Monitoring of Manufacturing Processes Based on Econometric

Computational Control

DANIELA GHELASE, LUIZA DASCHIEVICI

Faculty of Engineering, Braila

Dunarea de Jos University

47, Domneasca St., Galati

ROMANIA

Abstract: - In order to address the present-day changes and tendencies, the industrial enterprise has advanced in its

manner of approaching the manufacturing processes by performing a high integration of operations, functions and

compartments. One can say that we are dealing with a single structure, with no organising levels, of the enterprise

which reacts to inputs and manufactures the requested product by almost immediately adapting to the customers’

requirements or to the requirements of other interested parties.

Key-Words: - Econometric modeling, manufacturing system, process, control, market, management.

Received: June 9, 2022. Revised: August 9, 2023. Accepted: September 23, 2023. Published: November 6, 2023.

1 Introduction

The most important feature of the present-day market is

the high level of customizing the products requested by

customers, which brings about a large variety of the

requested products and a small volume of the batches in

which these products are manufactured [1]. One of the

responses industry can give to this challenge is to

increase its responsiveness by continuously

reconfiguring the manufacturing systems in compliance

with the task to be carried out [2].

The manufacturing system is defined as being the

ensemble of technological systems which are used for

realization of certain product. Each of these

technological systems is made up of machine-tool, tools,

apparatus, parts, operator and it executes one of the

operations of technological process [3]. The

manufacturing system is made up when the manufacture

of the product is started up. After this, when is started up

another product, the problem of structure of the

manufacturing system is taken again from the begin.

This ad-hoc structure of manufacturing system is always

presented to the batches manufacturing, but it isn’t to the

serial manufacturing (the ensemble of technological

systems which make up the manufacture system is for

long time in the same structure) [4], [5].

The performance of the manufacturing system

depends on its control. The special literature [6], [7]

refers to the relationship between the parameters of the

cutting regimes and the technical performance of the

manufacturing system (pure technical aspects) and the

other [8], [9] refers to the relationship between the

product manufactured by the manufacturing system and

the market (relations of economic nature). In special

literature there aren’t approaches of the ensemble of

manufacture system and market. There are important

improving resources of the manufacturing system

performances because the technical and economic

aspects are approached separately.

Nowadays, the manufacturing systems are controlled

by means of numerically programmed machine tools

which are part of the system [10]. The control is

exclusively technical because there is no economic

variable, although this is actually the ultimate goal of

any processing process.

The dynamic changes and the overall progress of

society are reflected at company level by many orders

in number, small in volume, very diverse, obtained

through frequent auctions with short- term response ,

which leaves no time for a relevant analysis of said

orders. As a result, a long-term management is no

longer possible.

A sort of fluctuating (just like the market) on-line,

fastly responsive, prompt and rapid, however,

ephemeral management is called for [11].

In this context, an important problem is the

management of the customer - company relationship

with regard to the product contracting, on the one side,

and the management of the contracted product

manufacturing process on the other.

This paper proposes a solution to this problem, which

is based on building an economic model of the product

requested by the customer [12].

This economic model is to be used in two situations:

- the first situation is the one where the economic model

of the product is used in conducting the negotiations

between the enterprise and the customer who requests

the manufacture of the product in a certain number;

- the second situation occurs after entering upon the

contract and starting up the manufacture of the

Engineering World

DOI:10.37394/232025.2023.5.16

Daniela Ghelase, Luiza Daschievici

E-ISSN: 2692-5079

144

Volume 5, 2023

contracted product, when the optimal control of the

manufacturing machines is requested during the carrying

on of the manufacturing processes.

The present-day solutions to this problem are based

on the manufacturing systems modelling and on using

these models both for establishing the price quotation

and for the operational control of the manufacturing

processes. Some of the present-day solutions are given

below.

A first example is the one where a model, especially

created for small and medium enterprises, is the virtual

integrated manufacturing system which starts from the

computer-aided integrated manufacture. The architecture

of the manufacturing system, even though it is based on

three-level structured agents, has the advantage of

connecting the agents through the internet [13],

[19],[20].

Figure 1 suggests a network of small and medium

enterprises that are structured after the model of the

virtual integrated manufacturing system based on

hierarchically structured agents [14], [15], [16], [17].

Fig. 1 Model of virtual manufacturing system based on

hierarchically structured agents

A second example is the one where a multi-agent

structure using a fictitious market model is used; this

structure can be used in programming the CNC

equipment of the manufacturing machines. In this

structure the agents fulfil the sellers’ and buyers’

assignments on a virtual market.

The method uses the costs and proceeds calculated for

every activity of the agents, and, by the price strategy,

the coordination of the prices is accomplished. To this

purpose [18], the quoted price, the penalty cost and the

opportunity cost based on the delay times, the cost for

the previously determined time unit based on the task

(i) performance times and on the task (i) finishing times

at the work site (k) are calculated by such relations as:

BPi,k=OperationalCosti,k +

+PenaltyCosti,k+OportunityCosti,k (1)

PenaltyCosti,k=δi x Max(Ci,k – Di,0)+

+ εi x Max(Di – Ci,k,0) (2)

OprotunityCosti,k=Σj≠i δj x [Max(Cj,k(i,j) – Dj,0) –

-Max(t + pj,k – Dj,0)] / Lk (3)

BPi,k= pj,k + δi x Max(Ci,k – Di,0) + Σj≠i δj x [Max(Cj,k(i,j)

– Dj,0) – Max(t + pj,k – Dj,0)] / Lk (4)

A third example is the one where the concept of

“smart product” is elaborated, according to which every

product runs its own manufacture, thus allowing the

separation of the manufacture from the forwarded order,

in the same manner that the physical flow is separated

from the informational one.

The smart product requires services from the

manufacturing resources and, based on the cooperation

logic, aims at fulfilling its own requirements.

Nevertheless, another result pursued is, first of all,

dealing with the variety of factory’s resources.

Resources, as well as smart products, can be

presented as communication entities. In this case, the

organisation is considered to be the cooperation

between resources and communication entities, from

which calculation entities named rules result.

A meta-model of rules arranging and of performance

systems running by the product, as well as its

application to the designing is used for the control.

Figure 2 suggests a entities control model [19].

Fig. 2 Entities control model

The solution proposed in this paper consist in:

Engineering World

DOI:10.37394/232025.2023.5.16

Daniela Ghelase, Luiza Daschievici

E-ISSN: 2692-5079

145

Volume 5, 2023

- the econometric modelling of manufacturing

system;

- the econometric modelling of each element in the

structure of a manufacturing system and in using the

resulted models for building a global model of the

product requested by the customer under the

assumption that the product will be manufactured by

means of that manufacturing system;

- using the global model of the requested product in

drawing up successive offers during negotiations;

- using this model for drawing up optimal part

programmes related to the manufacturing operations to

be performed by the machines that make up the

manufacturing system.

This paper presents a method of econometric

modelling of the manufacturing machines which is

based on the machine operation monitoring, the

collected data storing and an analytical model building

meant to describe the machine economics. It also

presents the construction model of the global model of

the requested product as well as the manner of using

this model.

2 Econometric Modelling Method

Proposed

At the conceptual level, the proposed method consists in

monitoring and recording the relevant state variables of

the manufacturing machine in a data base.

Then the causality relationships between parameters

are identified. Based on these relationships, clusters of

independent variables are established. Further on, based

on the input data to be fed into the model for

interrogating it, a cluster of neighbouring states is made

up, at the centre of which is the state to which the

respective input data are related.

Finally, a linear model whose variables are the

variables of one of the clusters of identified variables is

fitted on the states cluster. The model interrogation will

consist in the calculation of the sought variables,

depending on the input data that describe the current

state of the machine. These input data are the ones

which have been previously considered in the procedure

of enclosing the states cluster.

It can be noted that, according to the method proposed,

the model construction and its operation are

accomplished within an integrated algorithm which is

run through upon each interrogation of the model.

At the operational level, the variable clustering is

based on using the “best model” facility which is offered

by the neural networks technique applied to a data set

recently obtained from monitoring the model. The state

cluster construction, the linear model is fitted to, first

implies the use of the 2nd rank Minkowski distance for

the classification of states, in the increasing order of

their distance to the state the data set to be used in

interrogating the model belongs to. That is why only the

variables representing these input data will be

considered in the calculation of Minskowski distance.

The states cluster is to be obtained either by

restricting the value of the distance or by restricting the

number, k, of retained states or using these two

conditions. The construction of the mathematical model

is made by linear regression. It can be noted that this is a

local model, as it is only valid in the vicinity of the

linking state by means of which the model is

interrogated, that this model is meant to be used just

once as, after the interrogation, it is given up.

3 Simulations and Discussions

During the experiment, first the data resulted over the

last 6 months have been collected, with regard to the

manufacturing machines that had been used for

manufacturing some important parts in the construction

of dump truck bins, namely the attachment plate of the

supplementary chassis of the dump truck bin. Data

regarding the actual work times, data referring to the

modes of operation, data regarding the amounts of

resulted wastes, data regarding all types of consumption,

as well as data regarding the orders for delivered

products were collected.

The case study under consideration is the one where

the customer has requested an offer referring to both the

manufacture of the part named “Attachment plate of the

supplementary chassis of the dump truck bin” shown in

figure 3, and to its welding onto the supplementary

chassis of dump truck bins.

Fig. 3 Attachment plate of the supplementary

chassis of the dump truck bin

180

350

100

 = 8

 16

Engineering World

DOI:10.37394/232025.2023.5.16

Daniela Ghelase, Luiza Daschievici

E-ISSN: 2692-5079

146

Volume 5, 2023

Table 1

Fig. 4 Table arranged according to the shortest distances

The number of ordered parts is 80 and the total time for

making them is 4.87 hours. The part will be welded on a

length of 2x175 mm to the supplementary chassis (three-

layer tee weld according to the drawing). The material to

be used for making the parts is OL 37.

Table 1Item nr.

Type of material

Type of weld

Length of welding

seam

Number of passes

Amperage (A)

Rate of welding

(mm/sec.)

Amount of welding

wire (m)

Number of pieces

Welding time (sec.)

Power

consumption

(kW/h)

Cost of operation

(RON)

Amount of wastes

(Kg)

v10

v11

v12

OL 52

Tee

501

200

10,2

4,2

1375

10,52

3.156,29

15,78

OL 52

Tee

562

198

9,2

5,25

3075

5,31

1.611,06

8,05

OL 42

Tee

498

185

8,2

3,25

3705

29,17

9.461,99

47,30

OL 42

Tee

589

211

10,25

6,2

3467

57,16

16.256,3

81,28

Engineering World

DOI:10.37394/232025.2023.5.16

Daniela Ghelase, Luiza Daschievici

E-ISSN: 2692-5079

147

Volume 5, 2023

For the purpose of performing these operations, we will

describe the technological flow along with the required

processing, tools, equipment and devices to be used and

we will present the resulted data hereinafter:

1) Cutting to dimension from the semi-finished

product:

a - cutting to the length of 2x350 mm

b - cutting to the width of 2x180 mm

Equipment: Plasma jet cutting equipment PLASMA

Jackle

Verifying tools and devices: Plasma nozzle 1.5 to 1.7;

Compressed air; Acetylene

2) Drilling holes having the diameter of 16.4 m

a - drilling one hole after another

b - positioning for every hole

Equipment: general purpose drilling and milling

machine FUS 32

Verifying tools and devices: 16 mm diameter drill

3) Weld on the length of 175 mm on both sides, on the

a - positioning and fixing

b - 4-pass tee weld

Equipment: welding equipment MAG Jackle

Verifying tools and devices: wire of 1.6 mm.

In order to succeed in demonstrating the viability of

the solution to the problem of continuous identification

and of adaptive and optimal running of the holonically

modelled manufacturing systems, a practical data base

resulted from process measurements was obviously

required.

For this, some determining, measuring and

monitoring of the cutting-off, drilling and welding

processes was made, whose results are summarized in

the following table, as an example (see Table 1).

For the experimental implementation of the modelling

method proposed, an IT product was developed and

designed in the Visual FoxPro programming

environment, using the function libraries in Matlab and

C++.

The simulation implies two data input sequences: a

sequence referring to entering the customer’s

requirements and the second sequence referring to

entering the work restrictions.

Between these two sequences, the distance

calculation sequence, according to the above-

mentioned, as well as its comparison to a minimum

distance previously established are inserted, so that the

values of the variables could concentrate into a vicinity

of the values of the customer’s requirements.

Thus, a table arranged according to the shortest

distances, having values under 160 and 17 recordings,

results. Thus, the state cluster around the input data the

linear model will be fitted to is built (Fig. 4).

The performance of modelling consists in the fact

that, using a dedicated data set, the method has been

tested by comparing the values of the state variables

obtained by interrogating the model to the measured

values of those variables.

The modelling errors were within a tolerance range

of -15% to + 15%.

Also, the use of the econometric modelling method

for the optimal running of the three manufacturing

machines making up the manufacturing system related

to the item called “plate” has been simulated.

The target function was the cost and the restriction

enforced was the total duration of manufacturing the

batch of 80 parts.

The optimization had the distribution of the total

duration of 4.87 hours to the three operations as its

variable and the following five cases were considered

(see Table 2).

0.00

5,000.00

10,000.00

15,000.00

1 2 3 4 5

Minimum cost

Fig. 5 Histogram of the cost of a part manufacture

depending on the time distribution model

Table 2

Item nr.

Time allotted for the

welding operation

(sec.)

Time allotted for the

cutting-off operation

(sec.)

Time allotted for the

drilling operation

(sec.)

Total time (sec.)

9.500

5.500

2.500

17.500

9.000

7.000

1.500

17.500

8.500

6.500

2.500

17.500

8.000

7.500

2.000

17.500

7.500

6.000

4.000

17.500

Engineering World

DOI:10.37394/232025.2023.5.16

Daniela Ghelase, Luiza Daschievici

E-ISSN: 2692-5079

148

Volume 5, 2023

Figure 5 presents the results have been obtained after

running the algorithm.

4 Conclusion

The method proposed has the advantage of being

applicable to any manufacturing machine, regardless the

physical nature of the process and the product features.

Furthermore, the result of applying the method is not

significantly distressed by the inherent modification of

the behaviour of the modelled machine in the long run.

The method provides the extended modelling of the

manufacturing machine. The level of extension is only

limited by the number of the monitored state variables.

The level of the modelling accuracy satisfies both the

exacting requirements specific to a contract negotiation

and the ones specific to the operational management.

The implementation of the method implies the

implementation of an adequate informational system

and the completion of the CNC systems with new

facilities by comparison to their present-day versions.

References:

[1] A. Tharumarajah, 2002. A Self-Organising View of

Manufacturing Enterprises, Computers in Industry,

2003 Elsevier Science B.V., 2003, 185-196.

[2] G. Morel, H. Panetto, M. Zaremba, F. Mayer,

Manufacturing Enterprise Control and Management

System Engineering: Paradigms and Open Issues,

Annual Reviews in Control, Elsevier Ltd., 2003, 199-

209.

[3] E. P. Paladini, A Pattern Recognition and Adaptive

Approach to Quality Control, WSEAS Transaction on

System and Control, Vol. 3, 2008, 627-643.

[4] A. Epureanu, F.B. Marin, V. Marinescu, M. Banu, I.

Constantin, Reconfigurable Machine Tool

Programing- A New Approach, WSEAS Transaction

on System and Control, Vol. 5, 2008, 463-472.

[5] V. Marinescu, I. Constantin, A. Epureanu, M. Banu,

F.B. Marin, Online Adaptive Learning System for

Reconfigurable Machine Tool, WSEAS Transaction

on System and Control, Vol. 3, 2008, 473-482.

[7] M. Ozbayrak, Activity – Based Cost Estimation in

Push/Pull Advanced Manufacturing System,

International Journal of Production Economics 87

(1), 2006, 49-65.

[6] F. H’nida, P. Martin, F. Vernadat Cost Estimation

in Mechanical Production:The Cost Entity Approach

Applied to Integrated Product Engineering,

International Journal of Production Economics,

[8] J. Wooldridge, Introductory Econometries: A

Modern Approach, Mason: Thomson South-Western,

2003.

[9] A. Rooney, Handbook on the Knowledge Economy,

Cheltenham: Edward Elgar, 2005.

[10] A. Epureanu, T. Virgil, On-Line Geometrical

Identification of Reconfigurable Machine Tool Using

Virtual Machining, Enformatica, vol. 15, Spain, 2006.

[11] L.Y. Steven, H. L. Rogelio, G. Robert Landers

(2004) – Machining Processes Monitoring and

Control: The State- of- the- Art, in Journal of

Manufacturing Science and Engineering, DOI

10.1115, 2004

[12] J. M. Simao, P. C. Stadzisz, G. Morel,.

Manufactuting Execution Systems for Customized

Production, Journals of Materials Processing

Tehnology, Elsevier B.V., 2006, 268-275.

[13] B. Huang, H. Gou, W. Liu, Yu Li, Min Xie, A

Framework for Virtual Enterprise Control with the

Holonic Manufacturing Paradigm, Computers in

Industry, Elsevier Science B.V., 2002, 299-310.

[14] D. Wang, Sev. V. Nagalingam, Grier C.I. Lin.

Development of an Agent-based Virtual CIM

Architecture for Small to Mediul Manufacturers,

Robotics and Computer-Integrated Manufacturing,

Elsevier Ltd., 2005

[15] P. Valckenaers, H. Van Brussel, Holonic

Manufacturing Execution Systems, Annals of the

CIRP 2005, Elsevier Ltd., 2005.

[16] J. Zhang, L. Gao, F. T.S. Chan, P. Li, A Holonic

Architecture of the Concurrent Integrated Process

Planning System, Journal of Materials Processing

Technology, Elsevier Science B.V., 2003, 267-272.

[17] J. Jarvis, R. Ronnquist, D. McFarlane, L. Jain, A

Team-Based Holonic Approach to Robotic Assembly

Cell Control, Journal of Network and Computer

Apllications, Elsevier Ltd., 2004, 160-176

[18] X. Shao, Li Ma, Z. Guan, A Market Approach to

Decentralized Control of a Manufacturing Cell, Chaos

Solitons & Fractals, Elsevier Ltd. 2007

[19] P. Leitao, F. Restivo, ADACOR: A Holonic

Arhitecture for Agile and Adaptive Manufacturing

Control, Computers in Industry, Elsevier B.V., 2005,

121-130.

[20] Q. Wang, K. L. Yung, W. Hung Ip, A Hierarhical

Multi-View Modeling for Networked Joint

Manufacturing System, Computers in Industry,

Elsevier B.V., 2003, 59-73.

Engineering World

DOI:10.37394/232025.2023.5.16

Daniela Ghelase, Luiza Daschievici

E-ISSN: 2692-5079

149

Volume 5, 2023

Contribution of Individual Authors to the

Creation of a Scientific Article (Ghostwriting

Policy)

The authors equally contributed in the present

research, at all stages from the formulation of the

problem to the final findings and solution.

Sources of Funding for Research Presented in a

Scientific Article or Scientific Article Itself

No funding was received for conducting this study.

Conflict of Interest

The authors have no conflicts of interest to declare

that are relevant to the content of this article.

Creative Commons Attribution License 4.0

(Attribution 4.0 International, CC BY 4.0)

This article is published under the terms of the

Creative Commons Attribution License 4.0

https://creativecommons.org/licenses/by/4.0/deed.en

_US

Engineering World

DOI:10.37394/232025.2023.5.16

Daniela Ghelase, Luiza Daschievici

E-ISSN: 2692-5079

150

Volume 5, 2023