WO2013038262A1

WO2013038262A1 - A method and a system for scheduling of resources in a process industry

Info

Publication number: WO2013038262A1
Application number: PCT/IB2012/001792
Authority: WO
Inventors: Nareshkumar NANDOLA; Tarun Mathur
Original assignee: Abb Research Ltd
Priority date: 2011-09-15
Filing date: 2012-09-14
Publication date: 2013-03-21

Abstract

The invention relates to a method for scheduling of resources in a process industry. The method of the invention comprises the steps of simplifying at least one setup in a process industry, and modeling the simplified setup so as to obtain a model of the setup. Simplifying the setup include partitioning and / or rearranging the network of the setup. Discretizing the model of the setup to obtain at least one discrete time model; estimating the demand of the resources to obtain demand forecast; and estimating the supply of the resources to obtain supply forecast. Determining scheduling of resources based on the said model, said demand forecast and said supply forecast. Compensate for the difference in the supply forecast or demand forecast or both of the resources corresponding to discrete time model with smaller sampling interval in relation to scheduling of resources. Update the scheduling of resources to obtain optimal scheduling of resources. The method of the invention may be performed through simplifying the setup and determining the scheduling or resources. Also, the method of the invention can be co-extensively applied from discretizing the model of the setup through updating the scheduling of resources, in which the model employed herein may be a model that not been simplified depending on the need and complexity of the setup. The invention also relate to a system for scheduling of resources in a process industry in accordance with the method of the invention.

Description

A METHOD AND A SYSTEM FOR SCHEDULING OF RESOURCES IN

A PROCESS INDUSTRY

FIELD OF THE INVENTION

The invention relates to scheduling of resources in a process industry, and more particularly to a method and a system for scheduling of resources for a setup with uncertainty in process industry.

BACKGROUND

Scheduling of resources is inherently involved in any process industry. The problems associated with scheduling of resources, i.e., the scheduling problems do exist across various process industries. Examples for scheduling problem include but not limited to distribution of various petroleum products to various customers in a refinery, scheduling of various batch process, distribution of various high calorific byproduct gases to various in-house as well as external power plant in iron and steel industry, distribution of raw material to various processes, inventory management in warehouses, distribution of various utilities such as steam, water etc in process industries, etc. Each of the scheduling problems, at least as inferred from those referred above, differ from one another in respect of type of resource that need to be supplied, type of receivers and / or consumers, or the like.

In practice, supply and demand associated with scheduling problems and of scheduling of resources thereof, is highly uncertain. For example, generation of byproduct gases such as Blast Furnace Gas (BFG), Coke Oven Gas (COG), Corex Gas (CXG), etc in an iron and steel plant is highly uncertain with respect to time and it adversely affects the scheduling of resources if it is not accommodated systematically. Currently, scheduling of resources is performed based on heuristic rules derived from the experience of the operator, which is not capable of handling uncertainties in the supply and demand in an optimal manner. This leads to losses including economic loss due to inefficient scheduling of resources.

In the present scenario, the solution for scheduling of resources do not consider various operational aspects of the processes involved in a process industry and therefore do not address or account for the practical limitations or flexibilities, in its solution for scheduling of resources. Generally, scheduling of resources is done for longer period. However, in order to achieve a solution that is more practical and tractable, larger sampling time needs to be accounted. Also, in order to accommodate time varying uncertainties occurring at higher frequency, shorter sampling time needs to be considered. The invention is aimed at providing a solution for scheduling of resources that accounts for time varying uncertainties, considering the fact that such uncertainties are crucial in optimal distribution of various resources among various consumers.

OBJECTS OF THE INVENTION

It is an object of the invention to provide a method for scheduling of resources that considers operational aspects of the processes.

It is another object of the invention to provide a method for scheduling of resources that considers time varying uncertainties in the supply and demand.

It is also an object of the invention to provide a method for scheduling of resources that aid in re-planning of scheduling of resources.

Yet another object of the invention is to provide a system for performing the method for scheduling of resources in accordance with the method of the invention.

SUMMARY OF THE INVENTION

Accordingly, the invention provides a method for scheduling of resources in a process industry. The method of the invention comprises simplifying at least one setup in a process industry. Wherein, simplifying the setup includes partitioning the said setup and / or rearranging the setup or of the partitioned setup, to obtain a simplified setup. Modeling the simplified setup to obtain a model of the setup and determining scheduling of resources based on the model of the setup.

Accordingly, the invention also provide a method for scheduling of resources in a process industry. The method comprises discretizing the model of at least one setup in a process industry to obtain at least one discrete time model; estimating the demand of the resources to obtain demand forecast; estimating the supply of the resources to obtain supply forecast. Determining scheduling of resources based on the model, demand forecast and supply forecast. Compensating for the difference in the demand forecast or supply forecast or both corresponding to discrete time model with smaller sampling interval in relation to the scheduling of resources and updating the scheduling of resources to obtain optimal scheduling of resources.

Accordingly, the invention also provides a method for scheduling of resources in a process industry. The method of the invention comprises simplifying at least one setup in a process industry. Wherein, simplifying the setup includes partitioning the setup and / or rearranging the setup or of the partitioned setup, to obtain a simplified setup. Modeling the simplified setup to obtain a model of the setup. Discretizing the simplified model of the setup to obtain at least one discrete time model. Estimating the demand of the resources to obtain demand forecast; estimating the supply of the resources to obtain supply forecast. Determining scheduling of resources based on the model, demand forecast and supply forecast. Compensating for the difference in the demand forecast or supply forecast or both corresponding to discrete time model with smaller sampling interval in relation to the scheduling of resources and updating the scheduling of resources to obtain optimal scheduling of resources.

Accordingly, the invention also provides a system for scheduling of resources in a process industry. The system comprises a model simplifier unit for simplifying at least one setup in a process industry, and modeling the said simplified setup to obtain a model of the said setup. A sampling unit is provided for discretizing the model of the setup to obtain at least one discrete time model. A demand estimator unit for estimating the demand of the resources to obtain demand forecast; a resource estimator unit for estimating the supply forecast of the resources corresponding to the process of the said process industry and of the setup. An optimizer unit is provided for determining scheduling of resources based on the model, demand forecast and supply forecast. A compensator unit to compensate for the difference in the supply forecast or demand forecast or both of the resources corresponding to discrete time model with smaller sampling interval in relation to the scheduling of resources. An optimal scheduler unit for correcting and / or updating the scheduling of resources to obtain optimal scheduling of resources. A monitor unit for monitoring the scheduling of resources in order to enable reschedule of resources; and rescheduling of resources thereof

BRIEF DESCRIPTION OF THE DRAWING

With reference to the accompanying drawing in which:

Fig. 1 shows a system for scheduling of resources in accordance with the invention. DETAILED DESCRIPTION

The invention is explained referring to Fig. 1 and with reference to a non exhaustive exemplary embodiment. Industrial setup mentioned as setup in the specification is a system or the like in a process industry in which or by which processes is carried out. Such setup has network with varying complexities as applied to supply and demand of various resources involved in the processes in a process plant or to the extent applicable to a particular setup. It is required to simplify such networks in order to improve the scheduling of resources and of the efficiency thereof.

Practically, there persists limitation and certain other constraints that are either related or attached to the process or of the setup or both. This needs to be duly considered when simplifying the setup. Accordingly, in a system (100) for scheduling of resources as shown in Fig. 1, a model simplifier unit (101) is provided to simplify the setup. Such simplification of the setup is performed to partition the setup considering one or more related aspects such as nature and type of resources, physical parameters or of its equivalence thereof, operational constraints or limitations etc. The partitions of the system may be interlinked based on appropriate logical and / or operating conditions. Further, the network with one or more partitions of the setup is rearranged accordingly. Hence, the simplification of the setup besides having the setup simplified does not compromise on system or process dynamics with respect to one or more of supply or demand or inventory level of resources. A model needs to be developed for such setup that been simplified or for setup that not been simplified, in order to understand the workflow and characteristics of the setup and thereby to provide a suitable solution that eventually improve the functioning of the setup and of the processes and efficiency thereof. Accordingly, the setup is deduced to a model by the model simplifier unit (101) considering logical conditions or the like that would better define the setup. For instance, a continuous time model of supply and demand network can be deduced by pooling similar type of resources generated at different location in a setup or by different equipments, and considering physical limitations like pressure, distance with respect to the source of resources etc.

Further, the model is discretized by the sampling unit (102) so as to obtain two discrete time models, each for a sampling interval corresponding to high sampling frequency and other to a low sampling frequency. High sampling frequency may purport to sampling interval that is in the order of shorter time period e.g. seconds or minutes. Similarly, low sampling frequency may purport to sampling interval that is in the order of larger time period e.g. hour.

In current practice, most of the scheduling problem results in a Mixed Integer Program (MIP) that may be a Mixed Integer Linear Program (MILP) or Mixed Integer Non Linear Program (MINLP) or a Mixed Integer Quadratic Program (MIQP), and the complexity of the scheduling of resources depends on the scheduling prospect or scope and sampling frequency relating to sampling interval. For a scheduling scope that is fixed for a day, the size of MIP problem increases with increase in sampling frequency i.e with smaller sampling time, and lead to practically intractable solution. On the other hand, decreasing the sampling frequency by increasing the sampling time, the problem size gets reduced at the cost of losing effect of some important fast dynamics in supply and demand profile of the resources, which is crucial in deciding optimal distribution of various resources and reducing the wastage of resources when meeting the demand. Hence, it becomes imperative to capture the effect of such fast dynamics of the supply and demand profiles of the resources into the scheduling of resources besides maintaining the size of the MIP as smaller as possible for the purpose of practical implementation. Also, uncertainty in the demand needs to be considered. The invention through its proposal of having simplification of the setup made considering the uncertainties involved and having two discrete time model as stated herein above caters to the need as described before and of the solution thereof.

A demand estimator unit (103) is provided for estimating the demand of various resources in the processes so as to make a forecast of demand for the resources. The forecast of demand may be provided by a user or expert, or it can be estimated using a time series model and / or other equivalent model between production parameters of different consumers and their respective demands of the resources.

A resource estimator unit (104) is provided for estimating the forecast of supply of resources by employing time series or other equivalent modeling techniques to derive relation between various production parameters. Further, such profile or the forecast of resources is normalized through piecewise affine approximation of the profiles of various resources over a longer sampling interval and that over smaller sampling interval.

The optimizer unit (105) solves the MIP optimization problem for scheduling of resources based on the discrete time model with larger sampling interval provided by model simplifier unit (101), forecast of demand provided by demand estimator unit (103) and normalized forecast of supply of resources with respect to longer sampling interval provided by resource estimator unit (104), so as to determine scheduling of resources. Solving an optimization problem involves one or more of but not limited to: deciding on the manipulated variables, formulating cost (objective) function which needs to be minimized or maximized, formulating various logical and / or process constraints or limitations, supply and demand forecast over a scheduling horizon or scope that may correspond to larger sampling interval, scheduling solution purporting to scheduling of resources from MIP.

The scheduling of resources as determined by the optimizer unit ( 105) is provided to a compensator unit (106). The compensator unit (106) is also provided with discrete time model of high and low frequency sampling from the model simplifier unit (101), and with normalized forecast of supply of resources of longer and smaller sampling interval from the resource estimator unit (104). The compensator unit (106) provides compensation for the finer aspects of the scheduling of resources and for the fast dynamics of the supply forecast or demand forecast or both of the resources as compared to the counterparts purporting to coarser aspects and slow dynamics respectively. This is done through the difference in the supply and / or demand forecast of resources, and is been provided by the compensator unit (106). Compensator unit (106) can also solve a simplified optimization problem to provide compensation. The simplified optimization problem may be a linear program obtained by fixing the integer variable of MIP formulated for coarser aspects.

An optimal scheduler unit (107) provided herein, corrects and / or updates the scheduling of resources obtained from the optimizer unit (105) thereby capturing the effects of both fast and slow dynamics of the supply and / or demand forecast of resources. Such update can be made using simple heuristic rules or by solving a less complex optimization problem or both. Also, it becomes possible to solve optimization problem considering discrete time model of smaller interval for a specific sample of longer interval and update the scheduling of resources obtained from solving MIP optimization problem by the optimizer unit (105).

An optimal scheduling of resources is hereby obtained in the manner described here before in the description considering the practical limitations and constraints along with uncertainties in the supply and / or demand forecast of resources that been vested with the supply and demand of resources. A monitor unit (108) is provided for monitoring various resources involved in the processes and of the corresponding demand profiles, to enable re-planning or re-scheduling of resources or both. In practice, availability of resources and its demand is always varying due to some unknown or unpredictable disturbances. Hence, it is important to avoid re-planning that is unwarranted. Accordingly, it becomes essential to distinguish between the variations arising out of instantaneous disturbance and / or noise that marginally affects the scheduling of resources and that of variations arising out of sustainable disturbance of longer time that adversely affects the scheduling of resources. The monitor unit (108) re-plans or reschedules the scheduling of resources only when there is a prolonged difference between the actual supply and demand forecast of resources and of the initially forecasted supply and demand forecast of resources. Moreover, the monitor unit (108) monitors the estimated forecast or user provided forecast of supply and demand of resources with respect to their actual values. The monitor unit (108) distinguishes between smaller disturbances and / or noises and prolonged disturbances and / or noises in forecast of supply and demand of resources, using trend analysis or other equivalent statistical methods and raises an alarm for re-planning or reschedule of scheduling of resources, or enable or cause the same, in case of prolonged disturbance or the like.

The invention can be applied effectively in scheduling problem that include but not limited to distribution of various petroleum products to various customers in a refinery, scheduling of various batch process, distribution of various high calorific byproduct gases to various in- house as well as external power plant in iron and steel industry, distribution of raw material to various processes, inventory management iri warehouses, distribution of various utilities such as steam, water etc in process industries, etc.

Considering, iron and steel industry, in the context of the invention as a non exhaustive and non restrictive exemplary embodiment or case, certain and other aspects of the invention is referred for better understanding. Partitioning of the network during simplification of setup can be done for byproduct gas distribution having different sources of gas generator producing same gas at same pressure and stored in multiple gas holders via a common header, by considering single pool of this gas with one virtual gas holder. One dynamic equation can be obtained instead of obtaining separate dynamic equation for each gas holder. Considering, multiple gases are produced at different pressure, then partition of the network can be done between similar pressures while keeping one virtual gas holder for each partition. This way the setup can be simplified and modeled.

Going through the same consideration with respect to iron and steel industry, the invention can be coextensively applied thereto as described before. With the pretext of heuristic rules for the scheduling of resources, heuristic approach may involve certain criteria or the like such as a) Open loop simulation of gas holder dynamics over a scheduling horizon considering shorter sampling interval, supply of gases to various consumers from the results of MIP formulated for larger sampling time and actual gas generation profile at shorter sampling interval; b) Check for violation of various criteria based on upper and lower bounds on holder levels, rate of raise or rate of fall in gas holder level; c) Calculate adjustment quantity from difference in gas holdup with respect to above mentioned criteria; d) Redistribute the adjustment quantity among the possible consumers if any e.g. if some consumer is able to accommodate more or less amount of gas then such consumer can be considered for redistribution; e) If MIP solution indicates flaring of some gas then flaring quantity of gas should also be considered as an adjustment quantity during redistribution; f) Check for constraint violation for corrected solution. Alternatively, an optimization based approach can also be adopted independently or along with the heuristic approach, through a linear program obtained by fixing the integer variable of MIP formulated for coarser aspects. The linear program referred herein provide compensation for the quantity of particular gas to be supplied to each customer.

The invention is not restricted by the preferred embodiment described herein in the description. It is to be noted that the invention is explained by way of exemplary embodiment and is neither exhaustive nor limiting. Certain aspects of the invention that not been elaborated herein in the description are well understood by one skilled in the art. Also, the terms relating to singular form used herein in the description also include its plurality and vice versa, wherever applicable. Any relevant modification or variation, which is not described specifically in the specification are in fact to be construed of being well within the scope of the invention.

Claims

WE CLAIM

1. A method for scheduling of resources in a process industry, the method comprising the steps of:

simplifying at least one setup in a process industry, wherein partitioning the said setup, and / or rearranging the said setup or of the said partitioned setup to obtain a simplified setup; modeling the said simplified setup to obtain a model of the said setup; and

determining scheduling of resources based on the said model.

2. A method for scheduling of resources in a process industry, wherein the method comprising the steps of:

discretizing the model of at least one setup in a process industry to obtain at least one discrete time model;

estimating the demand of the resources to obtain demand forecast;

estimating the supply of the resources to obtain supply forecast;

determining scheduling of resources based on the said model, said demand forecast and said supply forecsat;

compensating for the difference in the said demand forecast or supply forecast or both corresponding to discrete time model with smaller sampling interval in relation to the said scheduling of resources; and

updating the said scheduling of resources to obtain optimal scheduling of resources.

3. A method for scheduling of resources in a process industry, wherein the method comprising the steps of:

simplifying at least one setup in a process industry, wherein partitioning the said setup, and / or rearranging the said setup or of the said portioned setup to obtain a simplified setup;

modeling the said simplified setup to obtain a model of the said setup;

discretizing the model of the said setup to obtain at least one discrete time model;

estimating the demand of the resources to obtain demand forecast;

estimating the supply of the resources to obtain supply forecast;

determining scheduling of resources based on the said model, said demand forecast and said supply forecast; compensating for the difference in the said demand forecast or supply forecast or both corresponding to discrete time model with smaller sampling interval in relation to the said scheduling of resources; and

4. The method as claimed in claim 1 or 3, wherein partitioning the said setup includes combining and / or splitting the network of the said setup with relevance to one or more aspects corresponding to nature and type of resources, physical parameters or of its equivalence thereof, operational constraints or the like.

5. The method as claimed in claim 1 or 3, wherein the said simplified setup accounts for process dynamics in relation to one or more of supply or demand or inventory level of the said resources.

6. The method as claimed in any one of the claims 1 to 3, wherein the said model of the said setup is a continuous time model of the said setup in relation to supply and / or demand of the corresponding said resources.

7. The method as claimed in claim 2 or 3, wherein the said discrete time model purports to a model with sampling interval having high sampling frequency and / or low sampling frequency.

8. The method as claimed in claim 7, wherein the said discrete time model corresponding to high sampling frequency purports to capturing fast and / or finer dynamics of the supply and demand profile of the said resources.

9. The method as claimed in claim 7, wherein the said discrete time model corresponding to low sampling frequency purports to capturing slow and / or coarser dynamics of the supply and demand profile of the said resources.

10. The method as claimed in claim 2 or 3, wherein estimating the demand of the resources includes obtaining demand forecast from the user or by using time series model or the like.

1 1. The method as claimed in claim 2 or 3, wherein estimating the supply of the resources includes making a forecast of the supply of the said resources by using a time series model or the like to determine the relation between various production parameters in respect of the said resources.

12. The method as claimed in claim 1 1, further comprising generating a piecewise affine approximation of the profile of the said resources over a long sampling interval.

13. The method as claimed in claim 1 1, further comprising generating a piecewise affine approximation of the profile of the said resources over a short sampling interval.

14. The method as claimed in any one of the preceding claims, wherein compensating for the difference in the demand forecast or supply forecast or both of the said resources include capturing the effect of the fast dynamics of the demand forecast or supply forecast or both of the said resources.

15. The method as claimed in any one of the claims 1 to 3, wherein updating the said scheduling of resources includes capturing the effects of fast dynamics and slow dynamics of the demand forecast or supply forecast or both of the said resources.

16. The method as claimed in any one of the claims 1, 2, 3 or 15, wherein updating the said scheduling of resources includes using heuristic rules or solving an optimization problem or both.

17. The method as claimed in any one of the preceding claims, further comprising monitoring the scheduling of resources in order to enable reschedule of resources; and rescheduling of resources thereof.

18. A system for scheduling of resources in a process industry, wherein the system comprises:

a model simplifier unit for simplifying at least one setup in a process industry, and modeling the said simplified setup to obtain a model of the said setup;

a sampling unit for discretizing the said model of the said setup to obtain at least one discrete time model; a demand estimator unit for estimating the demand of the resources to obtain demand forecast;

a resource estimator unit for estimating the supply of the said resources to obtain supply forecast;

an optimizer unit for determining scheduling of resources based on the said model, said demand forecast and said supply forecast;

a compensator unit to compensate for the difference in the demand forecast or supply forecast or both corresponding to discrete time model with smaller sampling interval in relation to the said scheduling of resources;

an optimal scheduler unit for correcting and / or updating the said scheduling of resources to obtain optimal scheduling of resources; and

a monitor unit for monitoring the scheduling of resources in order to enable reschedule of resources; and rescheduling of resources thereof.