A scheduling optimization method, device and equipment for chemical materials

Abstract

The embodiment of the specification relates to the technical field of chemical products, in particular to a scheduling optimization method, device and equipment for chemical materials, and the method comprises the following steps: initializing a plurality of brand sequences corresponding to a plurality of chemical material production devices; calculating daily production of each batch of chemical materials according to production rate and production duration of each batch in each brand sequence; determining loss index of the chemical material production device and daily consumption and utilization rate of the chemical raw material according to daily production and physical property parameters of each batch of chemical materials; taking maximizing the utilization rate and minimizing the loss index as the target, taking the daily consumption satisfying the daily inventory constraint and the starting time of each batch satisfying the switching time window constraint as the condition, constructing a mixed integer linear programming model, the mixed integer linear programming model comprising a production rate linear term and a production duration linear term for representing daily production of each batch of chemical materials; and optimizing the plurality of brand sequences according to the mixed integer linear programming model.

Description

TECHNICAL FIELD

[0001] Embodiments of the present specification relate to the technical field of chemical products, in particular to a scheduling optimization method, device and equipment for chemical materials. BACKGROUND

[0002] Polyethylene, polypropylene, polystyrene, butadiene rubber, cis-butadiene rubber and other chemical polymer materials are important foundations of modern industrial systems. With the continuous expansion of the chemical industry, the product categories are increasingly diverse, and the production scheduling business of chemical enterprises is facing unprecedented complexity and refinement requirements. In particular, in the multi-grade production environment, how to develop a grade switching and production scheduling scheme that can quickly respond to market changes while meeting multiple production constraints has become a core bottleneck that needs to be solved in the industry.

[0003] In the production process of chemical materials, different grades of chemical materials will have different effects on the service life and energy consumption of the production device due to their physical and chemical differences. Currently, production scheduling mainly relies on human experience, and it is difficult to achieve systematic collaborative optimization of multiple targets such as service life, energy consumption and raw material utilization rate, resulting in problems such as significant device service life loss, high overall energy consumption, and low raw material utilization efficiency. SUMMARY

[0004] The purpose of the embodiments of the present specification is to provide a scheduling optimization method, device and equipment for chemical materials to overcome the problems of significant device service life loss, high overall energy consumption, and low raw material utilization efficiency in existing methods.

[0005] To solve the above technical problems, the specific technical solutions of the embodiments of the present specification are as follows:

[0006] On the one hand, the embodiments of the present specification provide a scheduling optimization method for chemical materials, comprising:

[0007] initializing a plurality of grade sequences corresponding to a plurality of chemical material production devices, each grade sequence including the start time, production rate and production duration of a plurality of batches, and each batch corresponding to a grade with specific physical and chemical properties;

[0008] According to the production rate and production duration of each batch in each grade sequence, the daily output of each batch of chemical materials is calculated;

[0009] According to the daily output and physical property parameters of each batch of chemical materials, the loss index of the chemical material production device and the daily consumption and utilization rate of the chemical raw material are determined, and the loss index is used to represent the influence degree of producing each batch of chemical materials on the service life of the chemical material production device;

[0010] construct a mixed integer linear programming model, the mixed integer linear programming model including a production rate linear term and a production duration linear term for representing daily production of each batch of the chemical material, the mixed integer linear programming model being constructed with a target of maximizing the utilization rate and minimizing the loss index, and a condition that the daily consumption meets a daily inventory constraint and starting time of each batch meets a switching time window constraint;

[0011] optimize the plurality of grade sequences according to the mixed integer linear programming model.

[0012] In another aspect, the embodiments of the present specification provide a scheduling optimization device for chemical materials, comprising:

[0013] an initialization module configured to initialize a plurality of grade sequences corresponding to a plurality of chemical material production devices, each grade sequence including starting time, production rate and production duration of a plurality of batches, and each batch corresponding to a grade having specific physical and chemical properties;

[0014] a calculation module configured to calculate daily production of each batch of the chemical material according to the production rate and the production duration of each batch in each grade sequence;

[0015] a determination module configured to determine a loss index of the chemical material production device and daily consumption and utilization rate of the chemical raw material according to the daily production of each batch of the chemical material and the physical property parameter, the loss index being used to represent the influence degree of producing each batch of the chemical material on the service life of the chemical material production device;

[0016] a construction module configured to construct a mixed integer linear programming model with a target of maximizing the utilization rate and minimizing the loss index, and a condition that the daily consumption meets a daily inventory constraint and starting time of each batch meets a switching time window constraint, the mixed integer linear programming model including a production rate linear term and a production duration linear term for representing daily production of each batch of the chemical material;

[0017] an optimization module configured to optimize the plurality of grade sequences according to the mixed integer linear programming model.

[0018] In another aspect, a computer device is provided, a memory is configured to store a computer program, and a processor is configured to execute the computer program to implement the above-mentioned scheduling optimization method for chemical materials.

[0019] In another aspect, the embodiments of the present specification further provide a computer readable storage medium storing a computer program, the computer program being executed by a processor to implement instructions of the above-mentioned any one of the scheduling optimization method for chemical materials.

[0020] From the technical solutions provided by the embodiments of the present specification, it can be seen that the embodiments of the present specification can initialize a plurality of grade sequences corresponding to a plurality of chemical material production devices, each grade sequence including starting time, production rate and production duration of a plurality of batches, and each batch corresponding to a grade with specific physical and chemical properties; calculating the daily output of each batch of chemical materials according to the production rate and production duration of each batch in each grade sequence; determining the loss index of the chemical material production device and the daily consumption and utilization rate of the chemical raw material according to the daily output and physical property parameters of each batch of chemical materials, the loss index being used to represent the degree of influence of producing each batch of chemical materials on the service life of the chemical material production device; taking maximizing the utilization rate and minimizing the loss index as the target, and taking the daily consumption satisfying the daily inventory constraint and the starting time of each batch satisfying the switching time window constraint as the condition, a mixed integer linear programming model is constructed, the mixed integer linear programming model including a production rate linear term and a production duration linear term used to represent the daily output of each batch of chemical materials; and optimizing the plurality of grade sequences according to the mixed integer linear programming model. By decomposing the grade sequence into a plurality of batches with accurately defined starting time, rate and duration, the transformation from macro planning to executable operation instructions is realized, and the traditional manual production scheduling problem is solved. By combining daily output with physical property parameters to quantitatively evaluate the device loss, the scheduling can actively balance the production task and the equipment health, delay the device aging from the source, and reduce the operation and maintenance risk. At the same time, by establishing the correlation between the daily output and the daily consumption, it is ensured that the daily inventory constraint can still be met under continuous time optimization, and the risk of raw material shortage or overflow is avoided. Taking maximizing the raw material utilization rate as the target, the model drives to reduce the transition material and improve the yield. By introducing the rate and duration linear term to construct the mixed integer linear programming model, the complex nonlinear problem is transformed into a form that can be efficiently solved. Finally, by cooperatively optimizing the raw material utilization rate maximization and the device loss minimization, and under the constraints of inventory and time window, a multi-dimensional optimization model covering material, time and equipment state is constructed, and a comprehensive optimal and executable scheduling decision is realized. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present specification or the prior art, a brief introduction will be given below to the drawings needed to be used in the embodiments or prior art description.

[0022] Figure 1 is a flowchart of a scheduling optimization method for a chemical material provided by the embodiments of the present specification;

[0023] Figure 2 is a whole logic flow schematic diagram of a scheduling optimization method for a chemical material provided by the embodiments of the present specification;

[0024] Figure 3 is a grade switching schematic diagram provided by the embodiments of the present specification;

[0025] Figure 4 is a chemical material brand production scheduling optimization result schematic diagram provided by an embodiment of the present specification;

[0026] Figure 5 is a raw material ethylene daily balance calculation result schematic diagram provided by an embodiment of the present specification;

[0027] Figure 6 is a structure composition schematic diagram of a chemical material scheduling optimization device provided by an embodiment of the present specification;

[0028] Figure 7 is a structure composition schematic diagram of a computer device provided by an embodiment of the present specification. DETAILED DESCRIPTION

[0029] The technical solutions in the embodiments of the present specification will be described clearly and completely below with reference to the drawings in the embodiments of the present specification. Obviously, the described embodiments are only part of the embodiments of the present specification, rather than all the embodiments. Based on the embodiments in the present specification, all other embodiments obtained by those of ordinary skill in the art without creative labor should fall within the scope of protection of the present specification.

[0030] It should be noted that the terms "first", "second", and the like in the present specification and claims and the above-described drawings are used to distinguish similar objects, and do not necessarily have to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present specification described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, device, product, or apparatus that includes a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products, or apparatuses.

[0031] In some embodiments, the chemical raw material can be a basic chemical or monomer that participates in a chemical reaction and is consumed to generate a target product. The chemical raw material is the input and material basis of the production process. In the production process of the regenerated chemical material, the inventory of the chemical raw material has dynamic fluctuation. They are warehoused, stored, consumed in devices, and possibly supplied externally, so their inventory level is always changing. The chemical raw material can specifically include liquid or gaseous basic chemical monomers such as ethylene, propylene, butadiene, and styrene.

[0032] In some embodiments, a grade can be a specific specification or model of a chemical material produced on the same set of production devices by adjusting process parameters (such as catalyst, temperature, pressure, raw material ratio, etc.) with different quality indicators and performance. It represents a specific operating state of the device. The grade has switchability and difference. Different grades correspond to different product standards, and the switching between grades is not instantaneous, involving a complex transition process.

[0033] In some embodiments, a chemical material can be a final solid or liquid polymer product converted from a chemical raw material through chemical synthesis, polymerization, etc. The chemical material is the output of the production process. The chemical material is a physical entity collection of various grades. The chemical material can specifically include polyethylene, polypropylene, polystyrene, styrene-butadiene rubber, cis-butadiene rubber, etc.

[0034] In some embodiments, a chemical material production device can be a general term for large-scale continuous flow production equipment and systems that realize the conversion from a chemical raw material to a chemical material, including reactors, towers, heat exchangers, conveying systems, etc. A set of chemical material production devices can have the ability to produce multiple grades, but its production capacity (such as maximum / minimum feed rate) has a physical upper limit. The chemical raw material can be put into the chemical material production device according to the specific grade formula and process, and the corresponding specification of the chemical material can be produced.

[0035] The embodiments of the present specification provide a scheduling optimization method for a chemical material. Referring to Figure 1 and Figure 2 , in specific implementation, the following steps are included:

[0036] S101: Initialize a plurality of grade sequences corresponding to a plurality of chemical material production devices, each grade sequence including the start time, production rate and production duration of a plurality of batches, and each batch corresponding to a grade with specific physical and chemical properties.

[0037] In some embodiments, a grade can be a specific specification or model of a chemical material produced on the same set of production devices by adjusting catalyst, hydrogen concentration, temperature, pressure, etc. Process conditions to produce a chemical material product specification with predetermined and mutually different physical and chemical properties. Physical and chemical properties can include melt index, density, molecular weight distribution physical property parameters, etc.

[0038] In some embodiments, a grade sequence can be a collection of a plurality of production batches arranged in time sequence designed for a production device. Each sequence specifies which grades are produced in sequence by the chemical material production device within the scheduling period, and the switching relationship between the grades.

[0039] In some embodiments, a batch can be a time period in which a chemical material production device continuously produces a specific grade, and is a basic unit of the production event of the grade in the time dimension. Each batch can be described by three parameters: start time, production duration, and production rate.

[0040] For a subsequent batch (i.e., a non-first batch) in any grade switching sequence, the start time thereof strictly corresponds to the end time of the previous batch, and the time point at which the switching operation from the previous grade to the current grade occurs.

[0041] In some embodiments, the step S101 can further specifically include: constructing a grade switching directed graph of each chemical material production device, the grade switching directed graph including a plurality of nodes and a plurality of directed edges, each node being associated with a grade, and each directed edge being used to represent a grade switching relationship between two connected nodes; and generating a grade sequence set of each chemical material production device based on the grade switching directed graph, each grade sequence in the grade sequence set corresponding to a production scheme of the chemical material production device.

[0042] In some embodiments, the grade switching directed graph can be a data structure for expressing grade switching rules by using a graph theory method. The nodes in the graph represent all possible grades, and the directed edges represent allowed switching directions. This structure can intuitively reflect the process compatibility between different grades.

[0043] In some embodiments, basic information of all producible grades of each chemical material production device can be collected and used as nodes of the graph. Then, based on process knowledge and historical data, feasible switching paths between grades are determined to establish directed edges. It is particularly important to note that the directed graph is usually not fully connected, which means that there is no direct switching process feasibility between any two grades. For example, switching from a high-melt-point grade to a low-melt-point grade can be prohibited, and vice versa. Accurate expression of such constraints is the basis for ensuring the executability of the generated scheduling scheme.

[0044] The complex process switching constraints are accurately expressed by the graph structure, and the switching feasibility that originally depends on the experience of engineers is converted into a formalized rule that can be processed by a computer. This expression method not only ensures that all generated sequences meet the basic process requirements, but also provides a structured knowledge base for subsequent automatic optimization.

[0045] In some embodiments, after the directed graph is constructed, a graph traversal algorithm (such as depth-first search or breadth-first search) can be employed to explore all possible production paths from the current production mix (initial state). During the traversal, multiple business constraints can be considered simultaneously: including minimum production duration requirements for each grade, grades that must be produced (market demand), grades that need to be paused (inventory rising demand), etc. In this way, tens to hundreds of feasible grade switching sequences that comply with all basic business rules can be generated for each device, forming the search space for the subsequent optimization model.

[0046] Compressing the infinite permutation and combination space into a finite set of feasible sequences greatly reduces the complexity of the optimization problem. This preprocessing mechanism makes the originally difficult-to-solve large-scale scheduling problem tractable, while ensuring that all candidate solutions are practically operable, avoiding the generation of infeasible schemes from the source.

[0047] In some embodiments, the step S101 can specifically include: selecting the grade sequence corresponding to each chemical material production device from the grade sequence set of each chemical material production device.

[0048] In some embodiments, before the optimization process starts, an initial grade sequence can be selected for each chemical material production device from its corresponding set of feasible sequences. These initial sequences can be selected based on historical scheduling experience or randomly selected. The initialization process provides a starting search point for the subsequent optimization algorithm, which plays an important role in improving optimization efficiency and ensuring the quality of the solution. In the mixed integer programming model, the decision variables corresponding to these initial sequences will be assigned initial values, guiding the solver to search in the promising region.

[0049] By using domain knowledge for intelligent initialization, blind search in meaningless solution space can be avoided, improving the solving efficiency and increasing the possibility of finding the global optimal solution.

[0050] S102: Calculate the daily production of each batch of chemical materials according to the production rate and production duration of each batch in each grade sequence.

[0051] In some embodiments, the production rate can be the number of a specific grade product produced by a chemical material production device per unit time. The production rate is limited by factors such as device physical characteristics and grade process requirements, and has a clear upper and lower limit.

[0052] In some embodiments, the production duration can be the total time length of a chemical material production device continuously producing a certain grade product, which is the span representation of the batch on the continuous time axis.

[0053] In some embodiments, the step S102 can specifically include: decomposing the production duration of each batch in each brand sequence to each natural day to obtain the daily production duration of each batch in each natural day; and determining the daily production of each batch in each natural day according to the daily production duration of each batch in each natural day and the production rate of each batch.

[0054] In some embodiments, the daily production duration can be the total production duration of the batch on the continuous time axis, which is divided according to the natural days it spans, and the effective production time in each natural day. Since the batch can start and end at any time, the production duration of the batch in each natural day needs to be accurately calculated.

[0055] In some embodiments, the step of decomposing the production duration of each batch in each brand sequence to each natural day to obtain the daily production duration of each batch in each natural day can specifically include: decomposing the production duration of each batch into a first mode production duration, one or more second mode production durations, and a third mode production duration according to natural days; the first mode production duration is the daily production duration of the batch that starts in a natural day and does not end in the natural day; the second mode production duration is the daily production duration of the batch that does not start in a natural day and does not end in the natural day; and the third mode production duration is the daily production duration of the batch that does not start in a natural day and ends in the natural day.

[0056] In some embodiments, the first mode production duration can be the period when the batch starts production in a natural day but does not end in the natural day. The duration is the time from the start time of the batch to the end time of the day (i.e., 24:00).

[0057] In some embodiments, the second mode production duration can be the period when the batch produces 24 hours completely in a natural day. This mode requires the batch to neither start nor end in the natural day, but to be in a continuous production state.

[0058] In some embodiments, the third mode production duration can be the period when the batch ends production in a natural day but does not start in the natural day. The duration is the time from the start time of the day (i.e., 00:00) to the end time of the batch.

[0059] In some embodiments, based on the two continuous time variables of the start time and the end time of each batch, all the natural days spanned by the batch can be automatically identified. Then, for each spanned natural day, according to the time characteristics of the batch in the day, the production duration of the batch is classified into the corresponding mode:

[0060] For the natural day in which the batch starts (and does not end in the same day), the time difference from the batch start time to the end of the day is calculated and recorded as the first mode production duration. This duration is necessarily less than 24 hours, representing the initial production phase of the batch on that day.

[0061] For the natural day in which the batch ends (and does not start in the same day), the time difference from the start of the day to the batch end time is calculated and recorded as the third mode production duration. This duration is also less than 24 hours, representing the final production phase of the batch on that day.

[0062] For the intermediate natural days that the batch passes through entirely (neither starting nor ending on the same day), the production duration is directly identified as 24 hours, recorded as the second mode production duration. These natural days represent the stable production phase of the batch.

[0063] By establishing standardized three production modes, the conversion problem between continuous time scheduling and discrete day accounting is successfully solved. This decomposition method transforms complex continuous time calculation into a normative pattern recognition problem, ensuring the accuracy of time calculation and providing a structured framework for establishing a unified mathematical model. In particular, this method can accurately handle the situation where the batch starts and ends at any time on the natural day boundary, eliminating the calculation errors brought by traditional approximation methods.

[0064] In some embodiments, after completing the multi-mode decomposition, the production of each mode can be calculated according to the basic principle of production = production rate x production duration. For the first and third modes, since the production duration is a variable, linearization techniques are used to handle the variable multiplication term; for the second mode, since the production duration is a fixed value of 24 hours, the production calculation is simplified to the product of the production rate and 24.

[0065] In some embodiments, the total daily production of each batch on each natural day is the cumulative sum of the production of its three modes. Through this decomposition calculation method, even if the batch spans multiple natural days, the independent production of each batch on each natural day can be accurately calculated, providing accurate data basis for subsequent daily material accounting.

[0066] The daily production calculation method based on multi-mode decomposition realizes the accurate mapping from continuous time variables to discrete daily production. This accurate calculation capability ensures the accuracy of subsequent material balance, providing reliable data support for fine scheduling. Compared with traditional methods, this technology can accurately reflect the true production distribution of each batch on each natural day, providing a more detailed reference for production decisions.

[0067] S103: Determine the loss index of the chemical material production device and the daily consumption and utilization rate of the chemical raw material according to the daily output and physical property parameters of each batch of chemical material. The loss index is used to represent the influence degree of producing each batch of chemical material on the service life of the chemical material production device.

[0068] In some embodiments, the determination of the loss index of the chemical material production device in step S103 can specifically include: calculating the unit loss of each batch of chemical material to the chemical material production device according to the physical property parameters of each batch of chemical material; and determining the loss index of each chemical material production device according to the unit loss and the daily output of each batch of chemical material.

[0069] In some embodiments, the physical property parameters can be a set of measurable physical and chemical property indexes inherent to the grade of chemical material, including melt index, density, molecular weight distribution, catalyst residual activity, monomer conversion characteristics, etc. These parameters are determined by the formula and process of the grade, and directly affect the fluid mechanics characteristics, heat transfer efficiency, catalyst deactivation rate and equipment fouling tendency in the production process.

[0070] In some embodiments, the unit loss can be the degree of wear, corrosion or fatigue of the key components (such as the inner wall of the reactor, the stirring paddle, the heat exchanger tube bundle, and the flow part of the pump) in the production device when producing a unit mass (such as per ton) of a certain grade of chemical material due to its unique physical property parameters. It is a conversion coefficient that quantifies the grade physical property into the equipment impact intensity.

[0071] In some embodiments, the loss index can be a quantitative evaluation value of the cumulative equipment loss of a single chemical material production device due to the execution of the production task of each batch in the selected grade sequence. The index is calculated by weighting and accumulating the unit loss of each batch and the actual output of the batch, and directly reflects the influence of the specific production plan on the mechanical integrity and expected remaining life of the device.

[0072] In some embodiments, a physical property-loss mapping model can be built-in or accessed. The model is established based on principles of equipment engineering, material corrosion and chemical reaction engineering. For each grade, a complete list of physical property parameters can be obtained and input into the model. The model outputs one or more unit loss coefficients. Specifically, a high melt index grade can result in higher shear stress, and its unit mechanical wear coefficient is higher; some grades containing specific monomers may be more corrosive, and their unit chemical corrosion coefficient is higher. Finally, these coefficients can be integrated into a unit loss value representing the overall wear severity of the grade.

[0073] In some embodiments, the calculation of the loss index for a single device is a batch-wise accumulation process. First, the grade information and optimized daily production plan for each batch scheduled for the device in the optimization model are obtained. Then, the following calculation is performed for each batch: the loss contributed by the batch = unit loss of the batch grade x total production of the batch. The total production of the batch is obtained by summing up its daily production. Finally, the loss contributed by all batches of the device is added up, which is the loss index of the device in the entire scheduling period. The index is a comparable numerical value directly related to the production plan.

[0074] The traditional scheduling cannot quantify the long-term impact of different production schemes on the equipment life. By establishing a unit loss model based on physical parameters, the implicit and long-term equipment health factors are converted into explicit, immediate and quantifiable indicators (loss index). After calculating the loss index, it is used as an optimization target together with the raw material utilization rate, forcing the optimization model to find the best balance between improving the original utilization rate and protecting the equipment safety.

[0075] In some embodiments, the above step S103 can further include: calculating a loss fluctuation index of the plurality of chemical material production devices according to the loss index of each chemical material production device, the loss fluctuation index being used to represent the difference in the degree of life impact between the plurality of chemical material production devices.

[0076] In some embodiments, the loss fluctuation index can be a measure of the dispersion or difference of the respective loss indices of a plurality of parallel running chemical material production devices in the entire production system. It represents the imbalance of the aging or wear rate between devices caused by the scheduling scheme.

[0077] In some embodiments, after obtaining the loss indices (L1, L2,..., Ln) of all N devices, the statistical dispersion of these indices is calculated as the loss fluctuation index. One way of calculation can be the standard deviation or the coefficient of variation. The larger the index value, the more uneven the loss conditions between devices.

[0078] By minimizing this loss fluctuation index, the optimization model will tend to distribute high-wear production tasks more evenly among devices or arrange appropriate rotation, rather than allowing a device to bear high loss for a long time. This helps to avoid the short board effect, prolong the overall overhaul cycle of the device group, disperses the unplanned shutdown risk caused by excessive wear of a single device from the system level, and significantly improves the robustness and reliability of the system.

[0079] In some embodiments, the determination of the daily consumption of the chemical raw material in step S103 can specifically include: determining the daily consumption of the chemical raw material according to the daily production of each batch of chemical material and the unit consumption of the chemical raw material of each batch.

[0080] In some embodiments, the unit consumption of the chemical raw material can be a standard amount of a certain raw material required for producing a unit mass of a certain grade of chemical material.

[0081] In some embodiments, the daily consumption can be the total amount of a certain raw material consumed by all production devices to complete the production task of each batch within a single natural day.

[0082] In some embodiments, based on the accurate production data of each batch in each natural day obtained through multi-mode decomposition, and in combination with the pre-established grade raw material unit consumption database, the systematic calculation of the daily consumption is performed.

[0083] Specifically, for each natural day and each raw material, the calculation process is as follows: all batches being produced in the natural day (including the normal grade and the transition material grade) are traversed, the daily production of each batch is multiplied by the unit consumption of the specific raw material corresponding to the grade of the batch to obtain the daily consumption of the raw material by the batch; and then the consumptions of all batches are added to obtain the total consumption of the raw material in the natural day.

[0084] In some embodiments, the improvement of the utilization rate of the chemical raw material can be achieved by maximizing the utilization rate of the raw material through the optimization model, and the core connotation is to maximize the conversion of the scarce chemical raw material into the normal grade product while minimizing the generation of the transition material.

[0085] In some embodiments, the above grades can include the normal grade and the transition material grade, the normal grade corresponds to the target chemical material product of the chemical material production device, and the transition material grade corresponds to the non-target chemical material product of the chemical material production device. Based on this, the determination of the utilization rate of the chemical raw material in step S103 can specifically include: determining the utilization rate of the chemical raw material according to the daily production of each batch of chemical material corresponding to the normal grade.

[0086] In some embodiments, the normal grade can be a target chemical material grade that fully meets the predetermined quality standard and can be sold as a qualified product, and is the target chemical product in the production process.

[0087] In some embodiments, the transition material grade can be an intermediate product grade produced in the grade switching process, and the quality indicators of which do not meet the normal product standard, and the generation thereof is an inevitable byproduct in the grade switching process.

[0088] In some embodiments, the optimization of chemical raw material utilization is not achieved through simple formula calculation, but through the construction of an optimization function targeting the maximization of raw material value, and is realized in an iterative solving process. The optimization function guides the model to intelligently allocate raw material resources under the premise of meeting all constraint conditions.

[0089] Specifically, the production of positive brand numbers can be prioritized, and the production of positive brand numbers can be ensured to meet market requirements through demand constraints; at the same time, the production of transition materials can be minimized through optimization of switching sequence and production parameters; in addition, the stability of raw material supply can be ensured through daily inventory constraints, and production interruptions caused by raw material shortages can be avoided.

[0090] In each iterative solving process, the raw material utilization efficiency under different production schemes can be evaluated, and the production parameters and sequence selection of each batch can be continuously adjusted to gradually approach the optimal state of raw material utilization. This optimization is global and considers the configuration efficiency of raw materials in positive product production, transition material production, and inventory management in multiple dimensions.

[0091] By establishing a daily granularity-based consumption calculation system, accurate tracking and management of raw material flow can be achieved. This fine management helps to discover potential risks in material balance in a timely manner and provides accurate data support for production decisions.

[0092] S104: Construct a mixed integer linear programming model with the conditions that the daily consumption meets the daily inventory constraints and the start time of each batch meets the switching time window constraints, with the goal of maximizing the utilization rate and minimizing the loss index, the mixed integer linear programming model including production rate linear terms and production duration linear terms for representing daily production of each batch of chemical materials.

[0093] In some embodiments, the mixed integer linear programming model can be a mathematical optimization model that includes both continuous variables and integer variables, where the objective function and constraint conditions are linear expressions. In this scheduling problem, continuous variables represent production rates, times, etc., and integer variables (especially 0-1 variables) are used to represent sequence selection, mode selection, etc. discrete decisions.

[0094] In some embodiments, the above step S104 can specifically include: constructing a mixed integer linear programming model with the conditions that each brand sequence meets the sequence unique selection constraint, and the daily consumption meets the daily inventory constraint, and the start time of each batch in the brand sequence meets the switching time window constraint, and the production rate of each batch in the brand sequence meets the production capacity constraint, and the production of the chemical material meets the demand constraint, with the goal of maximizing the utilization rate, minimizing the total loss index of the plurality of chemical material production devices, and minimizing the loss fluctuation index, the mixed integer linear programming model including production rate linear terms and production duration linear terms for representing the production.

[0095] In some embodiments, the total loss index can be the sum of all the loss indices of the production system consisting of multiple chemical production devices. It quantifies the total wear, corrosion, or fatigue load on the key equipment assets of the entire plant caused by a particular scheduling scheme. Minimizing this objective aims to reduce the total life-time depreciation of the equipment population from a global perspective.

[0096] In some embodiments, the loss fluctuation index can be the statistical dispersion (e.g., standard deviation or coefficient of variation) among the loss indices of the individual devices. Minimizing this objective aims to pursue an even distribution of equipment wear, avoiding an uneven situation where some devices are over-worn, corroded, or fatigued while others are lightly loaded, thus improving the overall reliability and maintenance smoothness of the equipment population.

[0097] In some embodiments, the daily inventory constraint can be a mandatory requirement on the raw material inventory, including that the end-of-day inventory must be maintained between a pre-set high limit and low limit, and a daily material balance relationship based on daily consumption, daily supply, and initial inventory.

[0098] In some embodiments, the switching time window constraint can be a limitation on the switching time (i.e., the start time of a batch), requiring that the switching operation must occur within an allowed time range, such as being limited to the daytime shift period of a workday, to avoid performing complex switching operations at night or on weekends.

[0099] In some embodiments, the production capacity constraint can be a physical limitation on the production rate of a device, requiring that the production rate of each grade must be within the minimum and maximum production capacity range set by the device for that grade.

[0100] In some embodiments, the demand constraint can be a market limitation on the production of chemical materials, requiring that the total production of each grade must satisfy the upper and lower limit range of the market demand forecast.

[0101] In some embodiments, the sequence unique selection constraint can be a mandatory condition requiring that each production device must select only one grade switching sequence from its feasible sequence set.

[0102] In some embodiments, a normalization coefficient or decision weight (w1, w2, w3) can be set for each of the three sub-goals, i.e., raw material utilization (U), total loss index (L_total), and loss fluctuation index (L_variance). Then, a comprehensive objective function can be constructed, for example, using a linear weighting method: Maximize Z = w1*U - w2*L_total - w3*L_variance. Here, U is to maximize the utilization, and L_total and L_variance are to minimize the total loss index and loss fluctuation index, so they are added in the form of a negative sign when maximizing Z. The weights reflect the importance of different goals. Further, a complete linear constraint system can be established, including: (1) sequence selection constraints, ensuring that each device selects and only selects one feasible sequence through 0-1 variables; (2) time-related constraints, including batch-to-batch timing connection, production time range limitation, and switching time window limitation; (3) material balance constraints, establishing daily raw material inventory balance formula based on the daily consumption obtained by multi-mode decomposition and ensuring inventory safety; (4) production capacity constraints, limiting the production rate of each grade within the range allowed by the process; (5) demand constraints, ensuring that the total production of each grade meets the upper and lower limits of the production requirement.

[0103] By constructing a multi-objective optimization model, short-term raw material utilization, long-term equipment life loss, and system operation risk are placed in the same framework at the scheduling level for automated trade-off and solution. This forces the final solution to achieve a Pareto optimal balance in these three dimensions, which is far superior to any method that only considers a single objective or relies on human experience compromise, achieving a fundamental improvement in scheduling decision-making from single dimension to multi-dimensional coordination.

[0104] In some embodiments, the production rate linear term includes a linear term corresponding to each batch based on a production rate linear coefficient; and the production time linear term includes a linear term corresponding to each batch based on a first mode production time linear coefficient and a third mode production time linear coefficient. Based on this, step S104 can further include: using Taylor expansion algorithm to linearize the daily production according to the first mode production time, the second mode production time, and the third mode production time of each batch in each natural day, and the production rate of each batch, to obtain the first mode production time linear coefficient corresponding to the first mode production time, the third mode production time linear coefficient corresponding to the third mode production time, and the production rate linear coefficient corresponding to the production rate.

[0105] In some embodiments, the production rate linear term can be a linear expression containing a production rate variable obtained by linearization, and the coefficient is a production rate linear coefficient. This linear term can be used to approximate the contribution of production rate to daily production in the optimization model.

[0106] In some embodiments, the production duration linear terms can be derived from linearizing the linear expressions involving the production duration variables, including the first-mode production duration linear term and the third-mode production duration linear term, using different production mode duration variables.

[0107] In some embodiments, the Taylor expansion algorithm can be a mathematical method that uses the first-order Taylor formula to linearly approximate a nonlinear function. The Taylor expansion algorithm can be used to handle the nonlinear terms formed by the multiplication of the production rate and the production duration, by linearizing around the operating point to construct an approximate linear expression.

[0108] In some embodiments, the production rate linear coefficient can be a coefficient multiplied by the production rate variable in the Taylor expansion linearization process, which is related to the production duration.

[0109] In some embodiments, the first-mode production duration linear coefficient can be a coefficient multiplied by the first-mode production duration variable in the Taylor expansion linearization process, which is related to the production rate.

[0110] In some embodiments, the third-mode production duration linear coefficient can be a coefficient multiplied by the third-mode production duration variable in the Taylor expansion linearization process, which is related to the production rate.

[0111] In some embodiments, the first-mode duration (from the start time to the end of the day) and the third-mode duration (from the start of the day to the end time) of a batch in a given natural day are continuous decision variables. Because they depend on the exact start time and end time of the batch, which are themselves the core continuous variables that the model needs to optimize, their specific values are not known before solving. The second-mode duration (a constant) indicates that the batch is produced for 24 hours in the natural day. It is a state of a Boolean property (yes or no). If a batch is identified to be in the second mode in a natural day, its production duration is fixed at 24 hours. Therefore, it is a known constant, not a variable. The fundamental purpose of introducing the linear coefficients is to use Taylor expansion to approximate the handling of the variable x variable, a nonlinear term. For the first and third modes, the production is calculated as: production = production rate variable x first / third-mode duration variable, which is a nonlinear term, and therefore must be linearized by Taylor expansion, thereby introducing the first / third-mode production duration linear coefficient and the production rate linear coefficient. For the second mode, the production is calculated as: production = production rate variable x 24 hours. This is essentially a linear term of variable x constant. It is already linear by itself, and there is no nonlinear problem that needs to be linearized. Therefore, it can be directly included in the model without introducing an additional second-mode production duration linear coefficient.

[0112] In some embodiments, initial guess values can be provided for the production rate, the first mode production duration, and the third mode production duration of each batch. These guess values can be determined based on historical production data, process design values, or results of previous iterations. The quality of the guess values will directly affect the accuracy of the linearization approximation and the convergence speed of the subsequent optimization process.

[0113] In some embodiments, for the nonlinear term, production rate x production duration, in the daily production calculation for each batch, a first order Taylor expansion can be employed for linearization. Specifically, for the nonlinear term, a first order Taylor expansion at the guess point is: .

[0114] In some embodiments, according to the Taylor expansion formula, the linear coefficients can be calculated and assigned to the corresponding variable terms. The production rate linear coefficient is determined by the guess values of the first mode production duration and the third mode production duration; the first mode production duration linear coefficient is determined by the guess value of the production rate; and the third mode production duration linear coefficient is also determined by the guess value of the production rate. These coefficients remain constant during a single iteration process and are used as known parameters in the optimization calculation of the production rate, the first mode production duration, and the third mode production duration.

[0115] In some embodiments, after obtaining all the linear coefficients, a complete linearized daily production calculation model can be constructed. This model converts the original nonlinear optimization problem into a mixed integer linear programming problem, making it possible to use mature linear programming algorithms for efficient solution. The linearized model retains the physical meaning and engineering constraints of the original problem, while significantly reducing the computational complexity.

[0116] By Taylor expansion linearization, the complex nonlinear programming problem is converted into a mixed integer linear programming problem, making it possible to use mature and efficient linear programming algorithms for solution. This transformation greatly reduces the computational complexity of the problem, making it possible to solve large-scale practical industrial problems within an acceptable time. Further, although a linear approximation is used, through an iterative updating mechanism, the linearization error can be gradually corrected, and the solution that meets the engineering precision requirements can be finally converged. Taylor expansion has a high local approximation accuracy near the working point, and combined with the iterative optimization strategy, the technical reliability of the final solution can be guaranteed. In addition, linearization effectively avoids the numerical instability problems commonly encountered in nonlinear optimization, such as convergence failure, local optimal solution, etc. The mixed integer linear programming model has better mathematical properties and convergence guarantees, improving the robustness and reliability of the optimization process.

[0117] S105: optimizing the plurality of grade sequences according to the mixed integer linear programming model.

[0118] In some embodiments, the step S105 can specifically include: generating a plurality of new brand sequences of the plurality of chemical material production devices using a mixed integer linear programming model based on the first mode production duration linear coefficient, the third mode production duration linear coefficient and the production rate linear coefficient corresponding to each batch; updating the first mode production duration linear coefficient, the third mode production duration linear coefficient and the production rate linear coefficient corresponding to each batch according to the first mode production duration, the third mode production duration and the production rate corresponding to each batch in each new brand sequence; iteratively performing the steps of calculating the daily production, determining the loss index, the daily consumption and the utilization, generating the new brand sequence and updating the first mode production duration linear coefficient, the third mode production duration linear coefficient and the production rate linear coefficient until the optimization convergence condition is met.

[0119] In some embodiments, the mixed integer linear programming model solution can be a calculation process of finding the decision variable value that meets all the constraint conditions and optimizes the objective function by using professional mathematical optimization software (such as Gurobi, CPLEX) to calculate the established linear programming model.

[0120] In some embodiments, the new brand sequence can be an improved production arrangement scheme obtained after solving the optimization model, including updated sequence selection, accurate start time of each batch, optimized production duration and production rate, etc.

[0121] In some embodiments, the linear coefficient update can be to use the production parameter values obtained by this iteration as the new reference point for Taylor expansion linearization in the next iteration, specifically including the update of the first mode production duration linear coefficient, the third mode production duration linear coefficient and the production rate linear coefficient.

[0122] In some embodiments, the optimization convergence condition can be a quantitative standard for judging whether the iteration process can be terminated, including both the stability criterion of the solution and the limitation of the computing resources.

[0123] In some embodiments, the mixed integer linear programming model containing the latest linear coefficient can be input into the solver, and appropriate solving parameters (such as optimality tolerance, time limit, etc.) can be set. The solver uses branch and bound, cutting plane and other algorithm strategies to systematically search for the optimal value combination of the decision variables. This process automatically determines the optimal sequence selection of each device, optimizes the specific time arrangement and production parameters of each batch under this sequence, and outputs the complete new brand sequence scheme.

[0124] In some embodiments, after obtaining a new batch sequence, the key production parameters of each batch in the sequence can be extracted: the actual production rate, the first mode production time and the third mode production time of each batch are obtained from the solution. These actual values are directly used as the corresponding linear coefficient values in the next iteration, that is, the production rate value obtained in this solution is used to update the first mode production time linear coefficient and the third mode production time linear coefficient in the next iteration, and the first mode production time value and the third mode production time value in this solution are used to update the production rate linear coefficient in the next iteration.

[0125] In some embodiments, the sum of the first mode production time and the third mode production time of each batch in each new batch sequence can be used as the new production rate linear coefficient corresponding to the batch; the production rate corresponding to each batch can be used as the new first mode production time linear coefficient and the new third mode production time linear coefficient.

[0126] In some embodiments, the daily output of the chemical material, the loss index of the chemical material production device and the consumption of the chemical raw material can be calculated based on the current linear coefficients; then a new batch sequence is solved by solving the optimization model; then the linear coefficients are updated to the actual values obtained in this solution; finally, it is judged whether the convergence condition is met. If not, start a new round of iteration; if yes, terminate the iteration and output the final optimization scheme.

[0127] Through the iterative solution mechanism, the approximation error introduced by linearization is gradually corrected, ensuring that the final obtained scheduling scheme reaches or approaches the global optimum under the premise of meeting the engineering precision. This gradual optimization method effectively overcomes the problem of low solution quality caused by single linearization, providing reliable technical support for production decision-making.

[0128] In some embodiments, the above optimization convergence condition can specifically include: the variation of the production rate, the first mode production time and the third mode production time of each batch in the continuous N iterations is less than a preset tolerance, and N is a positive integer; or, the number of iterations reaches a preset maximum number of iterations.

[0129] In some embodiments, the variation can be a relative variation, which is obtained by calculating the ratio of the absolute difference between the current iteration value and the last iteration value to the last iteration value, and its calculation formula is |current value-last value| / |last value|.

[0130] In some embodiments, the preset tolerance can be a pre-set allowed error range. When the variation of the decision variable is less than this threshold value, it is considered that the solution is stable enough, for example, it is set to be between 0.1% and 1% according to the engineering precision requirement.

[0131] In some embodiments, the preset maximum iteration number can be a safety parameter set to prevent excessive consumption of computing resources, ensuring that the computing process can end within a reasonable time even in the case of slow convergence.

[0132] In some embodiments, the current iteration value of the three key decision variables (production rate, first mode production duration, and third mode production duration) is automatically extracted after each iteration is completed. For each variable, its relative change amount from the last iteration value is calculated, and whether the change amount is less than the preset tolerance is recorded. This calculation process is performed simultaneously for all corresponding variables of all batches, ensuring comprehensive monitoring of the optimization process.

[0133] In some embodiments, a convergence state counter can be maintained, and when the change amount of a certain decision variable is less than the preset tolerance in a single iteration, the convergence state of the variable is marked as true. Only when the convergence state of all key decision variables is true in consecutive N iterations, the entire optimization process is determined to have reached convergence. This continuous judgment mechanism effectively avoids false positives caused by accidental fluctuations.

[0134] In some embodiments, an iteration number counter can be maintained simultaneously, and the count is accumulated from the start of the optimization process. When the iteration number reaches the preset maximum iteration number, the optimization process is forcibly terminated regardless of the current convergence state. This mechanism serves as a safety guarantee to prevent infinite loops in abnormal situations.

[0135] The two convergence conditions work together with an OR logic relationship, and either condition being met triggers the termination of optimization. In normal cases, the optimization process is terminated through the change amount convergence condition; in abnormal or complex situations, the maximum iteration number condition ensures the end of the process. This double insurance mechanism ensures the reliability of the optimization process.

[0136] This precise monitoring of relative change amount and consecutive N times of stability requirement ensures that the obtained optimization solution truly reaches a stable state, effectively avoiding the problem of insufficient solution quality caused by premature termination, and providing a reliable guarantee for obtaining high-quality scheduling solutions.

[0137] In some embodiments, the step S105 described above can further include, after the optimization of the plurality of grade sequences, scheduling optimization of the plurality of chemical material production devices according to the optimized plurality of grade sequences.

[0138] In some embodiments, the optimized grade sequence can be the optimal production scheme obtained by iterative solving of the mixed integer linear programming model, including the selected unique grade switching sequence for each device, as well as the optimized parameters such as the starting time of each batch, the optimal production duration, and the best production rate.

[0139] In some embodiments, the scheduling optimization can be a process of converting the theoretical results of the output into executable production instructions and assigning them to specific production devices. The production instructions can be standardized operation commands generated based on the optimization results, which can be recognized and processed by the production execution system, including grade switching instructions, feed adjustment instructions, parameter setting instructions, etc.

[0140] For each device, a detailed grade switching schedule is generated, which specifies the exact time of each switching point, the target grade, and the expected transition duration. At the same time, production parameter settings are generated, including key process parameters such as production rate and raw material feed ratio for each grade. These instructions are arranged in chronological order to form a complete scheduling timeline.

[0141] Through the pre-established system interface, the generated scheduling instructions are automatically issued to the production execution system. This process can use a hierarchical issuance mechanism: first, the main framework plan for the entire cycle is issued, and then detailed operation instructions are issued near the execution time. The system establishes an instruction confirmation mechanism to ensure that each instruction is correctly received and understood.

[0142] Through scheduling instructions accurate to the hour level, fine-grained control of the production process is achieved. This precise control enables the device to operate stably under optimal process parameters, minimizing the production of transition materials during grade switching, while actively balancing production tasks and equipment health, delaying device aging from the source, and reducing operational risks. Through the entire cycle of material balance protection and inventory safety control, a stable raw material supply environment is provided for continuous production. This stability effectively avoids unplanned shutdowns due to raw material shortages and operational risks due to inventory overflows, ensuring long-term smooth operation of the production device.

[0143] In some embodiments, the above step S103 can further include determining the unit energy consumption of each chemical material production device for producing each batch of chemical material; and determining the energy consumption index of each chemical material production device based on the unit energy consumption and the production rate and production duration of each batch.

[0144] Based on this, the above step S104 can further include the condition that the daily consumption meets the daily inventory constraint, the start time of each batch meets the switching time window constraint, and the energy consumption index of each chemical material production device meets the energy consumption constraint.

[0145] In some embodiments, the unit energy consumption can be the comprehensive energy consumed by the chemical material production device when producing a unit mass (e.g., one ton) of a certain grade of chemical material, which is determined by the physical and chemical properties of the grade and the production process path. For example, producing a high melt index grade may require higher reaction temperature and longer residence time, resulting in significantly higher unit energy consumption than the base grade.

[0146] In some embodiments, the energy consumption index can be a quantitative value of the total energy predicted to be consumed by a single chemical material production device in the process of executing all batch production tasks within an entire scheduling period. It can be calculated by weighting and accumulating the unit energy consumption of each batch and the total production of the batch.

[0147] In some embodiments, the energy consumption constraint can be a limit condition set for the energy consumption index of a single or multiple chemical material production devices within an entire scheduling period. The constraint can be manifested as an upper limit constraint (such as not exceeding a certain absolute energy consumption value or budget amount), or as a target range constraint (such as being as close as possible to a certain energy efficiency target value).

[0148] In some embodiments, based on historical production data, process simulation models or first principle calculations, a unit energy consumption benchmark value is established or called for each grade that can be produced by each set of devices. The core of this process is to associate the physical property parameters with the energy consumption characteristics. Specifically, the built-in energy consumption model of the system will simulate and calculate the theoretical energy required to maintain certain reaction conditions (such as temperature, pressure) and complete subsequent separation, granulation and other processes according to the melt index, density and other parameters of the grade. Finally, a two-dimensional device-grade unit energy consumption parameter matrix is formed as an input for subsequent calculations.

[0149] In some embodiments, for any device, the calculation of its energy consumption index follows a clear accumulation rule. First, obtain the detailed information of all batches planned for the device from the optimization model, including: 1) the grade produced by each batch; 2) the optimized production rate and production time length of each batch. Then, perform the calculation for each batch: the batch energy consumption = the unit energy consumption of the grade × (production rate × production time length). Finally, add up the energy consumption of all batches of the device, and the energy consumption index of the device is obtained. This calculation directly depends on the time and rate variables of the optimization, making energy consumption a dynamic result of scheduling decisions.

[0150] In some embodiments, when constructing the mixed integer linear programming model in step S104, energy consumption can be integrated as a new hard constraint condition into the model. For each device u, its mathematical expression is: energy consumption index_u ≤ energy consumption upper limit_u, or, energy consumption target lower limit_u ≤ energy consumption index_u ≤ energy consumption target upper limit_u; where energy consumption index_u is an expression of variables containing production rate and time length defined by the above calculation formula. This constraint, together with the original daily inventory constraint, switching time window constraint, etc., constitutes the feasibility boundary of the model, ensuring that any solution meets the preset energy management requirements.

[0151] By incorporating energy consumption into the optimization framework, the model is prompted to automatically favor production paths with higher energy efficiency while meeting targets such as production yield and equipment wear. For example, the model can optimize the production sequence of grades to reduce the additional energy loss caused by drastic switching from high-energy-consumption grades to low-energy-consumption grades, or fine-tune the production rate to make the device run at a load closer to its optimal energy efficiency zone. This collaborative optimization goes beyond simple scheduling and rises to the level of integrated optimization of production process energy efficiency and production planning, tapping into hidden energy-saving potential.

[0152] Table 1

[0153]

[0154] In some embodiments, the above step S101 can further specifically include: for each chemical material production device, taking each grade as a node and switchable relationship as an edge, drawing a grade switching route diagram as shown in the figure. As can be seen, the directed graph is not fully connected, that is, not any two grades can be switched, so in the preprocessing of sequence generation, according to the feasible route of grade switching, combined with the constraints of the grade being produced at the beginning, the grade having sales demand, the grade having rising demand, and the single-batch minimum production time of each grade, a series of feasible switching sequences of each device conforming to the business rules are generated, denoted as set Figure 3 As can be seen, the directed graph is not fully connected, that is, not any two grades can be switched, so in the preprocessing of sequence generation, according to the feasible route of grade switching, combined with the constraints of the grade being produced at the beginning, the grade having sales demand, the grade having rising demand, and the single-batch minimum production time of each grade, a series of feasible switching sequences of each device conforming to the business rules are generated, denoted as set For any feasible sequence , each production event contained is called a batch.

[0155] According to specific business conditions, the feasible sequences of each production device generally vary from several to several hundred, and each sequence is composed of several batches, as shown in Table 1.

[0156] Table 2 provides a set symbol explanation provided by the embodiments of the present specification, Table 3 provides a parameter symbol explanation provided by the embodiments of the present specification, and Tables 4 and 5 provide variable symbol explanations provided by the embodiments of the present specification.

[0157] Table 2

[0158]

[0159] Table 3

[0160]

[0161] Table 4

[0162]

[0163] In some embodiments, the objective of the mixed integer linear programming model in step S104 above can be to maximize the utilization rate of chemical raw materials, minimize the total loss index of the chemical material production devices, and minimize the loss fluctuation index of the chemical material production devices. Through optimization, it is realized to convert chemical raw materials such as ethylene, propylene, butadiene, and styrene into chemical polymer materials that best match the demand while ensuring the safe and stable operation of multiple production devices, and dispersing the unplanned shutdown risk caused by excessive loss of a single device.

[0164] Table 5

[0165]

[0166] In some embodiments, the conditions in step S104 above can include sequence unique selection constraints, switching time window constraints, daily inventory constraints, production capacity constraints, and demand constraints.

[0167] In some embodiments, for each production device corresponding to multiple grade sequences, a binary variable can be used to let the model automatically determine which sequence is the optimal sequence, so that each device can only select one sequence:

[0168]

[0169] wherein, is a binary variable indicating whether sequence s of device u is selected.

[0170] In some embodiments, the start time of the first batch of each production device is 0:

[0171]

[0172] wherein, is the start time of the first batch of sequence s of device u.

[0173] In some embodiments, the end time of the last batch of each device is the total duration of the scheduling period:

[0174]

[0175] wherein, is the start time of the last batch of sequence s of device u, is the duration of the last batch of sequence s of device u, and ct is the total duration of the scheduling period.

[0176] In some embodiments, the start time of each batch is equal to the end time of the previous batch:

[0177]

[0178] wherein, is the start time of the batch b of the sequence s of the device u, is the start time of the last batch, is the duration of the last batch.

[0179] In some embodiments, the start time of each batch is between the earliest allowed start time and the latest allowed start time.

[0180]

[0181] wherein, is the start time of the batch b of the sequence s of the device u, is the earliest start time of the batch b of the sequence s of the device u, is the latest start time of the batch b of the sequence s of the device u.

[0182] In some embodiments, the duration of each batch is between the shortest allowed duration and the longest allowed duration.

[0183]

[0184] wherein, is the duration of the batch b of the sequence s of the device u, is the shortest allowed duration of the batch b of the sequence s of the device u, is the longest allowed duration of the batch b of the sequence s of the device u.

[0185] In formula (5) and formula (6), the four parameters of the earliest start time, the latest start time, the shortest duration, and the longest duration are calculated by the sequence generation module according to the initial production model, the shortest duration of the initial model, the shortest production duration of each model, the maximum demand of each model, and the requirements for the production time of individual models.

[0186] In some embodiments, since the model switching is a complex work, it is necessary to limit it as much as possible within the daytime of the working day. The time period that cannot be switched can be marked as Two binary variables and are introduced to establish five constraint formulas to exclude the time to be limited from the time period, and the specific constraints are as follows:

[0187]

[0188]

[0189]

[0190]

[0191]

[0192] wherein, is the start time of the batch b of the sequence s of the device u, 、 is the start time of the batch b of the sequence s of the device u is not allowed to be at the start time, end time of the time interval k, M is a maximum value.

[0193] 、 are two binary variables, which are defined by the constraint equation (11) 、 There is and only one taking 1.

[0194] When takes 1, takes 0, the equations (7), (8), (9), (10) are transformed into:

[0195]

[0196]

[0197]

[0198]

[0199] That is, the switching time is limited after the non-switchable time period.

[0200] When is 0, is 1, the equations (7), (8), (9), (10) are transformed into:

[0201]

[0202]

[0203]

[0204]

[0205] That is, the switching time is limited before the non-switchable time period.

[0206] In summary, through the equations (7)-(11), the switching time is limited before or after the non-switchable time period.

[0207] In some embodiments, the material constraint is a core part of the model, involving two aspects of raw materials and products, including raw material storage, external supply and processing, product yield and demand, and raw material inventory restrictions, and the two aspects affect each other.

[0208] In some embodiments, on the raw material level, it is crucial for chemical enterprises to achieve daily material balance of liquid raw materials such as ethylene and propylene in the monthly plan. Raw material inventory is always in a fluctuating state due to the influence of storage, external supply, processing and other aspects. In particular, due to the different raw material consumption rates of different devices and different product grades, and the non-consumption of raw materials during grade switching due to shutdown and other factors, the fluctuation of raw material inventory is particularly frequent. Both excess and insufficient inventory will have an important impact on production, and chemical enterprises need to ensure that the end-of-day inventory of raw materials is within a reasonable range. On the other hand, in a monthly cycle and scheduling scheme that needs to be accurate to the hour, the start time and duration of each production event are continuous variables, and it is difficult to align at a unified time. Therefore, it is extremely difficult to perform daily material balance.

[0209] In some embodiments, the start time and duration of each batch can be established by establishing a constraint formula between the binary variables representing the three modes, and then establishing a calculation formula for the product yield of each batch per day, thereby indirectly obtaining the daily change in raw materials, and using an iterative method to obtain a valid solution within a specified time.

[0210] In some embodiments, the specific constraint formula includes raw material storage constraints, raw material supply constraints, raw material processing consumption constraints, and daily degree balance constraints of raw materials.

[0211] In some embodiments, the raw material storage constraint can be represented as:

[0212]

[0213]

[0214]

[0215] wherein, is the storage amount of raw material mr on day d, is the rate of batch b of raw material mr, is the duration of batch b of raw material mr on day d, is the total storage amount of raw material mr, , are the lower limit and upper limit of the storage rate of batch b of raw material mr, respectively.

[0216] Formula (12) means that the amount of raw material mr stored in d days is equal to the sum of the amount of each batch stored in d days, and formula (13) means that the amount of raw material mr stored is equal to the sum of the amount of each day stored. Formula (14) means that the rate of batch b of raw material mr stored should be between the upper and lower limits of the set rate.

[0217] In some embodiments, the raw material external supply constraint can be characterized as:

[0218]

[0219]

[0220]

[0221] wherein, is the amount of raw material mr supplied externally in d days, is the rate of batch b of raw material mr supplied externally, is the duration of batch b of raw material mr supplied externally in d days, is the total amount of raw material mr supplied externally, , are the lower limit and upper limit of the supply rate of batch b of raw material mr, respectively.

[0222] Formula (15) means that the amount of raw material mr supplied externally in d days is equal to the sum of the amount of each batch supplied externally in d days, and formula (16) means that the amount of raw material mr supplied externally is equal to the sum of the amount of each day supplied externally. Formula (17) means that the rate of batch b of raw material mr supplied externally should be between the upper and lower limits of the set rate.

[0223] In some embodiments, the raw material processing consumption constraint can be characterized as:

[0224]

[0225]

[0226]

[0227] wherein, is the amount of raw material mr consumed by the good mp1 produced by batch b of sequence s of device u in d days, is the yield of good mp1 produced by batch b of sequence s of device u in d days, is the unit consumption value of raw material mr of good mp1 in device u, is the amount of raw material mr consumed by the transition material mp4 produced by batch b of sequence s of device u in d days, is the yield of transition material mp4 produced by batch b of sequence s of device u in d days, the consumption of the raw material mr in day d, the consumption of the raw material mr in day d.

[0228] The formula (18) and (19) means that the amount of the raw material mr consumed by the good product mp1 or the transition material mp4 produced by the sequence s of the device u in the batch b in day d is equal to the product of the yield and the raw material unit consumption. The formula (20) means that the consumption of the raw material mr in day d is equal to the sum of the raw material consumed by each batch product of each device.

[0229] In some embodiments, the daily consumption balance constraint of the raw material can be characterized as:

[0230]

[0231]

[0232]

[0233]

[0234] wherein, the amount of the raw material mr stored in day d, the initial inventory of the raw material mr in day d, the amount of the raw material mr supplied externally in day d, the processing consumption of the raw material mr in day d, the final inventory of the raw material mr in day d, the final inventory of the raw material mr in day d-1, the final inventory of the raw material mr in day 1, the initial inventory of the raw material mr, and the lower limit and the upper limit of the inventory of the raw material mr.

[0235] The formula (21) means that the amount of the raw material mr stored in each day + the initial inventory = the amount supplied externally + the processing consumption + the final inventory, i.e. the daily material balance. The formula (22) means that the initial inventory of the raw material mr in day d is equal to the final inventory of the previous day, i.e. the linking constraint between inventories. The formula (23) means that the initial inventory of the raw material mr in day 1 is equal to the initial inventory in the whole period. The formula (24) means that the final inventory of the raw material mr in each day must be between the lower limit and the upper limit of the inventory.

[0236] In summary, through the formula (12)-(24), the correlation between the daily consumption of the raw material and the daily product yield of each batch of the device is established, and on this basis, the daily material balance series constraints of the raw material are established.

[0237] In some embodiments, since the starting time and the duration of each batch of the production device are variable, it is difficult to correspond to a specific date, and therefore, how to calculate the production of each batch in each day is a very difficult problem. In order to process batches spanning multiple days, three modes can be set: the batch is switched from another batch on the day, the batch lasts a whole day (i.e. 24 hours) on the day, and the batch is switched to another batch on the day. For this purpose, three auxiliary 0-1 variables are introduced to represent the three modes: wherein b_z1 (d) represents whether the batch b of sequence s of device u is mode one on day d, i.e. the batch is switched from another batch on the day, and the variable is 1, otherwise 0; b_z2 (d) represents whether the batch b of sequence s of device u is mode two on day d, i.e. the production duration of the day is 24 hours, and the variable is 1, otherwise 0; b_z3 (d) represents whether the batch b of sequence s of device u is mode three on day d, i.e. the batch is switched to the next batch on the day, and the variable is 1, otherwise 0. Wherein, if the batch switching occurs at the junction time of two days, such as the time of switching from batch A to batch B (i.e. the starting time of batch B) is 96 (unit: hour), then it is considered that the switching occurs on the 5th day, and on the 5th day, A’s b_z3 is equal to 1 and B’s b_z1 is equal to 1.

[0238] In addition, in order to establish constraints, two auxiliary binary variables , may be introduced.

[0239] The specific implementation method and constraint formula are as follows:

[0240]

[0241] The meaning of formula (25) is that the first batch on any day is 0, i.e. the first batch is not switched from another batch.

[0242]

[0243] In the above formula, represents the batch other than the first batch, i.e. the second and subsequent batches, is the earliest starting time of batch b of sequence s of device u, is the latest starting time of batch b of sequence s of device u.

[0244] The meaning of formula (26) is: for the second and subsequent batches, if the earliest starting time of this batch is greater than or equal to d*24, or the latest starting time is less than or equal to (d-1)*24, then the starting time of the batch must not be on day d. If the batch is not converted from another batch in day d, then b_z1 of the batch in day d is 0.

[0245]

[0246] In the above formula, is the earliest start time of batch b of sequence s of device u. The meaning of formula (26) is: for the second and subsequent batches, if the earliest start time of this batch is greater than (d-1)*24, then the duration in day d is certainly less than 24 hours, and then b_z2 of the batch in day d is 0, and an auxiliary variable is set to 0.

[0247]

[0248] In the above formula, is the latest end time of batch b of sequence s of device u. The meaning of formula (28) is: for any batch, if the latest end time of this batch is less than d*24, then the duration in day d is certainly less than 24 hours, and then b_z2 of the batch in day d is 0.

[0249]

[0250] The meaning of formula (29) is that the last batch in any day is 0, that is, the last batch cannot be converted into another batch.

[0251]

[0252] In the above formula, represents a batch other than the last batch, is the earliest end time of batch b of sequence s of device u, is the latest end time of batch b of sequence s of device u.

[0253] The meaning of formula (30) is: for a batch other than the last batch, if the earliest end time of this batch is greater than or equal to d*24, or the latest end time is less than or equal to (d-1)*24, then the end time of this batch is certainly not in day d, that is, it is not converted into another batch in day d, and then b_z3 of the batch in day d is 0.

[0254]

[0255]

[0256]

[0257]

[0258]

[0259]

[0260]

[0261]

[0262]

[0263]

[0264]

[0265]

[0266]

[0267]

[0268]

[0269]

[0270]

[0271]

[0272] In the above formulas (31)-(48), Let b be the duration of the batch b of the sequence s of device u. , , Let b be the duration of batch b of sequence s of device u in day d, corresponding to mode 1, mode 2, and mode 3. Let b be the start time of batch b in sequence s of device u. Let b be the duration of the batch b of the sequence s of device u. A binary variable to identify whether the sequence s of device u is selected. It is a very small number. It is a very large number.

[0273] Formulas (31) to (48) above are used to establish a correlation between the duration of the three modes of a batch and the start time and duration of the batch. More specifically: Formula (31) indicates that the production time of a certain batch of device u is the sum of the production time of the three modes on each day; Formulas (32) to (34) indicate that regardless of the mode, the production time of a certain batch on the first day is the sum of the production time of the three modes on each day. (35) - (48) are used to relate the start time of a batch to b_z1, b_z2, and b_z3, respectively. Formula (46) and (47) are used to relate the start time of a batch to b_z3. Formula (48) indicates that for a sequence of a device, the sum of b_z3 of any batch over all days is 1, i.e., there is only one day with b_z3 = 1, when the sequence is selected, and the sum of b_z3 of any batch over all days is 0, when the sequence is not selected.

[0274] In some embodiments, the production duration of each batch on each day according to different patterns can be obtained according to the above formulas (25) - (48), and the product yield of each batch on each day can be calculated in combination with the rate. The specific formulas are as follows:

[0275]

[0276]

[0277]

[0278]

[0279]

[0280]

[0281]

[0282]

[0283]

[0284] In the above formulas, , , are the yield of the good product mp1 produced by the batch b of the sequence s of the device u on the dth day according to the first pattern, the second pattern, and the third pattern, respectively, , , Transition material mp4 produced for batch b of sequence s of device u on day d according to mode one, mode two, mode three, , , respectively, the duration of batch b of sequence s of device u on day d according to mode one, mode two, mode three, the rate of producing good product mp1 for batch b of sequence s of device u, the rate guess of producing good product mp1 for batch b of sequence s of device u, the rate of producing transition material mp4 for batch b of sequence s of device u, , respectively, the rate guess of producing good product mp1 for batch b of sequence s of device u on day d according to mode one, mode three, , respectively, the lower limit and upper limit of the rate of good product mp1 of device u.

[0285] The above formulas (49), (50), (51) are respectively used to calculate the production of good product mp1 for batch b of sequence s of device u on day d according to mode one, mode two, mode three. In formulas (49) and (51), since the production rate and the duration are both variables, but the multiplication of two consecutive variables is a nonlinear term, in order to reduce the complexity of the model, Taylor expansion is used to linearize it, first give the initial rate and production duration, then through iteration on the solution level to continuously obtain the actual production rate and production duration that meet the tolerance requirements. Formula (52) is used to constrain the rate of good product mp1 of device u to be between the set upper and lower limits.

[0286] Formulas (53), (55), (55) are respectively used to calculate the production of transition material mp4 for batch b of sequence s of device u on day d according to mode one, mode two, mode three.

[0287] The meaning of formula (56) is that the amount of good product mp1 produced on day d for batch b of sequence s of device u is equal to the sum of the production according to the three modes.

[0288] The meaning of formula (57) is that the amount of transition material mp4 produced on day d for batch b of sequence s of device u is equal to the sum of the production according to the three modes.

[0289] Therefore, through the above formulas (25)~(57), the production of each product produced daily by each batch of each sequence of each device is obtained, which provides a basis for raw material daily calculation and total product yield calculation.

[0290] In some embodiments, a positive brand number can be constructed in the whole cycle of production, sales, and storage balance and constraint formula:

[0291]

[0292]

[0293]

[0294]

[0295] wherein, is the production of the positive product mp1 produced by batch b of sequence s of device u, is the production of the positive product mp1 produced by batch b of sequence s of device u on the dth day, is the total production of the positive product mp1, is the initial inventory of the positive product mp1, is the sales of the positive product, is the end-of-period inventory target value of the positive product mp1, and are the lower and upper limits of the saleable amount of the positive product mp1, respectively.

[0296] The meaning of formula (58) is that the production of the positive product mp1 produced by batch b of sequence s of device u is the sum of the production of each day of the batch.

[0297] The meaning of formula (59) is that the total production of the positive product mp1 is equal to the sum of the production of each batch of each device.

[0298] The meaning of formula (60) is that the production of the positive product mp1 + the initial inventory = the sales + the end-of-period inventory, i.e. the whole cycle material balance.

[0299] The meaning of formula (61) is that the sales of the positive product mp1 need to be between the upper and lower limits of the saleable amount.

[0300] In some embodiments, a transition material can be constructed in the whole cycle of production, sales, and storage balance and constraint formula:

[0301]

[0302]

[0303] wherein, is the production of the transition material mp4 produced by batch b of sequence s of device u, is the production of the transition material mp4 produced by batch b of sequence s of device u on the dth day, is the sales of the transition material mp4.

[0304] The meaning of formula (62) is that the production of the transition material mp4 of the batch b of the sequence s of the device u is the sum of the production of each day of the batch.

[0305] The meaning of formula (63) is that the total sales of the transition material mp4 is equal to the sum of the production of each batch of each device, that is, the transition material is considered to be all produced and all sold without considering the inventory factor.

[0306] In some embodiments, after establishing a model with formula (1)~(63) as constraints and maximizing the utilization rate of chemical raw materials as the objective, since the nonlinear terms (multiplication of two continuous variables: rate and duration) in the model have been linearized through formula (49) and formula (51), the constructed model is a mixed integer linear programming model, which can be solved using a commercial solver such as Gurobi. Then the values of the rate variable , the mode one duration variable , and the mode three duration variable are used to update the rate guess value , the mode one duration guess value , and the mode three duration guess value in formula (49) and formula (51), and the model is iteratively solved until the relative errors of the rate variable and the duration variable obtained by the previous and subsequent iterations are less than the set accuracy requirement (such as 0.001), and the iteration ends, obtaining the final optimization result.

[0307] A specific embodiment of the present specification is provided below:

[0308] For chemical material production devices A, B, and C, as shown in Table 6:

[0309] Table 6

[0310]

[0311] Using the above scheduling optimization method for optimization, the results are shown in Figure 4 and Figure 5 . Figure 4 The scheduling optimization results of the chemical material grades, three sets of devices have been switched 4 times, 4 times and 2 times, a total of 12 grades are selected from 36 grades for scheduling and production, and the switching time between grades meets the time window requirements in the white shift of working days. Figure 5 The daily balance result of the raw material ethylene, the model calculation result ensures that the daily ethylene inventory is between the high and low limits.

[0312] It can be seen from the scheduling optimization method of the chemical material provided in the embodiments of the present specification that the embodiments of the present specification can initialize a plurality of grade sequences corresponding to a plurality of chemical material production devices, each grade sequence including starting time, production rate and production duration of a plurality of batches, and each batch corresponding to a grade with specific physical and chemical properties; according to the production rate and production duration of each batch in each grade sequence, the daily output of each batch of chemical material is calculated; according to the daily output of each batch of chemical material and the physical property parameters, the loss index of the chemical material production device and the daily consumption and utilization rate of the chemical raw material are determined, and the loss index is used to represent the influence degree of producing each batch of chemical material on the service life of the chemical material production device; taking maximizing the utilization rate and minimizing the loss index as the target, and taking the daily consumption satisfying the daily inventory constraint and the starting time of each batch satisfying the switching time window constraint as the condition, a mixed integer linear programming model is constructed, and the mixed integer linear programming model includes a production rate linear term and a production duration linear term used to represent the daily output of each batch of chemical material; according to the mixed integer linear programming model, the plurality of grade sequences are optimized. By decomposing the grade sequence into a plurality of batches with accurately defined starting time, rate and duration, the conversion from macro planning to executable operation instruction is realized, and the traditional manual production scheduling problem is solved. By combining the daily output with the physical property parameters to quantitatively evaluate the device loss, the scheduling can actively balance the production task and the equipment health, delay the device aging from the source, and reduce the operation and maintenance risk. At the same time, by establishing the correlation between the daily output and the daily consumption, it is ensured that the daily inventory constraint can still be met under continuous time optimization, and the risk of raw material shortage or overflow is avoided. Taking maximizing the raw material utilization rate as the target, the model is driven to reduce the transition material and improve the yield. By introducing the rate and duration linear term to construct the mixed integer linear programming model, the complex nonlinear problem is converted into a form that can be efficiently solved. Finally, by cooperatively optimizing the raw material utilization rate maximization and the device loss minimization, and under the inventory and time window constraints, a multi-dimensional optimization model covering material, time and equipment state is constructed, and a comprehensive optimal and executable scheduling decision is realized.

[0313] Based on the above scheduling optimization method of the chemical material, the present specification further provides an embodiment of a scheduling optimization device for the chemical material. As shown in Figure 6 The scheduling optimization device 600 for the chemical material can specifically include the following modules:

[0314] The initialization module 601 is configured to initialize a plurality of grade sequences corresponding to a plurality of chemical material production devices, each grade sequence including starting time, production rate and production duration of a plurality of batches, and each batch corresponding to a grade with specific physical and chemical properties;

[0315] The calculation module 602 is configured to calculate the daily output of each batch of chemical material according to the production rate and production duration of each batch in each grade sequence;

[0316] The determination module 603 determines a loss index of the chemical material production device and a daily consumption and utilization rate of the chemical raw material according to daily production and physical property parameters of each batch of chemical material, and the loss index is used to represent the influence degree of producing each batch of chemical material on the service life of the chemical material production device.

[0317] The construction module 604 is configured to maximize the utilization rate and minimize the loss index, and construct a mixed integer linear programming model under the condition that the daily consumption meets the daily inventory constraint and the starting time of each batch meets the switching time window constraint, wherein the mixed integer linear programming model includes a production rate linear term and a production time length linear term used to represent the daily production of each batch of chemical material.

[0318] The optimization module 605 is configured to optimize the plurality of grade sequences according to the mixed integer linear programming model.

[0319] In some embodiments, the initialization module 601 can be specifically configured to:

[0320] construct a grade switching directed graph of each chemical material production device, wherein the grade switching directed graph includes a plurality of nodes and a plurality of directed edges, each node is associated with a grade, and each directed edge is used to represent a grade switching relationship between two nodes connected.

[0321] generate a grade sequence set of each chemical material production device based on the grade switching directed graph, and each grade sequence in the grade sequence set corresponds to a production scheme of the chemical material production device.

[0322] In some embodiments, the initialization module 601 can be specifically configured to:

[0323] select the grade sequence corresponding to each chemical material production device from the grade sequence set of each chemical material production device.

[0324] In some embodiments, the determination module 602 can be specifically configured to:

[0325] divide the production time length of each batch in each grade sequence into each natural day to obtain the daily production time length of each batch in each natural day;

[0326] determine the daily production of each batch in each natural day according to the daily production time length of each batch in each natural day and the production rate of each batch.

[0327] In some embodiments, the determination module 602 can be specifically configured to:

[0328] each batch is divided into a first mode production duration, one or more second mode production durations, and a third mode production duration according to natural days; the first mode production duration is a daily production duration of the batch in a natural day in which the batch starts but does not end; the second mode production duration is a daily production duration of the batch in a natural day in which the batch does not start and does not end; and the third mode production duration is a daily production duration of the batch in a natural day in which the batch does not start and ends.

[0329] In some embodiments, the production rate linear term includes a linear term based on a production rate linear coefficient corresponding to each batch; and the production duration linear term includes a linear term based on a first mode production duration linear coefficient corresponding to each batch and a linear term based on a third mode production duration linear coefficient corresponding to each batch.

[0330] In some embodiments, the determining module 603 can be specifically configured to:

[0331] According to the daily production of each batch of chemical materials and the unit consumption of each batch of chemical raw materials, the daily consumption of the chemical raw materials is determined.

[0332] In some embodiments, the determining module 603 can be specifically configured to:

[0333] According to the daily production of each batch of chemical materials corresponding to the positive brand number, the utilization rate of the chemical raw materials is determined.

[0334] Based on this, the constructing module 604 can be specifically configured to:

[0335] According to the first mode production duration, the second mode production duration, and the third mode production duration of each batch in each natural day and the production rate of each batch, the daily production is linearly processed using a Taylor expansion algorithm to obtain a first mode production duration linear coefficient corresponding to the first mode production duration, a third mode production duration linear coefficient corresponding to the third mode production duration, and a production rate linear coefficient corresponding to the production rate.

[0336] In some embodiments, the optimization module 605 can be specifically configured to: based on the first mode production duration linear coefficient, the third mode production duration linear coefficient, and the production rate linear coefficient corresponding to each batch, generate a plurality of new brand sequences of the plurality of chemical material production devices using a mixed integer linear programming model;

[0337] According to the first mode production duration, the third mode production duration, and the production rate of each batch in each new brand sequence, the first mode production duration linear coefficient, the third mode production duration linear coefficient, and the production rate linear coefficient corresponding to each batch are updated;

[0338] The steps of performing the above calculation of daily production, determining the loss index, daily consumption and utilization, producing the new brand sequence, and updating the first mode production duration linear coefficient, the third mode production duration linear coefficient and the production rate linear coefficient are iteratively performed until an optimization convergence condition is met.

[0339] In some embodiments, the optimization convergence condition includes that the production rate, the first mode production duration and the third mode production duration of each batch have a variation in the last N iterations that is less than a preset tolerance, N being a positive integer; or, the number of iterations reaches a preset maximum number of iterations.

[0340] In some embodiments, the condition that the daily consumption meets the daily inventory constraint and the start time of each batch meets the switching time window constraint can specifically include:

[0341] The condition that each brand sequence meets the sequence unique selection constraint, and the daily consumption meets the daily inventory constraint, and the start time of each batch in the brand sequence meets the switching time window constraint, and the production rate of each batch in the brand sequence meets the production capacity constraint, and the daily production of the chemical material of each batch in the brand sequence meets the demand constraint.

[0342] It can be seen from the scheduling optimization device for chemical materials provided by the above embodiments of the present specification that the embodiments of the present specification can initialize a plurality of grade sequences corresponding to a plurality of chemical material production devices, each grade sequence including a starting time, a production rate, and a production duration of a plurality of batches, and each batch corresponding to a grade having a specific physical and chemical property; according to the production rate and the production duration of each batch in each grade sequence, the daily output of each batch of chemical materials is calculated; according to the daily output of each batch of chemical materials and the physical property parameters, the loss index of the chemical material production device and the daily consumption and utilization rate of the chemical raw materials are determined, and the loss index is used to represent the influence degree of producing each batch of chemical materials on the service life of the chemical material production device; taking maximizing the utilization rate and minimizing the loss index as the target, and taking the daily consumption satisfying the daily inventory constraint and the starting time of each batch satisfying the switching time window constraint as the condition, a mixed integer linear programming model is constructed, and the mixed integer linear programming model includes a production rate linear term and a production duration linear term used to represent the daily output of each batch of chemical materials; according to the mixed integer linear programming model, the plurality of grade sequences are optimized. By decomposing the grade sequence into a plurality of batches with accurately defined starting time, rate, and duration, the transformation from macro planning to executable operation instructions is realized, and the traditional manual production scheduling problem is solved. By combining the daily output with the physical property parameters to quantitatively evaluate the device loss, the scheduling can actively balance the production task and the equipment health, delay the device aging from the source, and reduce the operation and maintenance risk. At the same time, by establishing the correlation between the daily output and the daily consumption, it is ensured that the daily inventory constraint can still be met under continuous time optimization, and the risk of raw material shortage or overflow is avoided. Taking maximizing the raw material utilization rate as the target, the model is driven to reduce the transition material and improve the yield. By introducing the rate and duration linear terms to construct the mixed integer linear programming model, the complex nonlinear problem is transformed into a form that can be efficiently solved. Finally, by cooperatively optimizing the raw material utilization rate maximization and the device loss minimization, and under the inventory and time window constraints, a multi-dimensional optimization model covering materials, time, and equipment state is constructed, and a comprehensive optimal and executable scheduling decision is realized.

[0343] The embodiment of the present specification further provides a computer device for scheduling optimization of chemical materials, comprising a processor and a memory for storing processor executable instructions, wherein the processor, when implemented, can perform the following tasks according to the instructions: initializing a plurality of grade sequences corresponding to a plurality of chemical material production devices, each grade sequence comprising starting time, production rate and production duration of a plurality of batches, and each batch corresponding to a grade having specific physical and chemical properties; calculating daily production of each batch of chemical materials according to the production rate and production duration of each batch in each grade sequence; determining the loss index of the chemical material production device and the daily consumption and utilization rate of the chemical raw materials according to the daily production of each batch of chemical materials and the physical property parameters, wherein the loss index is used to represent the influence degree of producing each batch of chemical materials on the service life of the chemical material production device; taking maximizing the utilization rate and minimizing the loss index as the target, and taking the daily consumption satisfying the daily inventory constraint and the starting time of each batch satisfying the switching time window constraint as the condition, constructing a mixed integer linear programming model, wherein the mixed integer linear programming model comprises a production rate linear term and a production duration linear term used to represent the daily production of each batch of chemical materials; and optimizing the plurality of grade sequences according to the mixed integer linear programming model.

[0344] In order to be able to more accurately complete the above instructions, referring to Figure 7 The embodiment of the present specification further provides another specific computer device 700, wherein the computer device 700 comprises a network communication port 701, a processor 702 and a memory 703, and the above structures are connected through internal cables so that each structure can specifically interact with data.

[0345] The processor 702 can be specifically used for: initializing a plurality of grade sequences corresponding to a plurality of chemical material production devices, each grade sequence comprising starting time, production rate and production duration of a plurality of batches, and each batch corresponding to a grade having specific physical and chemical properties; calculating daily production of each batch of chemical materials according to the production rate and production duration of each batch in each grade sequence; determining the loss index of the chemical material production device and the daily consumption and utilization rate of the chemical raw materials according to the daily production of each batch of chemical materials and the physical property parameters, wherein the loss index is used to represent the influence degree of producing each batch of chemical materials on the service life of the chemical material production device; taking maximizing the utilization rate and minimizing the loss index as the target, and taking the daily consumption satisfying the daily inventory constraint and the starting time of each batch satisfying the switching time window constraint as the condition, constructing a mixed integer linear programming model, wherein the mixed integer linear programming model comprises a production rate linear term and a production duration linear term used to represent the daily production of each batch of chemical materials; and optimizing the plurality of grade sequences according to the mixed integer linear programming model.

[0346] The memory 703 can be specifically used for storing corresponding instruction programs.

[0347] In the embodiment, the network communication port 701 can be a virtual port bound with different communication protocols, so as to send or receive different data. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for mail data communication. In addition, the network communication port can also be an entity communication interface or a communication chip. For example, it can be a wireless mobile network communication chip such as GSM, CDMA, etc.; it can also be a Wifi chip; it can also be a Bluetooth chip.

[0348] In the embodiment, the processor 702 can be implemented in any appropriate manner. For example, the processor can take the form of a microprocessor or processor and a computer readable medium storing computer readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, Application Specific Integrated Circuit (ASIC), programmable logic controller, and embedded microcontroller, etc. The present specification does not make any limitation.

[0349] In the embodiment, the memory 703 includes volatile memory and non-volatile memory. The memory 703 can include multiple levels, and in a digital system, as long as it can save binary data, it can be a memory; in an integrated circuit, a circuit without a physical form having a storage function is also called a memory, such as RAM, FIFO, etc.; in a system, a storage device with a physical form is also called a memory, such as a memory stick, a TF card, etc.

[0350] In another aspect, the embodiments of the present specification also provide a computer readable storage medium, the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the instructions of the method described above. Figure 1

[0351] It should be understood that in various embodiments of the present specification, the size of the sequence number of each process described above does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present specification.

[0352] It should also be understood that in the embodiments of the present specification, the term and / or is only an associated relationship between the associated objects, indicating that there can be three relationships. For example, A and / or B can represent three cases of A alone, A and B together, and B alone. In addition, the character / in the present specification generally represents the relationship of one or the other of the associated objects before and after.

[0353] ​Those skilled in the art will appreciate that embodiments of the present application can be readily used as software, hardware, or a combination of software and hardware. In a software embodiment, the present application can be implemented with computer programs (also referred to as software instructions, software code, computer code, and the like) that execute on programmable hardware including computer processors, digital signal processors, microprocessors, central processing units, microcontrollers, programmable hardware logic devices, and the like. Generally, the present application can be implemented in hardware, software, or any combination of hardware and software.

[0354] The present application is described in relation to flow diagrams and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present application. It is to be understood that each block of the flow diagrams and / or block diagrams, and combinations of blocks in the flow diagrams and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing device or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 one or more functions specified by one or more blocks Figure 1 one or more functions specified by one or more blocks.

[0355] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the flow diagrams and / or block diagrams block or blocks. Figure 1 one or more functions specified by one or more blocks Figure 1 one or more functions specified by one or more blocks.

[0356] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational tasks to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide tasks that implement the flow diagrams and / or block diagrams block or blocks. Figure 1 one or more functions specified by one or more blocks ​ one or more functions specified by one or more blocks.

[0357] The above description is provided as an example of specific embodiments of the present application. It is to be understood that the above description is merely descriptive of certain embodiments of the present application, and is not intended to limit the scope of the present application. Any modification, equivalent replacement, improvement, and the like made within the spirit and principle of the present application shall be included in the scope of the present application.

Claims

1. A method for optimizing the scheduling of chemical materials, characterized in that, include: Initialize multiple grade sequences corresponding to multiple chemical material production units. Each grade sequence includes the start time, production rate and production duration of multiple batches. Each batch corresponds to a grade with specific physicochemical properties. The production duration of each batch in each brand sequence is decomposed into a first-mode production duration, one or more second-mode production durations, and a third-mode production duration by calendar day; the first-mode production duration is the daily production duration when the batch starts on a certain calendar day but does not end on that calendar day; the second-mode production duration is the daily production duration when the batch does not start on a certain calendar day but does not end on that calendar day; the third-mode production duration is the daily production duration when the batch does not start on a certain calendar day but ends on that calendar day. The daily output of each batch on each natural day is determined based on the daily production time of each batch on each natural day and the production rate of each batch. Based on the daily output and physical property parameters of each batch of chemical materials, the loss index of the chemical material production unit and the daily consumption and utilization rate of chemical raw materials are determined. The loss index is used to characterize the degree of impact of producing each batch of chemical materials on the life of the chemical material production unit. With the goal of maximizing the utilization rate and minimizing the loss index, and assuming that the daily consumption meets the daily inventory constraint and the start time of each batch meets the switching window constraint, a mixed-integer linear programming model is constructed. This includes: linearizing the daily output using a Taylor expansion algorithm based on the first mode production time, second mode production time, and third mode production time of each batch on each natural day, as well as the production rate of each batch, to obtain the linear coefficients of the first mode production time, the third mode production time, and the production rate; the production rate linear coefficient is related to the production time; the first mode production time linear coefficient and the third mode production time linear coefficient are related to the production rate; the mixed-integer linear programming model includes a production rate linear term and a production time linear term to characterize the daily output of each batch of chemical materials; the production rate linear term includes linear terms based on the production rate linear coefficient for each batch; the production time linear term includes linear terms based on the first mode production time linear coefficient and the third mode production time linear coefficient for each batch; The multiple brand number sequences are optimized based on the mixed integer linear programming model.

2. The method according to claim 1, characterized in that, The method further includes: Construct a directed graph for grade switching for each chemical material production unit. The directed graph for grade switching includes multiple nodes and multiple directed edges. Each node is associated with a grade, and each directed edge is used to represent the grade switching relationship between the two connected nodes. Based on the directed graph of grade switching, a set of grade sequences is generated for each chemical material production unit, and each grade sequence in the set corresponds to a production scheme for the corresponding chemical material production unit.

3. The method according to claim 2, characterized in that, The initialization of multiple grade sequences corresponding to multiple chemical material production devices includes: Select the corresponding grade sequence for each chemical material production unit from the set of grade sequence for each chemical material production unit.

4. The method according to claim 1, characterized in that, The optimization of the multiple brand number sequences based on the mixed-integer linear programming model includes: Based on the linear coefficients of the first mode production time, the third mode production time, and the production rate corresponding to each batch, a mixed integer linear programming model is used to generate multiple new grade sequences for the multiple chemical material production devices. Based on the first mode production time, third mode production time and production rate corresponding to each batch in each new grade sequence, update the first mode production time linear coefficient, third mode production time linear coefficient and production rate linear coefficient corresponding to each batch. The process involves iteratively executing the steps described above: calculating daily output, determining the loss index, daily consumption and utilization rate, producing new grade sequences, and updating the linear coefficients of production time for the first mode, production time for the third mode, and production rate, until the optimization convergence condition is met.

5. The method according to claim 4, characterized in that, The optimization convergence conditions include: The changes in the production rate of each batch, the production time of the first mode, and the production time of the third mode in N consecutive iterations are all less than the preset tolerance, where N is a positive integer; or, the number of iterations reaches the preset maximum number of iterations.

6. The method according to claim 1, characterized in that, The determination of the loss index of chemical material production equipment includes: Based on the physical properties of each batch of chemical materials, calculate the unit loss caused by each batch of chemical materials to the chemical material production equipment; Based on the unit loss and the daily output of each batch of chemical materials, the loss index of each chemical material production unit is determined. Based on the loss index of each chemical material production unit, the loss fluctuation index of the multiple chemical material production units is calculated. The loss fluctuation index is used to characterize the differences in the degree of life impact among the multiple chemical material production units. The objective of maximizing the utilization rate and minimizing the loss index includes: The objective is to maximize the utilization rate and minimize the total loss index of multiple chemical material production units and the loss fluctuation index.

7. The method according to claim 1, characterized in that, The determination of the loss index of chemical material production equipment and the daily consumption and utilization rate of chemical raw materials includes: The daily consumption of chemical raw materials is determined based on the daily output of each batch of chemical materials and the unit consumption of each batch of chemical raw materials.

8. The method according to claim 1, characterized in that, The grade designation includes a primary grade designation and a transitional grade designation. The primary grade designation corresponds to the target chemical material product of the chemical material production unit, while the transitional grade designation corresponds to the non-target chemical material product of the chemical material production unit. The determination of the loss index of chemical material production equipment and the daily consumption and utilization rate of chemical raw materials includes: The utilization rate of chemical raw materials is determined based on the daily output of each batch of chemical materials corresponding to the brand name.

9. The method according to claim 1, characterized in that, The method further includes: Determine the unit energy consumption for each batch of chemical materials produced by each chemical material production unit; Based on the unit energy consumption and the production rate and production time of each batch, the energy consumption index of each chemical material production unit is determined; The condition that the daily consumption meets the daily inventory constraint and the start time of each batch meets the switching window constraint includes: The conditions are that the daily consumption meets the daily inventory constraint, the start time of each batch meets the switching window constraint, and the energy consumption index of each chemical material production unit meets the energy consumption constraint.

10. The method according to claim 1, characterized in that, The condition that the daily consumption meets the daily inventory constraint and the start time of each batch meets the switching window constraint includes: The conditions are as follows: each grade sequence satisfies the sequence unique selection constraint, the daily consumption satisfies the daily inventory constraint, the start time of each batch in the grade sequence satisfies the switching window constraint, the production rate of each batch in the grade sequence satisfies the production capacity constraint, and the daily output of each batch of chemical materials in the grade sequence satisfies the demand constraint.

11. A scheduling and optimization device for chemical materials, characterized in that, include: The initialization module is used to initialize multiple grade sequences corresponding to multiple chemical material production units. Each grade sequence includes the start time, production rate and production duration of multiple batches. Each batch corresponds to a grade with specific physicochemical properties. The calculation module is used to decompose the production time of each batch in each brand sequence into a first mode production time, one or more second mode production times, and a third mode production time by natural day; the first mode production time is the daily production time of the batch that starts on a certain natural day and does not end on that natural day; the second mode production time is the daily production time of the batch that does not start on a certain natural day and does not end on that natural day; the third mode production time is the daily production time of the batch that does not start on a certain natural day and ends on that natural day; based on the daily production time of each batch on each natural day and the production rate of each batch, the daily output of each batch on each natural day is determined; The determination module is used to determine the loss index of the chemical material production unit and the daily consumption and utilization rate of chemical raw materials based on the daily output and physical property parameters of each batch of chemical materials. The loss index is used to characterize the degree of impact of producing each batch of chemical materials on the life of the chemical material production unit. A construction module is used to construct a mixed-integer linear programming model with the objective of maximizing the utilization rate and minimizing the loss index, and with the conditions that the daily consumption meets the daily inventory constraint and the start time of each batch meets the switching window constraint. This includes: linearizing the daily output using a Taylor expansion algorithm based on the first mode production time, second mode production time, and third mode production time of each batch on each natural day, as well as the production rate of each batch, to obtain the linear coefficients of the first mode production time, the third mode production time, and the production rate; the production rate linear coefficient is related to the production time; the first mode production time linear coefficient and the third mode production time linear coefficient are related to the production rate; the mixed-integer linear programming model includes a production rate linear term and a production time linear term to characterize the daily output of each batch of chemical materials; the production rate linear term includes linear terms based on the production rate linear coefficient for each batch; the production time linear term includes linear terms based on the first mode production time linear coefficient and the third mode production time linear coefficient for each batch; An optimization module is used to optimize the multiple brand number sequences according to the mixed integer linear programming model.

12. A computer device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the method of any one of claims 1-10.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method as described in any one of claims 1 to 10.

Images