A steel production scheduling rule engine platform

Abstract

The application discloses a steel production scheduling rule engine platform and belongs to the technical field of steel production scheduling.The platform comprises a rule definition module, a real-time reading and writing module, a database reading and writing module, a rule execution module, a post-confirmation module and a rule evaluation module.The rule definition module, the post-confirmation module and the rule evaluation module are integrated on a Web server supporting a RestFul access interface in the form of Web components, and an interactive interface is provided for users; the rule execution module is separately arranged as an independent executable program and accesses working condition data through the real-time reading and writing module; and the real-time reading and writing module and the database reading and writing module exist as public components and are built-in in the Web server and the rule execution program.The application has the advantages that the above-mentioned problems, such as logical confusion in rule execution, insufficient customization and flexibility of an execution program, a long development cycle, lack of a linkage mechanism for rule execution of multiple processes and inconsistency, and lack of a rule evaluation module leading to insufficient evaluation of execution results of a scheduling system, are solved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of steel production scheduling, and particularly provides a steel production scheduling rule engine platform. The software platform is used for scheduling complex working condition events in the steel production process through a rule engine, and comprehensively uses complex event processing, working condition scheduling period, post-confirmation and scheduling evaluation indexes, and is used for assisting a production scheduling software module to intelligently judge and decide, and can also be used for post-reversing of the production process. BACKGROUND

[0002] RestFul interface

[0003] RestFul is a Web service interaction specification, and compared with other Web interaction specifications (such as SOAP [Simple Object protocol, Simple Object Access Protocol], XML-RPC [uses Http protocol as a transmission protocol, and data is in XML text mode]) and service interfaces, is more lightweight and faces the Http protocol. A client uses four verbs representing operation modes (GET, POST, PUT and DELETE) to operate resources on a server: GET is used to obtain resources, POST is used to newly build resources (and can also be used to update resources), PUT is used to update resources, and DELETE is used to delete resources; resources are operated through the form of expression of resources; the form of expression of resources is XML or HTML; the interaction between the client and the server is stateless between requests, and each request from the client to the server must contain information necessary for understanding the request.

[0004] Device working condition

[0005] A device working condition refers to a working state of a device under a condition directly related to the action of the device, and a condition or process condition in a production process is referred to as a working condition of the device. The working condition of the device is generally related to the relevant state of the device, and is related to some control signal variables when implemented to a control system, and these signals are referred to as device working condition signals. When the device working condition changes, some events are generally triggered to notify a related secondary system or a related device management system. The related secondary system can execute some production scheduling tasks by monitoring and evaluating the device working condition signals, in combination with the production situation on a production line and related process systems and constraint conditions.

[0006] Steel production scheduling

[0007] Steel production enterprises belong to mixed manufacturing enterprises, and steel production has its own logistics characteristics and production methods: (1) The production process has distinct stages: the smelting stage is mainly chemical change, and the key is to adjust the chemical composition of the product to ensure physical properties; the casting and rolling stage is mainly physical change, and the product's phase change and shape and size need to be considered, and the connection between adjacent production plans needs to be considered; (2) Production has concentration, and the production capacity of the enterprise is reflected in some key equipment such as rolling mills and steel furnaces, and the production capacity and connection capacity of the equipment need to be considered; (3) Part of the production has continuity, from iron (steel) smelting to casting and rolling, the entire stage is carried out in a production line manner, which requires high temperature and high energy consumption, and the product has continuity characteristics in the production process, and cannot be easily interrupted, otherwise the product may be scrapped or the product may be stacked and energy wasted.

[0008] The key to steel production scheduling is to reasonably allocate production around key equipment and distinguish different production stages to ensure safe, stable, smooth logistics, and maximize the production capacity of bottleneck equipment, improve equipment utilization, and achieve full-load, high-quality, and efficient production of enterprises.

[0009] Complex event processing (CEP)

[0010] Event processing is a method of tracking and analyzing information about events that occur, and complex event processing (CEP) consists of a set of concepts and technologies developed in the early 1990s to process real-time events and extract information from event streams as they arrive. The goal of complex event processing is to identify meaningful events (such as opportunities and threats) in real time and respond in real time. CEP helps check the situation through declarative language (event processing language) or other abstractions, making it easier and faster to detect situations. CEP is called complex event processing because it can detect specific event combinations in event streams and process them, combining individual simple events into a "complex event" through certain rule matching, and then converting and processing these complex events that meet the rules to get the desired results for output. In summary, the process of complex event processing (CEP) can generally be divided into three steps:

[0011] (1) Define a matching rule

[0012] (2) Apply the matching rule to the event stream to detect complex conditions that meet the rule.

[0013] (3) Process the detected complex event to get the result for output.

[0014] Production cycle

[0015] Steel production generally has periodic properties, steel production is generally organized according to the process, the production of each process includes tool preparation, material receiving, material loading, power on, running, shutdown and other steps, which have strict time constraints in production process, and production is often arranged according to customer order demand and production cycle time when planning production, so the periodic properties of steel production need to be considered when scheduling equipment production.

[0016] Rule engine

[0017] The rule engine is an important module of production scheduling, which is a complex logic processing engine developed from the inference engine. It is a component embedded in the application program, which realizes the separation of business decision from the application program code and uses the pre-defined semantic module to write business decision. In the production scheduling system mentioned in the application, the rule engine can help to realize the definition, execution and result evaluation of the scheduling rules based on the production conditions. Through the rule engine, the hard-coded logic in the production scheduling system is reduced. By defining the scheduling rules, statistical index rules, event triggering rules and evaluation rules in the production process, the system designer can quickly and efficiently develop the business logic in the production scheduling system, and improve the flexibility of the system.

[0018] Process path

[0019] The process route, English is Routing, is a technical document describing the operation sequence of material processing and part assembly, which is a sequence of multiple processes. A process is an action or a series of actions taken by production workers or machines to complete a specified task, which is the most basic processing operation mode of processing materials and assembling products, and is directly related to location information such as work center and external supplier. It is the basic unit of process route. For example, a production line is a process route, which contains many processes.

[0020] Surplus material acts as

[0021] In production, raw materials need to be selected for production processes. This raw material may be purchased or may be a semi-finished product produced by the previous process. By selecting existing semi-finished materials as raw materials for processing, the material can be sent to the next process immediately after production is completed, reducing additional inventory. This process is surplus material acting.

[0022] Gantt chart

[0023] Gantt chart is actually a line chart, which was developed by Henry Gantt in the early 20th century, hence the name Gantt chart. A general Gantt chart uses the horizontal axis to represent time and the vertical axis to represent activities or projects. The lines in the chart are used to represent the arrangement and completion of activities during the planning period. Through the Gantt chart, the time rhythm and state of each task and activity on the production line can be described.

[0024] Production process data set

[0025] The production process data set is a self-service data model specially constructed for production process analysis and visualization, which generally corresponds to data of standard production process information, including various signals, input and output materials, process parameters and procedure information in the production process. Through the production process data set, the production process can be replayed, the relevant rules of production process scheduling can be discovered, and the production process optimization can be guided.

[0026] In the existing steel production scheduling system, the equipment working condition is complex, the rule information lacks unified definition, the production element information is numerous and disordered, the corresponding steel production scheduling program needs to be specially developed for a specific production scene, lacks a unified rule processing engine, which leads to the lack of flexibility of the corresponding scheduling system, and the repeated development of programs for similar rules in different fields, increasing the software development cost and cycle; when performing rule scheduling, each process needs to develop a corresponding scheduling program for the scheduling rules and production data of the process, and the production scheduling between processes lacks a linkage mechanism, leading to the lack of consistency between the scheduling operations of different processes; when performing production scheduling, there is a lack of periodic calculation indicators, leading to inaccurate production rhythm during scheduling calculation, and the scheduling algorithm falls into the current optimal but globally poor condition; for the execution result of the production scheduling rule, there is a lack of good evaluation indicators and interactive interface, leading to insufficient evaluation of the function and efficiency of the scheduling system by the operator. SUMMARY

[0027] In the original steel production scheduling system, the equipment working condition is complex, the rule information lacks unified definition, the production element information is numerous and disordered, the corresponding steel production scheduling program needs to be specially developed for a specific production scene, lacks a unified rule processing engine, which leads to the lack of flexibility of the corresponding scheduling system, and the repeated development of programs for similar rules in different fields, increasing the software development cost and cycle; when performing rule scheduling, each process needs to develop a corresponding scheduling program for the scheduling rules and production data of the process, and the production scheduling between processes lacks a linkage mechanism, leading to the lack of consistency between the scheduling operations of different processes; when performing production scheduling, there is a lack of periodic calculation indicators, leading to inaccurate production rhythm during scheduling calculation, and the scheduling algorithm falls into the current optimal but globally poor condition; for the execution result of the production scheduling rule, there is a lack of good evaluation indicators and interactive interface, leading to insufficient evaluation of the function and efficiency of the scheduling system by the operator.

[0028] The present application aims to provide a steel production scheduling rule engine platform, which aims to solve the above-mentioned problems of logical confusion in rule execution, insufficient customization and flexibility of execution program, long software development cycle, lack of linkage mechanism for rule execution of multiple processes, lack of consistency, lack of rule evaluation module, and insufficient evaluation of scheduling system execution results, etc. By introducing the working condition event of the equipment into the software platform framework of complex event processing, the actual working condition of each equipment in each process of steel production can be combined with the current process production scheduling plan rules, periodic indicators, raw materials, in-process products, product and equipment constraint rules, real-time calculation by the rule engine, generation of real-time scheduling scheme, and delivery to the scheduling system for execution, which can be confirmed by the user as needed. In order to ensure the accuracy of the scheduling scheme, periodic calculation indicators are introduced into the platform, which can calculate the entire production rhythm and equipment state according to the periodicity of each process and schedule the production time; the rule engine platform can perform event processing logic for collaborative production scheduling of multiple production sequences according to the process connection relationship, each production sequence can share event processing signals, automatically calculate the priority of each production sequence, and perform linkage operation of rule calculation of multiple processes when the working condition event changes, to ensure the orderly execution of the production process. In addition, the rule engine platform saves all production signals and actual scheduling information, which can help users compare the platform calculation working condition with the actual working condition, can call evaluation indicators to evaluate the accuracy of the scheduling result, and can assist system developers in optimizing the scheduling program.

[0029] The platform comprises a rule definition module, a real-time read-write module, a database read-write module, a rule execution module, a post-confirmation module and a rule evaluation module. The rule definition module is mainly connected with the database read-write module to perform definition and configuration of rules. The post-confirmation module and the rule evaluation module are similar, and are respectively connected with the database read-write module and the real-time read-write module to read rule configuration and event information and process the event information. The rule execution module is a rule processing engine, which is generally a separate module connected with the database read-write module and the real-time read-write module to read rule configuration, perform rule calculation and store event records into real-time data. The rule definition module, the post-confirmation module and the rule evaluation module are integrated on a special Web server supporting Http RestFul access interface in the form of Web modules to provide an interactive interface for users, perform rule definition, post-confirmation, result evaluation and other functions, and call each other through RestFul interfaces. The rule execution module is separately deployed as an independent executable program, referred to as a rule execution program, which internally has a complex event rule engine responsible for rule execution functions of various processes, reads rule configuration through the database read-write module, accesses working condition data through the real-time read-write module, performs rule calculation, and writes the generated event records into the real-time database through the interface. The real-time read-write module and the database read-write module exist as public components and are internally built in the Web server and the rule execution program to provide real-time database access interfaces and database access interfaces for rule execution.

[0030] (1) Rule definition module: first, an interactive screen is provided for users to define working condition information, complex event rule definition and periodic index rule definition of multiple processes, and the rule definition information is stored in a database, which is provided to the rule engine module for loading through a database access interface in later use.

[0031] The main function of the rule definition module is to provide an interactive rule definition interface for users and be responsible for rule definition. The rule definition interface provides RestFul services in the background, and the rule definition client and the user side can remotely edit and save rule definition and access rule definition through the RestFul service interface; the configured rules meet the specifications of complex event processing and are stored in a special rule configuration database, which can be loaded from the database before use.

[0032] The mapping relationship from various working condition signals to event in complex event can be defined in the rule definition module. For simple working condition signal, the working condition signal and the corresponding event execution action, timeout time, execution frequency and execution time constraint can be mapped; for combined working condition signal, multiple different event objects can be mapped, the logical relationship of multiple combined working condition signals is mapped to the composite event object in complex event processing, and the corresponding event processing interface is defined again for the composite event object; the mapping of production process data set to complex event processing object can also be defined, the mapping relationship of data feature mode of process data set and event processing logic can be analyzed in the rule definition stage, and different data feature modes are mapped to different event processing objects; when the actual production data comes, the relationship between historical working condition signal and data set can be calculated, and the corresponding working condition event processing logic is executed.

[0033] The rule definition module is deployed on a separate Web server and interacts with remote Web users through Http Restful protocol.

[0034] The rule information configured in the rule definition module includes: (1) working condition signal calculation rule. Used for calculating the working condition state of the equipment, generally configured as a conditional expression tree, each node is a binary tree node, the root node of the binary tree is a conditional connection predicate AND, OR, NOT, NOT is a negation operation, the right node of the binary tree is empty, only the left node is reserved; then a complex expression tree can be constructed; (2) steel production process path rule. Each steel has its own unique process path, which can be divided into different sections, and then each section is composed of different equipment. For each steel product, there are relevant process requirements on each equipment, including production time, temperature, composition, size, supporting equipment and other constraints; the entire life cycle production process path can be described by a graph node; (3) scheduling and batching rules, generally provide corresponding steel, then integrate finished product specification steel information, order quantity information, remaining material quantity, production time, composition information, size information, processing equipment and other information; (5) equipment capacity constraint rule, including the maximum capacity and load of a single equipment, energy consumption constraint, equipment power constraint, to determine the throughput rate of the equipment per unit time; (6) equipment maintenance and rotation rule; through the equipment maintenance rule, it can be determined which equipment is available and which equipment is down during scheduling, and the appropriate equipment is selected to meet the production demand; (7) equipment furnace grouping and casting grouping rule, the purpose is to improve the utilization rate of equipment. In the case of batch production, the raw materials of adjacent steel can be batched together for production, one furnace of molten iron in the steelmaking process may correspond to multiple steel grades, and one pouring time in the casting process may correspond to multiple users, but the steel grades need to be similar. (8) Periodic rule: including the standard production time and period of each equipment, such as the furnace change time of steelmaking, the roll change time of continuous rolling process, and the period of heating and annealing process. Through the periodic rule of the equipment, the periodic cycle processing semantics is added to the complex event processing, and the corresponding equipment event information can be generated periodically according to the periodic time and added to the semantic chain of the complex event processing.

[0035] (2) Real-time read-write module: a basic component, mainly provided for the rule execution module, can read the production signals and working condition information of the current process and associated process equipment from the real-time database, and send the calculated working condition scheduling results to the real-time database.

[0036] The real-time read-write module is used for communication with the real-time database to provide the rule engine with the ability to read the real-time / historical curve data of the equipment working condition. The real-time module is generally deployed together with the rule execution module in the form of a component.

[0037] The equipment working condition information includes the running status and related equipment parameters of each equipment in the steel production line, such as electrical signals of the equipment such as current, voltage, power, and electrical quantity.

[0038] The historical curve access supports data reading of device working condition information and related device parameters and the like information from real time to various periods of data, and provides real time data support for calculation of the rule engine.

[0039] (3) Database read-write module: a basic component, providing 1) configuration information of various rules; 2) device working condition information configuration, including device working condition naming, state, calculation formula and trigger event information; 3) read-write interface of a production database, the production data mainly including a model of a certain steel production line, and production data information of the line such as operators, orders, materials, plans, inventory and the like;

[0040] The database read-write module provides a database read-write interface for the Web server and the rule execution program by providing read-write functions of production information and configuration information.

[0041] (4) Rule execution module: the module reads the latest device working condition information of a process through the real-time read-write module, then accesses information such as materials, plans, process definitions, operator information, product specifications and quality in production scheduling through the database module, calls a complex rule execution engine, and processes working condition events in real time to generate scheduling records and send them to the production scheduling system, and meanwhile archives corresponding scheduling event records in the database.

[0042] The rule execution module is responsible for executing rules, which is based on complex events and implemented by using Java programming language. In the initialization stage, all rule definition information is loaded through the database, and then working condition events are processed according to the complex event processing rules.

[0043] For the complex event processing of the working condition type, the following features are included: generally, multiple combined signals are included, including flow signals, pulse signals, start and stop signals, fault signals, alarm signals, etc. For different signals, different preprocessing methods are adopted. First, the data is transformed to generate time-varying event information, and then the complex event engine is called to comprehensively judge the working condition information. For example, the flow signal needs to be quantized according to the interval first, and then the data sampling interval, the sampling method, the threshold of the flow signal are specified, and sometimes the rules for calculating historical data need to be specified, and the signal point statistical value of a certain time interval needs to be read from the real-time database to calculate the event value. The pulse signal is generally a short trigger signal, which is represented as a sharp peak on the curve graph, and is judged according to the peak signal; the start and stop signal is generally a simple state; the fault signal and the alarm signal remain true when they occur, and the signal resets after the fault and the alarm are eliminated. A complex event processing engine is built in this module to map the calculation of the working condition signal to the mode of the complex event to realize the calculation and processing process of the working condition signal. In some cases, the scheduling needs to be combined with the mode of the historical production data set to calculate the scheduling rule, and when the working condition signal is judged, the signal historical data curve of a period of time needs to be read to calculate the related working condition mode data, and then the data mode of the corresponding historical data set is matched to obtain the corresponding event logic, generate the corresponding simple event, and then calculate the corresponding event processing logic combined with the processing logic of the combined event.

[0044] In the system, various signal-triggered scheduling rule information has been built in advance, and the rule execution module reads and writes the information through the real-time reading and writing module, reading the latest process information of the device condition from the real-time database, and The information of the operators, the multi-batch materials, the plans, the orders, the inventories, etc. is read through the database reading and writing module to prepare for generating real-time scheduling instructions.

[0045] Before executing real-time production scheduling instructions, the rule engine also needs to combine the production process path rules of the product steel grade, the delivery cycle of the order, the production scheduling and ingredient rules, the quantity of inventory and surplus material, combine the equipment group furnace group pouring rules, and generate the production scheduling plan of the current production batch. The algorithm used is a general multi-objective optimization algorithm to achieve this. In the case of multiple processes, each process and the downstream process are in a supply-customer relationship. When solving, according to the demand of the sales order, the quantity of the product in the current process is converted to the output of the final product by backtracking from the finished product process to the front process. Then, the theoretical total production quantity is obtained by subtracting the inventory of all products on the current production line. Then, according to the product quantity conversion relationship, the total output of each sequence in the new production scheduling cycle is obtained. Then, according to the order demand, the furnace group pouring operation is executed to generate a production scheduling plan for each production sequence according to time and priority. Then, according to the surplus material acting rule, the appropriate materials on the production line are selected for dynamic ingredient allocation of the production instruction. Then, the system generates actual production instructions for each production phase according to the equipment operation rules and the production ingredient information, and sends them to the device control module. In the case of device working condition readiness, the execution can be performed. In the production execution requiring human intervention, the rule triggering program needs to combine the artificial confirmation signal and the device working condition signal to execute.

[0046] For the surplus material acting rule problem, there is great flexibility in production practice. In the system, the product process attribute and product characteristics are used to generate a material demand keyword. This keyword can integrate the steel grade code, product specification, product purpose, and quality composition parameters to generate a product process degree matching index. In dynamic production scheduling, the rule engine calculates the matching degree between product material demand and inventory material based on the product process degree matching index to perform dynamic matching and provide a material feeding plan for the operators on site.

[0047] The rule engine platform can perform production scheduling event processing logic for multiple production sequences according to the priority of the process connection relationship. Each production sequence can share event processing signals, and multiple processes can be linked and operated when production events arrive, ensuring that production scheduling rules are executed in the order of the process. Each process uses a tuple of (steel grade, process ID, event type, [triggering process], execution time) to identify the execution condition. The triggering process is an array. Only when the processes in the triggering process have completed execution will the corresponding process rule be triggered for execution. In order to improve the execution efficiency of the system, during initialization, the priority of the corresponding process for each process signal rule execution needs to be calculated first. The method is to sort the processes according to the process triggering relationship between the processes, and then identify the priority of the processes as different numbers according to the order. Then, when each process signal arrives, it is executed from high to low priority.

[0048] (5)Post-confirmation module, according to the scheduling event record, according to the post-confirmation request, the confirmation window is popped up, the working condition information, the plan information, the material information and the generated scheduling information are displayed on the Web window, and the user confirms whether the scheduling result is correct and the corresponding error information, and the module is a supplement to the real-time scheduling result information.

[0049] The post-confirmation module pops up a confirmation window according to the event record, and displays working condition information, plan information, time information, material information and generated scheduling information on the window, so that the user confirms whether the scheduling result meets the needs. For the case that the production record and the actual result are inconsistent, according to the error and delay of the signal, the current module adjusts the time rhythm of all production sequences according to the definition of the event and the value of the delay, for example, the start time of all production processes in the current stage is increased by an offset, and then the corresponding production data is corrected; if there is an error in the metering of the material, a relatively large loss occurs, and the process parameters and the yield need to be confirmed.

[0050] The post-confirmation module and the rule definition module are deployed in the Web server, and provide a remote Http access interface for user access.

[0051] Production post-confirmation is completed by production workers, and is used for confirming the actual working condition of the production process. Through the comparison of the actual working condition information and the signal judged by the complex event, the system problem is determined, and the supplement of the production data and the working condition information by the supplementary recording information is provided for the post-production process tracing.

[0052] (6) Rule evaluation module, which reads and writes event records through the real-time data reading and writing module, obtains complex event rules and historical records processed according to the complex event rules through a database reading and writing interface, and evaluates the execution time, accuracy and timeliness of the rules.

[0053] The function of the rule evaluation module is to compare the historical records processed according to the complex event rules with the production records confirmed by the production, and evaluate the execution time, accuracy and timeliness of the rules.

[0054] In the subsequent production process tracing, the execution of the complex event can be replayed by calling the rule engine at any time point according to the working condition signal information stored in the real-time database and the historical records of the execution of the complex event, the difference between the event results calculated by the rule engine and the historical event execution records can be compared, and the related parameters of the complex event processing can be adjusted in the subsequent scheduling execution process, so that the execution algorithm is more accurate and timely.

[0055] The innovation of the application is:

[0056] The present application aims to provide a steel production scheduling rule engine platform, which can combine the actual working condition of each device in each process of steel production with the current process production scheduling planning rules, periodic indicators, raw materials, in-process products, product and device constraint rules, generate scheduling plans through the rule engine, and generate real-time scheduling schemes combined with real-time working condition signals, and hand over to the scheduling system for execution; in order to ensure the accuracy of the scheduling scheme, periodic calculation indicators are introduced into the platform, which can calculate the entire production rhythm and device state according to the periodicity of each process and schedule the production time; the rule engine platform can perform event processing logic for collaborative production scheduling of multiple production sequences according to the process connection relationship, each production sequence can share event processing signals, automatically calculate the priority of each production sequence, and perform multi-process rule calculation linkage operation when the working condition event changes, thereby ensuring the orderly execution of the production process. In addition, the rule engine platform saves all production signals and actual scheduling information, which can help users compare the platform calculation working condition with the actual working condition, can call evaluation indicators to evaluate the accuracy of the scheduling result, and assist system developers in optimizing the scheduling program. BRIEF DESCRIPTION OF DRAWINGS

[0057] Fig. 1 It is a rule engine platform module structure diagram.

[0058] Fig. 2 It is a working condition signal rule definition structure tree.

[0059] Fig. 3 It is a rule engine platform module deployment diagram.

[0060] Fig. 4 It is a real-time dynamic scheduling rule execution flowchart. DETAILED DESCRIPTION

[0061] Figs. 2-3 It is a specific embodiment of the present application.

[0062] As Fig. 1 It is a module structure diagram of the rule engine deployment platform.

[0063] As Fig. 1 The rule engine platform includes the following modules: rule definition module, database access module, real-time read-write module, database read-write module, rule execution module, post-confirmation module and rule evaluation module. The rule definition is the core of the entire rule engine.

[0064] As Fig. 2, the tree structure defined for the working condition rule of the rule engine, each rule node taking AND / OR / NOT predicate as root, can generate a complex working condition signal rule expression.

[0065] As Fig. 3 , the rule engine is mainly divided into two application programs for deployment, which are connected together through message event middleware. One application program is WebServer, in which rule definition, post-confirmation, rule evaluation and other modules are deployed, and there are two components, real-time data read-write module and database read-write module; the other application program is rule execution module, in which rule execution module, real-time read-write module and database read-write module are built-in.

[0066] As Fig. 4 is the scene of real-time production scheduling rule execution. In the starting stage, the system will carry out rule preparation, including working condition information, process path, batching, inventory, surplus material, group furnace and group pouring, equipment capacity, shutdown for maintenance and other rule configuration information, which are sent to the rule engine for initialization operation, and then the system performs the following operations: (1) the working condition information judgment logic is started, the system will start an independent thread to read the working condition rule, and judge the working condition of each production process in real time; (2) the dynamic scheduling logic is started, first read the production line configuration information, then read the order, plan and material information, calculate the output and batching amount according to the product output ratio, get the scheduling quantity of each process, then read the related configuration of each process according to the information of each process, combine with material matching calculation, the rule engine carries out multi-objective optimization calculation, gets the output of dynamic scheduling, and then combines with the calculation of equipment working condition signal to get real-time scheduling rule.

Claims

1. A steel production scheduling rules engine platform, characterized in that, The system comprises a rule definition module, a real-time read-write module, a database read-write module, a rule execution module, a post-confirmation module and a rule evaluation module. The rule definition module, the post-confirmation module and the rule evaluation module are integrated in a Web server supporting a RestFul access interface in the form of Web components, and provide an interactive interface for a user to define rules, confirm results and evaluate results. The rule execution module is separately deployed as an independent executable program, and is responsible for rule execution functions of various processes. The rule execution module reads rule configurations through the database read-write module, and accesses working condition data through the real-time read-write module. The real-time read-write module and the database read-write module exist as public components, and are built in the Web server and the rule execution program to provide real-time database access interfaces and database access interfaces for the rule execution. The rule definition module provides an interactive interface for a user to define working condition information of multiple processes, complex event rule definitions and periodic index rule definitions. The rule definition information is stored in a database, and is provided to the rule engine module through a database access interface for loading in later use. The real-time read-write module is a basic component, and is provided to the rule execution module to read production signals and working condition information of current processes and associated process devices from a real-time database, and to send the signals and information to a rule engine for calculation. The working condition scheduling results obtained by the calculation are sent to the real-time database. The database read-write module is a basic component, and provides configuration information of various rules. The configuration information includes device working condition information configuration, and read-write interfaces of a production database. The device working condition information configuration includes point names, states, calculation formulas and trigger event information of device working conditions. The production database includes model information of a certain steel production line, and production data information of operators, orders, materials, plans and inventories. The rule execution module reads the latest device working condition information of a process through the real-time read-write module, and accesses material information, plan information, process definition information, operator information, product specification information and quality information in production scheduling through the database module. The rule execution module calls a complex rule execution engine to process working condition events in real time, generates scheduling records and sends the scheduling records to a production scheduling system. The rule execution module also archives corresponding scheduling event records in a database. The post-confirmation module pops up a confirmation window on a Web window to display working condition information, plan information, material information and generated scheduling information according to a scheduling event record and a post-confirmation request, so that a user can confirm whether the scheduling results are correct and corresponding error information. The post-confirmation module is a supplement to real-time scheduling results. The rule evaluation module reads and writes event records through the real-time data read-write module, and obtains complex event rules and historical records processed according to the complex event rules through a database read-write interface. The rule evaluation module evaluates execution time, accuracy and timeliness parameters of the rules. The function of the rule definition module is that a user provides an interactive rule definition interface, is responsible for the definition of rules, a rule definition interface provides a RestFul service in the background, a rule definition client and a user terminal remotely edit and save rule definitions through the RestFul service interface, and perform access to the rule definitions; the configured rules meet the specifications of complex event processing, are stored in a special rule configuration database, and are loaded from the database before use; The rule definition module is deployed on a separate Web server and interacts with remote Web users through an Http Restful protocol.

2. The steel production dispatching rules engine platform of claim 1, wherein, The rule information configured in the rule definition module includes: A working condition signal calculation rule, which is generally configured as a conditional expression tree, each node is a binary tree node, the root node of the binary tree is a conditional connection predicate AND, OR or NOT, the NOT is a negation operation, the right node of the binary tree is empty, and only the left node is reserved; then a complex expression tree is constructed; A steel grade production process path rule, each steel grade has a unique process path, the process path is divided into different sections, and then each section is composed of different equipment; for each steel grade product, there are related process requirements on each equipment, including production time, temperature, composition, size, supporting equipment constraint conditions, and the production process path of the whole life cycle is described by a graph node; A production scheduling and batching rule, which generally provides corresponding steel grades, and then synthesizes finished product specification steel grade information, order yield information, remaining material quantity, production time, composition information, size information and processing equipment information; A device capacity constraint rule, including the maximum capacity and load of a single device, energy consumption constraint, device power constraint, and determination of the device throughput rate per unit time; A device maintenance and rotation rule; through the device maintenance rule, it can be determined which devices are available and which devices are out of service during production scheduling, and appropriate devices are selected to meet the production demand; A device furnace grouping and casting grouping rule, which aims to improve the utilization rate of equipment; in the case of batch production, the raw materials of adjacent steel grades are batched together for production, one furnace of molten iron corresponds to multiple steel grades in the steelmaking process, and one casting corresponds to multiple users in the casting process, but the steel grades need to be similar, and the complex event processing is extended; for complex event processing, the process data set of steel production is read, and then constraint conditions between the data set and the working condition signal are added to the process data set; Periodic rules: including standard production time and cycle of each device: furnace change time of steelmaking, roll change time of continuous rolling process, and cycle of heating and annealing process; through the device periodic rule, the periodic cycle processing semantics is added to the complex event processing, the corresponding device event information is generated periodically according to the cycle time, and is added to the semantic chain of the complex event processing.

3. The steel production dispatching rules engine platform of claim 1, wherein, The real-time read-write module is used for communication with a real-time database, provides the ability of the rule engine to read device working condition real-time / historical curve data, and the real-time module is deployed together with the rule execution module in the form of a component. The equipment working condition information includes the running status and related equipment parameters of each equipment in the steel production line, and the electrical signals of the equipment, such as current, voltage, power and electrical measurement information. The historical curve access supports data reading of the equipment working condition information and related equipment parameter information from real time to various periods of data, and provides real time data support for the calculation of the rule engine.

4. The steel production dispatching rules engine platform of claim 1, wherein, The rule execution module is responsible for executing rules, is based on complex events, and is implemented by using a Java programming language; in an initialization stage, all rule definition information is loaded through a database, and then a complex event processing rule is executed to process a working condition event; For complex event processing of the working condition type, a plurality of combined signals are included, such as a flow signal, a pulse signal, a start / stop signal, a fault signal and an alarm signal; for different signals, different preprocessing methods are adopted to first change the data to generate time-varying event information, and then a complex event engine is called to comprehensively judge the working condition information; the flow signal needs to be quantized according to an interval first, and then a data sampling interval, a sampling method and a threshold value of the flow signal are specified; sometimes, a rule for calculating historical data also needs to be specified, and an event value needs to be calculated by reading signal point statistical values of a certain time interval from a real time database; the pulse signal is a short trigger signal, which is represented as a sharp peak on a curve graph, and is judged according to a peak value signal; the start / stop signal is generally a simple state; the fault signal and the alarm signal maintain a true state when they occur, and the signals are reset after the fault and the alarm are eliminated; a complex event processing engine is built in the module, and the calculation of the working condition signal is mapped to a complex event mode to implement the calculation and processing process of the working condition signal; In the system, various signal-triggerable scheduling rule information is built in advance, the rule execution module reads the latest equipment working condition information of a working procedure from a real time database through a real time reading / writing module, and reads operation worker, multi-batch material, plan, order and inventory information through a database reading / writing module, to prepare for generating real time scheduling instructions; Before the execution of the real-time production scheduling instructions, the rule engine needs to combine the production process path rules of the product steel grade, the production scheduling and ingredient rules, the quantity of inventory and surplus material, combine the equipment group furnace group pouring rules, generate the production scheduling plan, and the algorithm adopted is a multi-objective optimization algorithm to achieve it; In the case of multiple processes, each process and the downstream process are in a supply-client relationship, and when solving, according to the demand of the sales order, the quantity of all processes is obtained by backstepping from the finished product process to the front process, and then the output of the final product is obtained by converting the quantity of all processes into the output of the final product, and then the theoretical total production quantity is obtained by subtracting the inventory of all products on the current production line, and then the total output of each sequence in the new production scheduling cycle is obtained according to the quantity conversion relationship of the products; Then, according to the order demand, the furnace group pouring operation is performed to generate a production scheduling plan for each production sequence according to time and priority; Then, according to the surplus material acting rule, the appropriate material on the production line is selected to dynamically allocate the production instruction; Then, the system generates actual production instructions for each production stage according to the operation rules of the equipment and the production ingredient information, and sends them to the control module of the equipment; For the surplus material acting rule problem, in production practice, materials of similar materials are used to act, and in the system, product process attributes and product characteristics are used to generate material demand keywords. The keywords can generate product process matching indexes by integrating steel grade codes, product specifications, product purposes, and quality composition parameters. In dynamic production scheduling, the rule engine calculates the matching degree between product material demand and inventory material through product process matching indexes to perform dynamic matching and provide a material feeding plan for the operators on site; The rule engine platform; according to the process connection relationship, the production scheduling event processing logic of multiple production sequences is performed according to priority, and each production sequence shares event processing signals. When production events arrive, multiple processes are linked to operate to ensure that production scheduling rules are executed according to the order of processes; Each process uses a steel grade, a process ID, an event type, a trigger process, and an execution time tuple to identify the execution condition. The trigger process is an array. Only when the processes in the trigger process have completed execution will the execution of the corresponding process rule be triggered.

5. The steel production dispatching rules engine platform of claim 1, wherein, The post-confirmation module pops up a confirmation window according to the event record, displays working condition information, plan information, time information, material information, and generated scheduling information on the window, and lets the user confirm whether the scheduling result meets the needs; For the case where the production record and the actual result are inconsistent, according to the error and delay of the signal, the current module adjusts the time rhythm of all production sequences according to the definition of the event and the value of the delay, increases the start time of all production processes in the current stage by an offset, and then corrects the corresponding production data; If there is an error in the measurement of the material, a relatively large loss occurs, and the process parameters and yield need to be confirmed; The post-confirmation module and the rule definition module are deployed in the Web server together to provide a remote Http access interface for user access; Production post-confirmation is completed by production workers to confirm the actual working conditions of the production process. Through the comparison of actual working condition information and signals determined by complex events, the location of system problems is determined. In addition, through the supplement of information, the supplement of production data and working condition information is improved to provide conditions for the traceability of post-production scheduling process.

Images