Method for adjusting concentration of pickling in metal container in combination with mechanism model

Abstract

The present application relates to the technical field of mechanism model, and particularly relates to a metal container pickling concentration adjustment method combined with a mechanism model. The method obtains the material category of a metal container batch to be pickled, the surface area to be treated, the oxide layer grade, the initial concentration of pickling liquid, the acid supplement record, the water supplement record, the pickling duration, the metal ion accumulation amount and the free acid detection value, constructs a pickling mechanism model based on the mass conservation relationship, the oxide dissolution relationship and the liquid supplement dilution relationship, calculates a theoretical acid consumption trajectory, performs residual registration on the actual concentration trajectory and the theoretical acid consumption trajectory, inverses the acid consumption equivalent state of the pickling liquid, and generates segmented adjustment instructions according to the deviation between the acid consumption equivalent state and the target pickling end state. The present application can distinguish between the states of concentration meeting the standard but reaction capacity being insufficient, concentration being low but acid consumption equivalent still meeting the requirement and metal ion occupation being high, and changes the basis of pickling concentration adjustment from instantaneous concentration to residual reaction capacity.

Description

Technical Field

[0001] This invention relates to the field of mechanism modeling technology, and more specifically to a method for adjusting the pickling concentration of metal containers based on mechanism models. Background Technology

[0002] During the forming, welding, heat treatment, or storage of metal containers, oxide scale, rust layers, and residual deposits easily form on their surfaces. Pickling is typically used to remove these surface layers and create a metal surface that meets the requirements of subsequent processing. Current pickling concentration adjustments often combine empirical acid preparation, sampling titration, or online acidity detection. This involves calculating the acid dosage based on the initial volume of the pickling tank and the target free acid concentration, then taking free acid detection values ​​at fixed time intervals during pickling and comparing them to a preset concentration range. If the detection value is below the lower limit, acid is added; if it is above the upper limit, water is added for dilution; and if the detection value is within the range, the current reaction is maintained. Some processes also record the amount of acid added, water added, pickling duration, and batch number, and correct the tank concentration using simple mass conservation calculations. Operators usually set the acid replenishment cycle or safety margin based on historical batch concentration change curves, equating the acid consumption of the metal oxide layer to a fixed acid consumption per unit time or per unit surface area. The control object of the above scheme is actually the instantaneous average concentration of the pickling solution. The judgment is mainly based on the free acid value at the detection point and the empirical rules set manually. The actual reaction capacity of the acid solution has not formed an independent calculation state.

[0003] In more similar computer-aided mixing schemes, the material of the metal container, the number of tanks, the estimated surface area, the pickling time, and the initial concentration are input into the acid mixing program. The program then generates a recommended mixing amount based on the amount of acid required for the oxide dissolution reaction, the balance between acid and water replenishment, and the target concentration range. This type of scheme typically multiplies the surface area to be treated by an empirical acid consumption coefficient to obtain the theoretical acid consumption, and then converts the amount of acid and water already added into the concentration increment or dilution, thus obtaining the estimated concentration value for the next moment. To correct for on-site deviations, some schemes substitute the free acid detection value into the model to linearly correct the subsequent acid replenishment amount; other schemes increase the frequency of acid replenishment when the pickling time is long or the batch surface area is large, so that the estimated concentration value returns to the target range as much as possible. Although this treatment method introduces a mechanistic model, the model output still focuses on whether the concentration meets the target. The impact of metal ion accumulation on the acid reaction capacity, the changes in acid consumption trajectory due to different oxide layer loads, and the influence of replenishment mixing lag on the detection value are usually not broken down into calculable state variables. The deviation between the actual concentration trajectory and the theoretical concentration trajectory is often directly used as the correction amount for acid replenishment without distinguishing the cause of the deviation. This makes the model calculation results easily influenced by short-term detection fluctuations, uneven replenishment, or local oxide layer load deviations.

[0004] The main technical problem with the existing technology is that when using free acid concentration as the core basis for formulation, it is impossible to characterize the remaining reaction capacity of the pickling solution in the subsequent pickling process of the current batch of metal containers. At the same free acid concentration, if the surface area to be treated of the metal container, the oxide layer grade, the accumulated amount of metal ions, the history of acid and water replenishment, and the pickling duration are different, the ability of the acid solution to continue dissolving the oxide layer will not be the same. Conversely, a short-term low free acid concentration may simply be due to a lag in replenishment mixing or fluctuations in detection, and does not necessarily indicate insufficient acid reaction capacity. Existing solutions do not unify the contribution of free acid, theoretical oxide layer consumption, metal ion occupation, and dilution disturbances into an invertible mechanism. The residual between the actual concentration trajectory and the theoretical acid consumption trajectory lacks classification processing, and a direct mapping is formed between the detected value and the amount of acid replenishment, rather than the formulation action being driven by the pickling reaction capacity. Therefore, the formulation instruction is prone to continuing pickling when the concentration surface meets the standard but the acid consumption equivalent is insufficient, or prematurely replenishing acid when the acid consumption equivalent still meets the reaction requirements, resulting in a mismatch between concentration control and the actual pickling reaction demand. This mismatch can cause the same concentration range to correspond to different pickling processes in different batches, becoming the main technical root cause of insufficient concentration adjustment stability in the batch pickling of metal containers. Summary of the Invention

[0005] The purpose of this invention is to provide a method for adjusting the pickling concentration of metal containers based on a mechanistic model, which can solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method for adjusting the pickling concentration of metal containers based on a mechanistic model includes: obtaining the material type, surface area to be treated, oxide layer grade, initial concentration of pickling solution, acid replenishment record, water replenishment record, pickling duration, cumulative metal ion amount, and free acid detection value of the batch of metal containers to be pickled; constructing a pickling mechanism model including free acid contribution, theoretical oxide layer consumption, metal ion occupancy, and dilution disturbance terms according to the mass conservation relationship, oxide dissolution relationship, and replenishment dilution relationship, and calculating the theoretical acid consumption trajectory; performing residual registration between the actual concentration trajectory formed by the free acid detection value and the theoretical acid consumption trajectory to obtain a concentration residual sequence; inverting the acid consumption equivalent state of the pickling solution based on the concentration residual sequence; and generating segmented adjustment instructions based on the deviation between the acid consumption equivalent state and the target pickling endpoint state.

[0008] Preferably, the calculation of the theoretical acid consumption trajectory includes: mapping the surface area to be treated, the material type, and the oxide layer grade to an equivalent oxide layer load, and calculating the initial theoretical consumption according to the acid consumption equivalent coefficient of different oxide materials; discretizing the pickling duration into multiple calculation intervals, subtracting the oxide layer theoretical consumption term based on the free acid contribution term in each calculation interval, and updating the metal ion occupancy term based on the accumulated metal ions; converting the acid replenishment record and the water replenishment record into interval mass increments, and superimposing them on the dilution perturbation term to form the theoretical concentration value of the corresponding calculation interval.

[0009] Preferably, the residual registration includes: establishing a unified time axis according to the acid replenishment time, water replenishment time, and free acid detection time; resampling the actual concentration trajectory and the theoretical acid consumption trajectory within the same interval; segmenting and classifying the trajectory differences after resampling to form detection fluctuation residuals, oxide layer load deviation residuals, metal ion accumulation deviation residuals, and replenishment mixing hysteresis residuals; wherein, the segmentation and classification are determined based on the duration of the residuals, directional consistency, distance from the replenishment time, and synchronization relationship with the cumulative changes of metal ions.

[0010] Preferably, the inversion of the acid consumption equivalent state includes: constructing a state vector containing the remaining free acid, the occupied acid, the equivalent load of the remaining oxide layer, and the dilution disturbance; mapping the concentration residual sequence to correction components of the state vector according to the residual type, retaining the association identifier between each correction component and the corresponding residual type, and limiting the value range of the correction components by the mass conservation constraint of the acid washing mechanism model; and using the corrected state vector to calculate the acid consumption equivalent available for oxide layer dissolution within a subsequent preset calculation interval, as the acid consumption equivalent state.

[0011] Preferably, the mapping of the equivalent load of the oxide layer includes: dividing the surface area to be treated into acid consumption calculation partitions according to material type and oxide layer grade, setting a corresponding oxide composition weight for each acid consumption calculation partition; converting each acid consumption calculation partition into a standard oxide layer load according to the oxide composition weight, and including the weld heat-affected zone, corner area and ordinary surface area into different load correction terms; summarizing each standard oxide layer load and its load correction terms into a batch oxide layer equivalent load, and inputting it into the pickling mechanism model.

[0012] Preferably, the update of the metal ion occupancy item includes: determining the acid occupancy amount formed by metal salts in the acid solution based on the accumulated amount of metal ions, and storing the acid occupancy amount separately from the free acid contribution item; when the increment of the accumulated amount of metal ions in adjacent calculation intervals exceeds the theoretical increment derived from the theoretical consumption item of the oxide layer, the excess portion is marked as the non-oxide layer acid consumption increment; the non-oxide layer acid consumption increment is written into the additional occupancy sub-item of the metal ion occupancy item, and the corresponding free acid contribution is deducted in the subsequent calculation of the theoretical concentration value.

[0013] Preferably, the segmented classification includes: setting a sliding calculation window on a unified time axis, calculating the mean sign, slope direction, and duration of the residual within each sliding calculation window; when the residual appears only at a single detection moment and does not continue with adjacent moments, it is classified as detection fluctuation residual; when the residual forms a delayed peak after the replenishment moment, it is classified as replenishment mixing hysteresis residual; when the residual expands synchronously with the cumulative change of metal ions, it is classified as metal ion accumulation deviation residual; and the remaining residuals with unidirectional continuous expansion are classified as oxide layer load deviation residuals.

[0014] Preferably, the generation of the segmented dispensing instructions includes: dividing the acid consumption equivalent state into four categories: low concentration and sufficient acid consumption equivalent, qualified concentration and insufficient acid consumption equivalent, high metal ion occupation, and unstable dilution disturbance; generating a homogenization waiting instruction for the state of low concentration and sufficient acid consumption equivalent, generating a small-dose, multi-stage acid replenishment instruction for the state of qualified concentration and insufficient acid consumption equivalent, generating a drainage dilution and acid replenishment linkage instruction for the state of high metal ion occupation, and generating a replenishment pause instruction for the state of unstable dilution disturbance.

[0015] Preferably, the method further includes pre-allocation consistency verification: establishing a corresponding verification table between the residual types obtained from the segmentation and classification and the four states; before generating the segmented allocation instruction, reading the current dominant residual type and matching it with the state category corresponding to the acid consumption equivalent state; when the dominant residual type does not match the state category, re-executing residual registration and reducing the correction weight of the most recent free acid detection value in the state inversion; when two consecutive calculation intervals match, writing the corresponding segmented allocation instruction into the instruction sequence to be executed.

[0016] Preferably, the method also includes closed-loop update after adjustment: after executing the sequence of instructions to be executed, the free acid detection value, metal ion accumulation and replenishment record after execution are obtained and compared with the interval theoretical concentration value predicted by the acid washing mechanism model before execution; when the newly added residual formed by the comparison is consistent with the dominant residual type, the newly added residual is written into the historical residual set of the corresponding residual type; when the newly added residual is inconsistent with the dominant residual type, the newly added residual is written into the residual set to be confirmed and is only used as a transient correction amount for the dilution disturbance term in the state inversion of the next calculation interval.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0018] 1. By converting the pickling solution preparation basis from instantaneous free acid concentration to acid consumption equivalent state, this invention can directly address the problem that concentration control cannot characterize the remaining reaction capacity. The mechanistic model is established according to the mass conservation relationship, oxide dissolution relationship, and replenishment / dilution relationship. It incorporates the free acid contribution term, theoretical oxide layer consumption term, metal ion occupation term, and dilution disturbance term into the same calculation chain, ensuring that the amount of acid that can participate in oxide layer dissolution during pickling, the amount of acid already occupied by metal ions, the concentration disturbance caused by replenishment, and the theoretical consumption caused by oxide layer load have a common state expression. After registering the actual concentration trajectory with the theoretical acid consumption trajectory, a concentration residual sequence is generated. The residual is no longer simply converted into replenishment amount but is used to invert the acid consumption equivalent that the pickling solution can currently continue to complete oxide layer dissolution. The preparation calculation is driven by the deviation between the acid consumption equivalent state and the target pickling endpoint state, which can distinguish between states such as low concentration but still meeting the acid consumption equivalent, concentration meeting the standard but insufficient acid consumption equivalent, high metal ion occupation, and unstable dilution disturbance. For situations where the concentration is low but the acid consumption equivalent is still sufficient, the preparation process can enter a homogenization waiting phase to avoid adding acid due to short-term low detection values. For situations where the concentration meets the standard but the acid consumption equivalent is insufficient, the preparation process can enter a phase of small-dose, multi-stage acid addition to avoid delays in acid addition due to the average concentration meeting the standard. For situations where metal ion occupancy is high, the preparation process can link drainage dilution and acid addition to prevent the occupied acid from being mistakenly considered as the amount of acid that can continue to react. A closed-loop relationship is formed between the concentration detection value, the mechanism calculation value, and the state inversion value, enabling the pickling solution preparation to shift from single-point concentration correction to process control based on the remaining reaction capacity. The preparation action is jointly constrained by the theoretical acid consumption trajectory and the actual residual sequence, ensuring that concentration changes, acid consumption, and replenishment actions are calculated from the same source, which can reduce concentration overshoot, acid addition lag, and misjudgment of the pickling endpoint.

[0019] 2. Residual registration, state vector correction, and closed-loop update after adjustment further improve the computational stability of pickling concentration adjustment. A unified time axis maps acid replenishment, water replenishment, and free acid detection to the same calculation interval, avoiding direct subtraction between theoretical and actual trajectories due to different sampling sequences. Residual segmentation and classification separates detection fluctuations, oxide layer load deviations, metal ion accumulation deviations, and replenishment mixing lags, ensuring each type of deviation enters its corresponding state correction component, rather than all entering the same acid replenishment correction amount. The acid consumption equivalent state consists of the remaining free acid, occupied acid, remaining oxide layer equivalent load, and dilution disturbance, and is subject to mass conservation constraints, reducing the direct impact of short-term abnormal detection values ​​on the adjustment command. The oxide layer equivalent load forms acid consumption calculation partitions according to material type, oxide layer grade, and surface area to be treated, and incorporates the weld heat-affected zone, corner areas, and ordinary surface areas into different load correction items, ensuring a correspondence between the theoretical acid consumption trajectory and the differences on the metal container surface. The pre-mixing consistency check matches the dominant residual type with the state category. If there is a mismatch, the correction weight of the most recent free acid detection value in the state inversion is reduced to prevent unhomogenized replenishment or single-point detection deviation from directly triggering acid replenishment, water replenishment, or drainage commands. The post-mixing closed-loop update compares the executed free acid detection value, accumulated metal ion amount, and replenishment record with the predicted concentration value. New residuals are written to the historical residual set or the unconfirmed residual set based on whether they match the dominant residual type, ensuring that subsequent calculation intervals retain traceable residual sources. Therefore, subsequent state inversions can use the confirmed residual types and limit unconfirmed residuals to transient correction values, reducing the cumulative amplification of random deviations in continuous mixing. This ensures good continuity of mixing results under different batches, different oxide layer loads, and different replenishment disturbance conditions, preventing the same detection deviation from being repeatedly amplified as acid replenishment requirements in multiple calculation intervals. Attached Figure Description

[0020] Figure 1 This is an overall flowchart of the metal container pickling concentration adjustment method based on the combined mechanism model of the present invention;

[0021] Figure 2 This is a flowchart of the theoretical acid consumption trajectory calculation and oxide layer equivalent load mapping of the present invention;

[0022] Figure 3 This is a flowchart of the residual segmentation classification, acid consumption equivalent state inversion, and closed-loop update process of the present invention. Detailed Implementation

[0023] In one embodiment, a method for adjusting the pickling concentration of metal containers using a mechanistic model is provided, comprising: obtaining the material type, surface area to be treated, oxide layer grade, initial concentration of pickling solution, acid replenishment record, water replenishment record, pickling duration, cumulative metal ion amount, and free acid detection value of the batch of metal containers to be pickled; constructing a pickling mechanism model including a free acid contribution term, a theoretical oxide layer consumption term, a metal ion occupancy term, and a dilution disturbance term according to the mass conservation relationship, oxide dissolution relationship, and replenishment dilution relationship, and calculating the theoretical acid consumption trajectory; performing residual registration between the actual concentration trajectory formed by the free acid detection value and the theoretical acid consumption trajectory to obtain a concentration residual sequence; inverting the acid consumption equivalent state of the pickling solution based on the concentration residual sequence; generating a segmented adjustment instruction based on the deviation between the acid consumption equivalent state and the target pickling endpoint state. In this embodiment, the above data is used as the state input of the same pickling batch, and a single free acid detection value is not directly equated with the adjustment basis, but rather the surface area of ​​the metal container is considered as the state input. The mechanistic model incorporates the surface load to be treated, the historical acid replenishment process, the reaction time progression, and the accumulation of metal ions in the acid, enabling a calculable correspondence between the instantaneous concentration of the pickling solution, the theoretical acid consumption trajectory, and the remaining reaction capacity. The material category is used to determine the acid consumption properties of the metal oxide, the surface area to be treated is used to determine the reaction contact load, the oxide layer level is used to determine the acid consumption requirement per unit area, the acid replenishment record and the water replenishment record are used to reduce the mass change of the acid system, the pickling duration is used to divide the reaction interval, the accumulated metal ions are used to characterize the influence of metal salts that have already occupied the acid reaction capacity, and the free acid detection value is used to form the actual concentration trajectory and correct the model calculation deviation. The advantage of this embodiment is that it transforms the pickling concentration adjustment from a single concentration judgment to a closed-loop calculation based on the mechanistic model and the acid consumption equivalent state, avoiding incorrect adjustment when the concentration surface meets the standard but the reaction capacity is insufficient, or the concentration is low for a short period of time but the reaction capacity is still sufficient.

[0024] refer to Figure 1In this embodiment, the pickling mechanism model uses discrete time intervals as the calculation unit. The pickling duration is divided into several adjacent calculation intervals. Each calculation interval has a start time, end time, acid replenishment amount, water replenishment amount, detection concentration, and accumulated metal ions. After reading the process records of the same batch of metal containers, the system converts the material category to a material code, the oxide layer level to an equivalent oxide layer load coefficient, the surface area to be treated to a reaction area input, and the initial concentration of the pickling solution to an initial free acid amount. The acid replenishment and water replenishment records are assigned to the corresponding calculation intervals according to their occurrence time. The mass conservation relationship is used to calculate the total mass change of the acid solution and the oxide dissolution. The relationship is used to calculate the amount of acid required for oxide layer consumption per unit time. The replenishment and dilution relationship is used to calculate the effect of replenishment of acid and water on the free acid concentration. The free acid contribution term represents the amount of free acid that can participate in the pickling reaction within the calculation interval. The theoretical consumption term of oxide layer represents the theoretical acid consumption calculated according to material type, surface area and oxide layer grade. The metal ion occupancy term represents the amount of metal salt that has been formed in the acid solution and reduces the subsequent reactive capacity. The dilution disturbance term represents the transient effect on the detection concentration when replenishment of water and acid is not completely homogenized. The advantage of this embodiment is that it carries data from different sources in the same calculation interval, so that the detection value, replenishment record and mechanism consumption are calculated under the same time reference.

[0025] In this embodiment, the theoretical acid consumption trajectory can be calculated using the following formula:

[0026]

[0027] in, This represents the theoretical free acid concentration at the end of the k-th calculation interval. This indicates the initial concentration of the pickling solution. This indicates the initial mass of the pickling solution. This represents the amount of acid replenishment in the i-th calculation interval. This represents the concentration of acid added in the i-th calculation interval. This represents the water replenishment mass for the i-th calculation interval. This represents the theoretical acid consumption of the oxide layer in the i-th calculation interval. This represents the amount of acid occupied by metal ions in the i-th calculation interval. This represents the dilution perturbation concentration correction term for the k-th calculation interval. The numerator represents the theoretical total amount of free acid after considering acid replenishment, oxide layer consumption, and metal ion occupation, and the denominator represents the total mass of the pickling solution after considering acid and water replenishment. For example, the initial concentration of a certain batch... initial mass The quality of acid replenishment in the first interval Add acid solution concentration Hydration quality Theoretically, the oxide layer consumes a certain amount of acid. Metal ions occupy acid amount Dilution perturbation concentration correction term ,but The calculation result is used as the theoretical concentration trajectory point for the first interval.

[0028] refer to Figure 2 Furthermore, the calculation of the theoretical acid consumption trajectory includes: mapping the surface area to be treated, the material type, and the oxide layer grade to an equivalent oxide layer load, and calculating the initial theoretical consumption according to the acid consumption equivalent coefficient of different material oxides; discretizing the pickling duration into multiple calculation intervals, subtracting the oxide layer theoretical consumption term based on the free acid contribution term in each calculation interval, and updating the metal ion occupancy term based on the accumulated metal ion amount; converting the acid replenishment record and the water replenishment record into interval mass increments, superimposing them on the dilution perturbation term to form the theoretical concentration value of the corresponding calculation interval. Specifically, the system establishes a batch data table for each batch, and the fields of the batch data table include at least the batch number, material type, surface area to be treated, oxide layer grade, initial concentration of pickling solution, acid... The initial mass of the washing solution, the mass of acid replenishment at each time, the concentration of acid added at each time, the mass of water replenishment at each time, the cumulative amount of metal ions at each time, and the free acid detection value at each time are mapped. The acid consumption equivalent coefficient is determined by the material type, the surface load coefficient is determined by the oxide layer level, and the total reaction area is determined by the surface area to be treated. The three are multiplied to form the basic acid consumption input for the interval. The length of the calculated interval can be determined based on the sampling period, the replenishment event, or the record point of the acid washing process. When the acid replenishment event or water replenishment event occurs between two detection times, the system uses the time of the event as the new interval boundary to ensure that the replenishment record is not averaged into the wrong time period. The advantage of this embodiment is that the theoretical concentration value is no longer derived from a fixed empirical acid replenishment amount, but is determined by the surface load, reaction consumption, ion occupation, and replenishment dilution.

[0029] In one embodiment, the mapping of the equivalent oxide load includes: dividing the surface area to be treated into acid consumption calculation partitions according to material type and oxide layer grade; setting corresponding oxide composition weights for each acid consumption calculation partition; converting each acid consumption calculation partition into a standard oxide layer load based on the oxide composition weights; and including the weld heat-affected zone, corner area, and ordinary surface area into different load correction terms; summarizing each standard oxide layer load and its load correction terms into a batch oxide layer equivalent load, and inputting it into the pickling mechanism model. In specific implementation, the system does not change the structure of the metal container, but only performs data partitioning based on the surface area information and material information available in the process record. The acid consumption calculation partitions can be based on the same material. Furthermore, when the same oxide layer grade is formed, and the same batch of containers contains weld heat-affected zones, corner areas, and ordinary surface areas, the system records them into different surface area categories respectively, so that different surface states within the same batch will not be averaged into a single acid consumption load. The oxide composition weight is used to represent the relative contribution of different oxide components to acid consumption. The standard oxide layer load is used to convert different materials and different oxide layer grades into a unified calculation scale. The load correction term is used to compensate for the density of the oxide layer in the weld heat-affected zone, the difference in acid contact in the corner areas, and the difference in the uniformity of reaction on the ordinary surface. The advantage of this embodiment is that it enables the subsequent theoretical acid consumption trajectory to correspond to the differences in the surface of the metal container, reducing the trajectory deviation caused by the uniform unit area acid consumption coefficient.

[0030] Table 1 is an equivalent load mapping table for oxide layers, which illustrates the correspondence between material type, oxide layer grade, surface area, and model input quantity.

[0031] Table 1. Equivalent Load Mapping of Oxide Layer

[0032] Material category Distinguishing the acid consumption properties of metal oxides Acid consumption equivalent coefficient Batch records are converted into material codes and then used in calculations. Surface area to be processed Characterizing reaction contact load Acid consumption calculation zone area Together with material type and oxide layer grade, they form a zoning system. Oxidation layer level Characterizing the load to be removed per unit area Surface load factor Convert to standard oxide layer load Weld heat-affected zone Characterizing the density differences of local oxide layers Load correction item Separately incorporated into the equivalent load of the batch oxide layer Corner area Characterizing local reaction contact differences Load correction item Calculated separately from ordinary surface areas ordinary surface area Characterizing the reactive load of the reference surface Baseline zone load As a component of batch load aggregation

[0033] In this embodiment, the equivalent load of the oxide layer can be calculated using the following formula:

[0034]

[0035] in, Indicates the equivalent load of the batch oxide layer. This indicates the total number of zones for calculating acid consumption. This represents the surface area to be processed in the r-th acid consumption calculation partition. This represents the acid consumption equivalent coefficient for the r-th acid consumption calculation zone, determined by material type. This represents the surface load coefficient of the r-th acid consumption calculation zone, determined according to the oxide layer level. This represents the load correction term for the r-th acid consumption calculation zone, determined by the weld heat-affected zone, corner area, or ordinary surface area; for example, a batch is divided into 2 acid consumption calculation zones. , , , , , , , ,but This result is included in the calculation of the theoretical consumption term of the oxide layer.

[0036] Furthermore, the updating of the metal ion occupancy item includes: determining the acid occupancy amount formed by metal salts in the acid solution based on the accumulated metal ions, and storing the acid occupancy amount separately from the free acid contribution item; when the increment of the accumulated metal ions in adjacent calculation intervals exceeds the theoretical increment derived from the theoretical consumption item of the oxide layer, the excess portion is marked as the non-oxide layer acid consumption increment; the non-oxide layer acid consumption increment is written into the additional occupancy sub-item of the metal ion occupancy item, and the corresponding free acid contribution is deducted in subsequent theoretical concentration value calculations. Specifically, the system uses the accumulated metal ions as characterizing data that the acid solution reaction capacity has been occupied, and does not mix it into the free acid contribution item; the accumulated metal ions come from the same batch. Acid analysis records or process records are converted into acid occupancy within each calculation interval. When the cumulative increase of metal ions in adjacent calculation intervals is consistent with the increase of metal dissolution corresponding to the theoretical consumption of the oxide layer, the metal ion occupancy item is updated according to the theoretical occupancy. When the cumulative increase of metal ions is higher than the theoretical increase, the system determines that there is additional acid consumption from sources other than the oxide layer and writes it into the additional occupancy sub-item. When calculating the subsequent theoretical concentration value, the corresponding amount is deducted from the contribution of free acid, so that the state of high metal ion occupancy but compliant concentration can be expressed separately. The advantage of this embodiment is that it separates the amount of acid in the acid solution that has been occupied by metal salts from the amount of free acid that can continue to react, and avoids the influence of metal ion accumulation being masked in the average concentration judgment.

[0037] refer to Figure 3In one embodiment, the residual registration includes: establishing a unified time axis based on the acid replenishment time, water replenishment time, and free acid detection time; resampling the actual concentration trajectory and the theoretical acid consumption trajectory within the same interval; and segmenting and classifying the trajectory differences after resampling to form detection fluctuation residuals, oxide layer load deviation residuals, metal ion accumulation deviation residuals, and replenishment mixing hysteresis residuals. The segmentation and classification are determined based on the duration of the residuals, directional consistency, distance from the replenishment time, and synchronization with the cumulative changes in metal ions. Specifically, the system does not directly compare any adjacent detection points with theoretical points, but instead constructs a system based on acid replenishment events, water replenishment events, and detection events. A unified time axis is used so that each node on the same time axis represents a state boundary. The actual concentration trajectory forms interval concentration points between adjacent boundaries through the free acid detection value. The theoretical acid consumption trajectory forms theoretical concentration points on the same boundary through the mechanism model. During resampling, for boundary points without detection values, the system can use linear interpolation of adjacent detection times to form actual trajectory points. For boundary points without replenishment events, the theoretical trajectory is continuously advanced according to reaction consumption. The trajectory difference is formed by subtracting the theoretical trajectory point from the actual trajectory point and is classified into different categories according to the continuous performance of the residual and the correlation of events. The advantage of this embodiment is that it avoids false residuals caused by different sampling times from entering the state inversion.

[0038] In this embodiment, the concentration residual sequence can be generated according to the following formula:

[0039]

[0040] in, This represents the concentration residual at the k-th unified time axis node. This represents the actual free acid concentration at the k-th unified time axis node. This represents the theoretical free acid concentration at the k-th unified time axis node; for example, the actual free acid concentration at the 3rd unified time axis node. Theoretical free acid concentration ,but A negative value indicates that the actual concentration is lower than the theoretical concentration. If the negative value persists in multiple subsequent nodes and corresponds to the record of high oxide layer load, it will be entered into the oxide layer load deviation residual. If the negative value only appears in a short period of time after water replenishment and then returns, it will be entered into the liquid replenishment mixing hysteresis residual.

[0041] Preferably, the segmentation and classification includes: setting a sliding calculation window on a unified time axis, calculating the mean sign, slope direction, and duration of the residuals within each sliding calculation window; when the residual appears only at a single detection moment and does not continue with adjacent moments, it is classified as a detection fluctuation residual; when the residual forms a delayed peak after the replenishment moment, it is classified as a replenishment mixing hysteresis residual; when the residual expands synchronously with the cumulative change of metal ions, it is classified as a metal ion accumulation deviation residual; and the remaining residuals with unidirectional continuous expansion are classified as oxide layer load deviation residuals. In specific implementation, the sliding calculation window covers at least several adjacent unified time axis nodes, and the system calculates the sign of the mean residual within the window. The direction of adjacent residual differences and the length of continuous residual existence are considered. Detection fluctuation residuals usually exhibit single-point abrupt changes and are not continuous between nodes. Liquid replenishment mixing hysteresis residuals usually appear after acid or water replenishment events and gradually decrease during subsequent homogenization. Metal ion accumulation deviation residuals usually expand synchronously with the increase of metal ion accumulation. Oxide layer load deviation residuals usually change continuously in a single direction when there is no obvious liquid replenishment event. The system writes the classification results into the residual type identifier and stores them together with the residual values. The advantage of this embodiment is that it allows deviations from different sources to enter different correction paths, avoiding the uniform treatment of detection fluctuations, mixing hysteresis, and actual acid consumption differences as acid replenishment requirements.

[0042] Table 2 is a residual segmentation classification table, used to illustrate the correspondence between residual shape and state correction direction.

[0043] Table 2 Residual Segmentation Classification Table

[0044] Detecting fluctuation residuals Residual at adjacent detection points Single point appears and does not continue Reduce the correction weight of the corresponding detection point Oxide layer load deviation residual Residual duration and direction One-way continuous expansion Correction of the equivalent load of the remaining oxide layer Metal ion cumulative deviation residual cumulative changes in metal ions Expanding in tandem with the cumulative amount Correcting the amount of acid already used Hysteresis Residue of Liquid Replenishment Distance from the time of fluid resuscitation Post-refill delayed peak Correcting dilution disturbance

[0045] In one embodiment, the inversion of the acid consumption equivalent state includes: constructing a state vector containing the remaining free acid amount, the occupied acid amount, the equivalent load of the remaining oxide layer, and the dilution perturbation amount; mapping the concentration residual sequence to correction components of the state vector according to the residual type, retaining the association identifier between each correction component and the corresponding residual type, and limiting the value range of the correction components by the mass conservation constraint of the pickling mechanism model; using the corrected state vector to calculate the acid consumption equivalent available for oxide layer dissolution within a subsequent preset calculation interval, as the acid consumption equivalent state. Specifically, the remaining free acid amount in the state vector comes from the joint calculation of the initial free acid amount, the added acid amount, the theoretical consumption of the oxide layer, and the metal ion occupation amount. The occupied acid amount comes from the accumulated metal ions and its additional occupied sub-items. The remaining equivalent oxide layer load comes from the batch oxide layer equivalent load minus the completed reaction load. The dilution disturbance comes from the transient correction after the acid and water replenishment event. There is a fixed mapping relationship between the residual type and the state component. The detection fluctuation residual does not directly change the main component of the state vector. The oxide layer load deviation residual corrects the remaining equivalent oxide layer load. The metal ion accumulation deviation residual corrects the occupied acid amount. The replenishment mixing hysteresis residual corrects the dilution disturbance. The mass conservation constraint is used to limit the relationship between the remaining free acid amount, occupied acid amount and replenishment mass after correction. The advantage of this embodiment is that the residual correction does not deviate from the acid washing reaction and mass conservation boundary.

[0046] In this embodiment, the acid consumption equivalent state can be calculated using the following formula:

[0047]

[0048] in, This represents the acid consumption equivalent state at the end of the k-th calculation interval. This indicates the amount of remaining free acid after correction. This indicates the corrected amount of acid already used. This indicates the equivalent load of the remaining oxide layer after correction. This represents the corrected amount of dilution perturbation. This represents the acid consumption conversion factor corresponding to the equivalent load per unit of remaining oxide layer. This represents the acid consumption equivalent conversion factor corresponding to the dilution disturbance; for example, the endpoint of the 4th calculation interval. , , , , , ,but This value represents the acid consumption equivalent that can still be used for subsequent oxide layer dissolution after taking into account the occupancy, remaining load, and disturbance conversion.

[0049] Furthermore, the generation of the segmented dispensing instructions includes: dividing the acid consumption equivalent status into four categories: low concentration and sufficient acid consumption equivalent, qualified concentration but insufficient acid consumption equivalent, high metal ion occupancy, and unstable dilution disturbance; generating a homogenization waiting instruction for the low concentration and sufficient acid consumption equivalent status, generating a small-dose, multi-stage acid replenishment instruction for the qualified concentration but insufficient acid consumption equivalent status, generating a drainage dilution and acid replenishment linkage instruction for the high metal ion occupancy status, and generating a replenishment pause instruction for the unstable dilution disturbance status. Specifically, the system simultaneously reads the current free acid detection value and acid consumption equivalent status. When the free acid detection value is lower than the target range but... When the concentration is still sufficient to cover the equivalent load of the remaining oxide layer, the system will classify the state as low concentration and sufficient acid consumption equivalent and generate a homogenization waiting instruction. When the free acid detection value is within the target range but When the concentration is below the target pickling endpoint, the system classifies the state as meeting the standard but insufficient acid consumption equivalent and generates a small-dose, multi-stage acid replenishment command. When the occupied acid amount is relatively high compared to the remaining free acid amount, the system classifies the state as high metal ion occupation and generates a draining dilution and acid replenishment linkage command. When the dilution disturbance amount has not yet returned to a stable state, the system classifies the state as unstable dilution disturbance and generates a replenishment pause command. The advantage of this embodiment is that it maps different reaction capacity states under the same concentration detection value to different mixing actions.

[0050] In a preferred embodiment, the segmented dispensing instructions are not directly written to the execution end, but are first written to the sequence of instructions to be executed. The sequence of instructions to be executed stores the state category, dominant residual type, acid consumption equivalent state, target acid washing endpoint state, type of action to be executed, and generation time. The homogenization waiting instruction includes the start time and end judgment condition for waiting for homogenization. The small-dose, multi-stage acid replenishment instruction includes the relative order of each acid replenishment and the detection and verification conditions after each acid replenishment. The drainage, dilution, and acid replenishment linkage instruction includes the sequential relationship between drainage, dilution, and acid replenishment. The replenishment pause instruction includes the free acid detection value and metal ion accumulation that need to be collected during the pause. When the system forms the sequence of instructions to be executed, it does not use a single acid replenishment amount as the only output, but binds and stores the dispensing action with the state category that generated the action, so that subsequent closed-loop updates can trace the judgment basis before dispensing. The advantage of this embodiment is that the dispensing action has a traceable state source, which facilitates subsequent interval judgment on whether the new residual is consistent with the original dominant residual.

[0051] In one embodiment, the method further includes a consistency check before allocation: establishing a corresponding check table between the residual types obtained from the segmented classification and the four states; before generating the segmented allocation instruction, reading the current dominant residual type and matching it with the state category corresponding to the acid consumption equivalent state; when the dominant residual type does not match the state category, re-executing residual registration and reducing the correction weight of the most recent free acid detection value in the state inversion; when two consecutive calculation intervals match, writing the corresponding segmented allocation instruction into the instruction sequence to be executed. Specifically, the corresponding check table associates the oxide layer load deviation residual with the state where the concentration meets the standard and the acid consumption equivalent is insufficient or the concentration is too low and the acid consumption equivalent is insufficient, and associates the metal ion cumulative deviation residual with the state where the metal ion occupancy is too high. The system correlates residuals of replenishment mixing with unstable states of dilution disturbance, and correlates residuals of detection fluctuations with states where the correction weight of detection points needs to be reduced. The dominant residual type can be determined by the residual amplitude, duration, and number of classifications within the most recent calculation intervals. If the dominant residual type is residuals of replenishment mixing and the state category is concentration met but acid consumption equivalent is insufficient, the system re-executes residual registration and reduces the correction weight of the most recent free acid detection value to avoid the unhomogenized detection value directly triggering acid replenishment. Only when the dominant residual type matches the state category within two consecutive calculation intervals will the execution instruction sequence receive the corresponding allocation instruction. The advantage of this embodiment is that it filters out unstable judgments by matching the residual source and the state category.

[0052] In this embodiment, the consistency check before allocation can be performed by calculating the matching flag using the following formula:

[0053]

[0054] in, This represents the matching flag for the k-th computation interval. This indicates the dominant residual type for the k-th computation interval. This represents the acid consumption equivalent state category for the k-th calculation interval. Representation and state category The corresponding set of allowed residual types; for example, if To dilute the unstable state of the disturbance, Includes the residual after hysteresis of the replenishment mixture, and To compensate for the hysteresis residual of the liquid mixing, then ,like If the residual is the oxide layer load deviation, then ,when Residual registration is re-executed and the correction weight of the most recent free acid detection value is reduced.

[0055] Furthermore, it also includes closed-loop updates after adjustment: after executing the sequence of instructions to be executed, the free acid detection value, metal ion accumulation, and replenishment record after execution are obtained and compared with the interval theoretical concentration value predicted by the pickling mechanism model before execution; when the newly added residual formed by the comparison is consistent with the dominant residual type, the newly added residual is written into the historical residual set of the corresponding residual type; when the newly added residual is inconsistent with the dominant residual type, the newly added residual is written into the unconfirmed residual set and is only used as a transient correction amount for the dilution disturbance term in the state inversion of the next calculation interval. In specific implementation, the system establishes a post-execution data packet after the sequence of instructions to be executed. The post-execution data packet includes the execution time, execution action category, free acid detection value after execution, metal ion accumulation after execution, and newly added replenishment record during execution. The pickling mechanism model has been prepared before the instruction execution. The system generates a theoretical concentration value for the predicted interval. After execution, the actual free acid detection value is subtracted from the theoretical concentration value to form a new residual. The type of the new residual is determined according to the direction, duration, distance from the replenishment time, and synchronization with the cumulative change of metal ions. If the type of the new residual is consistent with the dominant residual type before execution, it indicates that the original residual source continues after execution. The system writes it into the historical residual set and uses it as a confirmed correction source in subsequent inversions of similar states. If the type of the new residual is inconsistent with the dominant residual type before execution, the system writes it into the unconfirmed residual set. The next calculation interval only allows it to enter the transient correction amount of the dilution disturbance term, avoiding accidental fluctuations after one execution from changing the equivalent load of the remaining oxide layer or the amount of acid already occupied. The advantage of this embodiment is that the new deviation after mixing can be absorbed in stages, reducing the accumulation of occasional deviations in continuous mixing.

[0056] In a preferred embodiment, the historical residual set is stored separately according to residual type. The detection fluctuation residual set stores the detection time, residual value, and recovery status of adjacent nodes. The oxide layer load deviation residual set stores the duration interval length, directional consistency, and corresponding acid consumption calculation partition. The metal ion cumulative deviation residual set stores the cumulative change in metal ions, the corresponding occupied acid amount, and additional occupied sub-items. The replenishment mixing hysteresis residual set stores the replenishment time, delayed peak position, and fallback interval. The unconfirmed residual set stores residual records that have not yet met the continuous matching condition. Before each state inversion, the system reads the records in the historical residual set that are the same type as the current dominant residual and forms state correction components according to the continuity of the same type of residual. For the unconfirmed residual set, it is only used as a transient input of dilution disturbance in the next calculation interval. If the unconfirmed residual appears continuously with the same type of residual in subsequent intervals, it is converted into the corresponding historical residual set. The advantage of this embodiment is that it stores confirmed deviations and unconfirmed deviations separately, so that subsequent calculation intervals can inherit the real source of deviations without solidifying single abnormal detection results into long-term mechanism parameters.

[0057] In this embodiment, the data processing flow for pickling concentration adjustment consists of batch input, mechanism calculation, trajectory registration, state inversion, segmented adjustment, consistency verification, and closed-loop update. Each processing step is connected via a unified time axis and state vector. The batch input step reads the material type, surface area to be treated, oxide layer grade, initial concentration, acid replenishment record, water replenishment record, pickling duration, accumulated metal ions, and free acid detection value. The mechanism calculation step outputs the theoretical acid consumption trajectory. The trajectory registration step outputs the concentration residual sequence and residual type. The state inversion step outputs the acid consumption equivalent state. The segmented mixing process outputs corresponding mixing instructions, the consistency verification process determines whether the mixing instructions enter the sequence of instructions to be executed, and the closed-loop update process writes the newly added residuals formed after execution into the historical residual set or the residual set to be confirmed. The data in the entire process is based on the remaining reaction capacity of the acid solution. It does not use data sources unrelated to the pickling of metal containers, nor does it introduce control objects that are not directly related to the mixing of pickling concentration. The advantage of this embodiment is that the data flow is consistent with the core technical issues, and it can cover the impact of batch differences, replenishment disturbances, ion occupancy and detection deviations on the mixing process.

[0058] Specifically, when a batch of metal containers enters the pickling process, the system reads the batch material category as the target material category, reads the surface area to be treated and forms acid consumption calculation partitions according to the weld heat-affected zone, corner area, and ordinary surface area, reads the oxide layer level and converts each calculation partition into a standard oxide layer load, and then inputs the initial concentration, initial mass, acid replenishment record, water replenishment record, and pickling duration into the mechanism model. The model calculates the theoretical acid consumption trajectory segment by segment according to the calculation interval. If the free acid concentration is detected to be temporarily lower than the theoretical value in the first interval, but the water replenishment event has just occurred and the metal ion accumulation has not increased synchronously, the residual is attributed to the replenishment. The liquid mixing hysteresis residual, the dilution disturbance in the state vector is transiently corrected, the segmented dispensing step generates a liquid replenishment pause command instead of an acid replenishment command. If the residual in the subsequent interval falls back, the new residual is written into the set of residuals to be confirmed and will no longer affect the equivalent load of the remaining oxide layer in the next calculation interval. If the residual after the second interval is continuously negative and matches the oxide layer load record, the system classifies the residual into the oxide layer load deviation residual and corrects the equivalent load of the remaining oxide layer. The acid consumption equivalent state is reduced accordingly. The segmented dispensing step generates a small dose of acid replenishment command in stages. The advantage of this embodiment is that the same low concentration phenomenon can enter different dispensing paths due to different residual sources.

[0059] Preferably, during continuous batch pickling, the system does not directly use the acid replenishment amount of the previous batch as the acid replenishment amount of the next batch. Instead, it retains the confirmed residual type, residual duration, and state correction component of the previous batch. Only when the next batch has the same material type, similar oxide layer grade, and the same surface area composition will the corresponding historical residual set be used as the initial correction reference. If the initial detection trajectory of the next batch is inconsistent with the direction of the reference residual, the system will write the reference residual into the set of residuals to be confirmed and reduce its weight in state inversion. If the next batch maintains the same type of matching with the reference residual in two consecutive calculation intervals, the system will restore it as a usable record in the historical residual set. This processing method ensures that historical data only participates in the current calculation in the form of mechanistic residuals, rather than replacing the current batch inversion with a fixed formula. The advantage of this embodiment is that it avoids the distortion of the acid consumption equivalent state of the new batch caused by directly applying batch experience.

[0060] In a preferred embodiment, the pickling mechanism model, residual registration, and state inversion can be implemented by the same calculation program, or by multiple software processes respectively, with intermediate results passed through a batch data table. The inputs received by the calculation program are process record data and pickling solution detection data, and the outputs are theoretical concentration values, concentration residual sequences, residual types, acid consumption equivalent status, and sequences of instructions to be executed. The calculation program establishes independent data records for each batch in storage to avoid mixing replenishment records and detection records between different batches. If there are multiple pickling tanks or multiple pickling processes in the same batch, the system establishes an independent and unified time axis according to the pickling process identifier, and only summarizes the acid consumption equivalent status of the same metal container batch when summarizing batches, without directly averaging the detection values ​​of different pickling processes. The advantage of this embodiment is that it ensures that the software data structure is consistent with the boundary of the mechanism calculation, and is suitable for stable implementation in a computer program.

[0061] Furthermore, to ensure sufficient disclosure, this embodiment sets calculation boundaries for the correction components in the state inversion. The correction of the remaining free acid amount must not cause the total acid amount of the acid solution to exceed the sum of the initial free acid amount and the added acid amount. The correction of the occupied acid amount must not be less than the occupied amount already converted from the accumulated metal ions. The correction of the equivalent load of the remaining oxide layer must not be less than the remaining load obtained by subtracting from the interval after the acid washing reaction has been completed. The correction of the dilution disturbance amount gradually decreases with the advancement of the unified time axis after the replenishment event. If there is no replenishment event to support it, it cannot be retained as the dominant correction source for a long time. The system calculates and judges the above boundaries before outputting the acid consumption equivalent state. If a certain correction component exceeds the boundary, the corresponding residual type is rewritten as the residual to be confirmed and the residual registration is re-executed. The advantage of this embodiment is that it makes the inversion process run within the mass conservation and acid washing reaction boundaries, avoiding the state vector from being unimplementable due to abnormal detection data.

[0062] In this embodiment, the target pickling endpoint state is not a single concentration value, but is jointly expressed by the near completion of the remaining equivalent load of the oxide layer, the acid consumption equivalent state meeting the requirements of the subsequent calculation interval, the contribution of the unmasked free acid in the occupied acid volume, and the stable state of the dilution disturbance. When generating segmented allocation instructions, the system calculates the deviation between the current acid consumption equivalent state and the target pickling endpoint state. If the deviation mainly comes from the remaining equivalent load of the oxide layer, a segmented acid replenishment instruction is generated first. If the deviation mainly comes from the occupied acid volume, a draining dilution and acid replenishment linkage instruction is generated first. If the deviation mainly comes from the dilution disturbance, a replenishment pause or homogenization waiting instruction is generated first. The deviation calculation does not rely on human experience judgment, but is determined by the numerical relationship of each component of the state vector and the residual type identifier. The advantage of this embodiment is that the endpoint control and process allocation use the same acid consumption equivalent state, avoiding the use of different criteria for endpoint judgment and replenishment allocation.

[0063] Preferably, when the acid replenishment command is generated in stages, the system divides the acid consumption equivalent difference of the target pickling endpoint state into several acid replenishment actions. After each acid replenishment action is executed, the free acid detection value, metal ion accumulation and replenishment record must be re-acquired, and a theoretical concentration value prediction comparison must be completed. If the new residual is consistent with the replenishment mixing lag residual, the subsequent acid replenishment action is suspended and enters the homogenization waiting phase. If the new residual is consistent with the oxide layer load deviation residual, the subsequent acid replenishment action is retained and the remaining oxide layer equivalent load is updated. If the new residual is consistent with the metal ion accumulation deviation residual, the simple acid replenishment path is stopped and changed to the drainage dilution and acid replenishment linkage path. The advantage of this embodiment is that the staged acid replenishment is not a fixed segmentation of acid replenishment amount, but is bound to the residual classification and state inversion after execution, reducing continuous overshoot of acid replenishment.

[0064] In a preferred embodiment, the draining, dilution, and acid replenishment linkage command is generated when the metal ion occupancy is high. Before generating the linkage command, the system reads the occupied acid amount, the remaining free acid amount, and the metal ion cumulative deviation residual. When the occupied acid amount continuously deducts the contribution of free acid and the new residual increases synchronously with the metal ion accumulation, the system decomposes the mixing action into sequential actions of draining, dilution, and acid replenishment. After each sequential action, the theoretical concentration value and acid consumption equivalent status are updated. If the metal ion accumulation decreases after draining but the free acid concentration also decreases, the system generates the subsequent acid replenishment action based on the difference in acid consumption equivalent status. If the residual of the mixing lag after dilution increases, the system pauses acid replenishment and waits for the dilution disturbance to fall back. The advantage of this embodiment is that the metal ion occupancy problem is not masked by simple acid replenishment, and the occupied acid amount is not mistakenly counted as the available acid amount.

[0065] Furthermore, both the homogenization waiting instruction and the replenishment pause instruction retain their state sources in the software implementation. The homogenization waiting instruction corresponds to a state where the concentration is low and the acid consumption equivalent is met, while the replenishment pause instruction corresponds to a state where the dilution disturbance is unstable. During the waiting or pause period, the system continues to receive the free acid detection value and the accumulated amount of metal ions. If the concentration residual drops during the waiting period and the acid consumption equivalent remains met, no new acid replenishment action will be added to the instruction sequence to be executed. If the concentration residual does not drop during the waiting period and the oxide layer load deviation residual forms a continuous match, the system ends the homogenization waiting and regenerates the phased acid replenishment instruction. If the residual changes from the replenishment mixing lag residual to the metal ion accumulation deviation residual during the replenishment pause period, the system re-identifies the state where the metal ion occupancy is high and generates a linkage instruction. The advantage of this embodiment is that both the waiting and pause have state verification logic and are not fixed time delays.

[0066] In this embodiment, the method for adjusting the pickling concentration of metal containers based on the mechanistic model can be written into a computer executable program. When the program runs, it processes data in the order of batch data reading, model calculation, residual classification, state inversion, and instruction output. The program internally stores a material category mapping table, an oxide layer level mapping table, a surface area correction table, a residual type corresponding verification table, and a state category corresponding instruction table. The material category mapping table is used to determine the acid consumption equivalent coefficient, the oxide layer level mapping table is used to determine the surface load coefficient, the surface area correction table is used to determine the load correction item, the residual type corresponding verification table is used for consistency verification before adjustment, and the state category corresponding instruction table is used to generate instructions such as homogenization waiting, small-dose acid replenishment, drainage dilution and acid replenishment linkage, and replenishment pause. The program attaches a unified time axis node number to each batch of output data, so that subsequent queries can trace the correspondence between theoretical concentration value, actual concentration value, residual type, and acid consumption equivalent state. The advantage of this embodiment is that it is easy to reproduce the complete adjustment logic in a computer data processing system.

[0067] Specifically, when the system reaches a certain unified time axis node, the program first reads all acid replenishment and water replenishment records before the current node, calculates the theoretical free acid concentration according to the formula, and then reads the actual free acid detection value of the current node to form a residual. Subsequently, the program determines the residual type according to the sliding calculation window. If the residual type corresponds to oxide layer load deviation, the program takes the remaining oxide layer equivalent load as the main correction object. If the residual type corresponds to metal ion accumulation deviation, the program takes the occupied acid amount as the main correction object. If the residual type corresponds to replenishment mixing lag, the program takes the dilution disturbance amount as the main correction object. If the residual type corresponds to detection fluctuation, the program reduces the correction weight of the current detection value in the inversion and waits for the next node verification. After the state vector is updated, the acid consumption equivalent state is calculated and candidate allocation instructions are generated. The candidate allocation instructions enter the consistency verification. After the matching is passed, they are written into the instruction sequence to be executed. After execution, the detection data is obtained again and the closed-loop update is completed. The advantage of this embodiment is that each program action has a clear input, calculation object and output result.

[0068] In this embodiment, if there are missing detections or delayed entry of replenishment records, the system does not directly terminate the dispensing calculation. Instead, it sets a data integrity flag on a unified time axis. Missing free acid detection values ​​do not participate in residual calculations. The theoretical acid consumption trajectory is only advanced by the mechanistic model. Delayed entry of replenishment records or water replenishment records is backfilled into the corresponding calculation interval according to the actual occurrence time, and the theoretical acid consumption trajectory after that interval is recalculated. After recalculation, the system regenerates the concentration residual sequence and acid consumption equivalent state. If a sequence of instructions to be executed has been formed before recalculation but has not yet been executed, the recalculated result replaces the sequence of instructions to be executed. If the instructions have been executed before recalculation, the data after execution is retained and processed as a new source of residuals in the closed-loop update. The advantage of this embodiment is that missing data and delayed entry will not disrupt the mass conservation calculation chain.

[0069] Preferably, the system performs batch consistency checks on all input data. Material type, surface area to be treated, and oxide layer grade belong to batch fixed data, pickling solution initial concentration belongs to batch initial data, and acid replenishment record, water replenishment record, free acid detection value, metal ion accumulation, and pickling duration belong to process data. If the timestamp of the process data is earlier than the batch initial time or later than the pickling end time, it is not written into the unified time axis. If there are multiple free acid detection values ​​at the same time, the system treats them as multi-value inputs of the same node and first determines whether they constitute detection fluctuation residuals. If there are acid replenishment and water replenishment records at the same time, the system generates adjacent sub-nodes according to the record order and calculates the dilution disturbance amount respectively. The advantage of this embodiment is that it prevents the theoretical acid consumption trajectory and the actual concentration trajectory from being mismatched due to abnormal records.

[0070] Furthermore, all state variables use a unified symbol and unit system within the same batch. The remaining free acid, occupied acid, theoretical acid consumption of the oxide layer, and acid consumption increment of the non-oxide layer are all measured using the acid quantity scale. The surface area to be treated and the area of ​​the acid consumption calculation zone are measured using the area scale. The concentration detection value and the theoretical concentration value are measured using the concentration scale. The dilution disturbance is converted to the acid quantity scale through a conversion factor before entering the acid consumption equivalent state formula. The system performs scale conversion on each data item before calculation and retains the original input value. The output result records both the original value and the converted value, so that there is no change in the meaning of the same variable between formula calculation, residual registration, and state inversion. The advantage of this embodiment is that it ensures that the calculation process is verifiable and that the meaning of variables in different processing stages is consistent.

[0071] In a preferred embodiment, if multiple target pickling endpoint states need to be selected during the pickling process of a metal container, the system only switches the target endpoint parameters based on the same acid consumption equivalent state, without changing the residual registration method and state vector composition. The target endpoint parameters can correspond to different oxide layer removal completion levels or different subsequent process requirements. However, each switch records the switching time, the target endpoint state before the switch, the target endpoint state after the switch, and the acid consumption equivalent state at the time of the switch. The system recalculates the deviation based on the target endpoint state after the switch and generates subsequent segmented allocation instructions. Allocation actions already executed before the switch are still updated into the historical residual set or the unconfirmed residual set through closed-loop updates. The advantage of this embodiment is that changes in the target endpoint will not change the core calculation chain, avoiding the break in state inversion caused by target switching.

[0072] In this embodiment, the complete implementation process can be applied to batch processing scenarios of metal containers centered on pickling solution concentration adjustment. The software processing objects are limited to batch data of metal container pickling, pickling solution concentration data, acid and water replenishment records, metal ion accumulation data, and mechanism model calculation data. The method does not rely on additional hardware structure modifications, nor does it require changes to the structure of the pickling tank, container body, or acid supply device. The key to implementation lies in expressing the state of acid reaction capacity through a computer program and generating adjustment instructions through the relationship between theoretical trajectory, actual trajectory, and residual type. The scheme discloses the implementation methods of input data source, theoretical acid consumption trajectory calculation, oxide layer equivalent load mapping, metal ion occupancy item update, residual segmentation and classification, acid consumption equivalent state inversion, consistency verification before adjustment, and closed-loop update after adjustment. The advantage of this embodiment is that it can cover the entire pickling concentration adjustment process with continuous data processing logic, so that acid replenishment, water replenishment, drainage dilution, homogenization waiting, and suspension of replenishment of metal container pickling solution are all driven by the same mechanism state.

Claims

1. A method for adjusting the pickling concentration of metal containers based on a mechanistic model, characterized in that, include: Obtain the material type, surface area to be treated, oxide layer grade, initial concentration of pickling solution, acid replenishment record, water replenishment record, pickling duration, cumulative metal ion amount and free acid detection value of the batch of metal containers to be pickled; Based on the mass conservation relationship, oxide dissolution relationship, and replenishment and dilution relationship, an acid washing mechanism model was constructed, which includes free acid contribution term, theoretical oxide layer consumption term, metal ion occupation term, and dilution disturbance term, and the theoretical acid consumption trajectory was calculated. The actual concentration trajectory formed by the free acid detection value is residually registered with the theoretical acid consumption trajectory to obtain a concentration residual sequence. The acid consumption equivalent state of the pickling solution was retrieved based on the concentration residual sequence. Based on the deviation between the acid consumption equivalent state and the target pickling endpoint state, a segmented allocation instruction is generated.

2. The method for adjusting the pickling concentration of metal containers based on a mechanistic model according to claim 1, characterized in that, The calculation of the theoretical acid consumption trajectory includes: The surface area to be treated, the material type, and the oxide layer grade are mapped to the equivalent oxide layer load, and the initial theoretical consumption is calculated according to the acid consumption equivalent coefficient of different oxide materials. The pickling duration is discretized into multiple calculation intervals. Within each calculation interval, the theoretical consumption of the oxide layer is deducted based on the contribution of free acid, and the metal ion occupancy is updated based on the accumulated amount of metal ions. The acid replenishment record and the water replenishment record are converted into interval mass increments, which are then superimposed on the dilution perturbation term to form the theoretical concentration value for the corresponding calculation interval.

3. The method for adjusting the pickling concentration of metal containers based on a mechanistic model according to claim 1, characterized in that, The residual registration includes: A unified time axis was established based on the acid replenishment time, water replenishment time, and free acid detection time, and the actual concentration trajectory and the theoretical acid consumption trajectory were resampled in the same interval. The trajectory differences after resampling are segmented and classified to form detection fluctuation residuals, oxide layer load deviation residuals, metal ion cumulative deviation residuals, and replenishment mixing hysteresis residuals. The segmentation and classification are determined based on the duration of the residual, the consistency of its direction, the distance from the time of liquid replenishment, and the synchronous relationship with the cumulative changes of metal ions.

4. The method for adjusting the pickling concentration of metal containers based on a mechanistic model according to claim 3, characterized in that, The inversion of the acid consumption equivalent state includes: Construct a state vector that includes the amount of remaining free acid, the amount of occupied acid, the equivalent load of the remaining oxide layer, and the dilution disturbance. The concentration residual sequence is mapped to the state vector correction components according to the residual type, and the association between each correction component and the corresponding residual type is retained. The range of values ​​of the correction components is limited by the mass conservation constraint of the pickling mechanism model. The acid consumption equivalent available for oxide layer dissolution within a subsequent preset calculation interval is calculated using the corrected state vector, and this is taken as the acid consumption equivalent state.

5. The method for adjusting the pickling concentration of metal containers based on a mechanistic model according to claim 2, characterized in that, The mapping of the equivalent load of the oxide layer includes: The surface area to be treated is divided into acid consumption calculation partitions according to material type and oxide layer grade, and a corresponding oxide composition weight is set for each acid consumption calculation partition. Based on the oxide composition weights, each acid consumption calculation zone is converted into a standard oxide layer load, and the weld heat-affected zone, corner area, and ordinary surface area are respectively included in different load correction terms; The standard oxide layer loads and their load correction terms are summarized into the batch oxide layer equivalent loads and input into the pickling mechanism model.

6. The method for adjusting the pickling concentration of metal containers based on a mechanistic model according to claim 2, characterized in that, The update of the metal ion occupancy item includes: The amount of acid occupied by metal salts in the acid solution is determined based on the accumulated amount of metal ions, and the amount of acid occupied is stored separately from the contribution of free acid. When the increase in the cumulative amount of metal ions in adjacent calculation intervals exceeds the theoretical increase derived from the theoretical consumption term of the oxide layer, the excess portion is marked as the non-oxide layer acid consumption increase. The increase in acid consumption of the non-oxidized layer is written into the additional occupancy sub-item of the metal ion occupancy item, and the corresponding free acid contribution is deducted in the subsequent calculation of the theoretical concentration value.

7. The method for adjusting the pickling concentration of metal containers based on a mechanistic model according to claim 4, characterized in that, The segmentation and classification include: Set up a sliding calculation window on a unified time axis, and calculate the mean sign, slope direction, and duration of the residuals within each sliding calculation window; When the residual appears only at a single detection time and does not continue with adjacent times, it is classified as detection fluctuation residual; When the residual forms a delayed peak after the replenishment time, it is classified as the replenishment mixing hysteresis residual; When the residual increases synchronously with the cumulative change of metal ions, it is classified as the cumulative deviation residual of metal ions; The remaining residuals that continuously increase in one direction are classified as oxide layer load deviation residuals.

8. The method for adjusting the pickling concentration of metal containers based on a mechanistic model according to claim 7, characterized in that, The generation of the segmented allocation instruction includes: The acid consumption equivalent state is divided into four categories: low concentration and sufficient acid consumption equivalent, high concentration and insufficient acid consumption equivalent, high metal ion occupation, and unstable dilution disturbance. For states where the concentration is low but the acid consumption equivalent is sufficient, a homogenization waiting instruction is generated. For states where the concentration meets the standard but the acid consumption equivalent is insufficient, a small-dose, multi-stage acid replenishment instruction is generated. For states where the metal ion occupancy is high, a draining dilution and acid replenishment linkage instruction is generated. For states where the dilution disturbance is unstable, a replenishment pause instruction is generated.

9. The method for adjusting the pickling concentration of metal containers based on a mechanistic model according to claim 8, characterized in that, It also includes consistency verification before allocation: Establish a corresponding verification table between the residual types obtained from the segmentation and classification and the four types of states; Before generating the segmented allocation instruction, the current dominant residual type is read and matched with the state category corresponding to the acid consumption equivalent state; When the dominant residual type does not match the state category, re-perform residual registration and reduce the correction weight of the most recent free acid detection value in the state inversion. When two consecutive calculation intervals match, the corresponding segment allocation instruction is written into the instruction sequence to be executed.

10. The method for adjusting the pickling concentration of metal containers based on a mechanistic model according to claim 9, characterized in that, This also includes closed-loop updates after allocation: After executing the sequence of instructions to be executed, the free acid detection value, metal ion accumulation and replenishment record after execution are obtained and compared with the interval theoretical concentration value predicted by the pickling mechanism model before execution. When the newly added residual formed by the comparison is consistent with the dominant residual type, the newly added residual is written into the historical residual set of the corresponding residual type; When the newly added residual is inconsistent with the dominant residual type, the newly added residual is written into the set of residuals to be confirmed, and is used only as a transient correction amount for diluting the disturbance term in the state inversion of the next calculation interval.

Images