit identifier to obtain the pulse width modulation duty cycle reduction. Read the target partition duty cycle value in the nonlinear electrothermal compensation duty cycle, and map the pulse width modulation duty cycle reduction to the same target partition duty cycle value according to the target partition identifier. Extract the difference between the target partition duty cycle value and the pulse width modulation duty cycle reduction, register the remaining duty cycle value of each target partition, extract the heat flow balance correction amount, and generate the collaborative compensation control amount.

8. The real-time compensation control method for hot plate temperature deviation according to claim 1, characterized in that, The steps for obtaining the hot plate drive output signal are as follows: Read the duty cycle values ​​of each target partition in the collaborative compensation control quantity, convert the duty cycle values ​​of each target partition into the conduction period identifier, off period identifier, and cycle length identifier corresponding to the electrical signal pulse width duty cycle format, and write the conduction period identifier, off period identifier, and cycle length identifier into the output register position of the corresponding relay driver terminal according to the partition channel number to generate the hot plate drive output signal.

Images