Sphere packing with a geometric based -main

Sphere packing with a geometric based compression algorithm

K.Han T ,Y .T.Feng,D.R.J.Owen

Civil and Computational Engineering Centre,School of Engineering,University of Wales Swansea,Singleton Park,Swansea,SA28PP ,UK

Received 17February 2004;received in revised form 14February 2005;accepted 28April 2005

Available online 23June 2005

Abstract

An efficient algorithm for the random packing of spheres can significantly save the cost of the preparation of an initial configuration often required in discrete element simulations.It is not trivial to generate such random packing at a large scale,particularly when spheres of various sizes and geometric domains of different shapes are present.Motivated by the idea of compression complemented by an efficient physical process to increase packing density,shaking,a new approach,termed compression algorithm,is proposed in this work to randomly fill any arbitrary polyhedral or cylindrical domains with spheres of various sizes.The algorithm features both simplicity and high efficiency.Tests show that it takes 181s on a 1.4-GHz PC to complete the filling of a cylindrical domain with a total number of 26,787spheres,achieving a packing density of 52.89%.

D 2005Elsevier B.V .All rights reserved.

Keywords:Sphere packing;Compression;Shaking;Discrete elements;Contact search

1.Introduction

In discrete element simulations,the pre-processing often involves the preparation of an initial random discrete object configuration that represents realistic situations.It is however not trivial to generate such a random packing at a large scale,particularly when discrete objects of various sizes and geometric domains of different shapes are present in many practical applications.

Random packing of disks/spheres is a research topic that has attracted considerable attention in different areas with different objectives over the past decades,and a number of packing approaches have been developed [1–5].A fairly comprehensive review on existing packing algorithms of both 2D and 3D is conducted in [6].A common feature of these algorithms,including the packing of 2D objects,is often the substantial CPU cost involved especially for large-scale problems.Therefore,the development of an effective packing procedure,in terms of computational costs,for a

large number of disks/spheres becomes an important numerical and practical issue in the discrete element simulation of many industrial applications.

Such an effort is reported in our previous work [6],which introduces a novel algorithm based on the idea of advancing front techniques for finite element mesh generations for the random packing of 2D disks.This is a pure geometric packing algorithm without physical forces involved.Further develop-ment has considered a primary packing direction,which emulates,for instance,gravitational compression,during the advancing packing process.As only the fronts,each comprising two disks,need to be maintained,the approach has proven to be very efficient with a linear complexity.It takes only a few seconds to pack 1,000,000disks on a normal desktop PC,and can achieve 80%packing density in general.The same procedure has also been successfully employed in the packing of ellipses and polygons [7].

In principle,this approach can also be extended to sphere packing.However,both the algorithm and implementation are much more complex.At the algorithmic level,the fronts consist of triangular facets,each obtained by joining the centres of three spheres.As the facet cannot be fully covered by the associated spheres,some spheres may penetrate through the facets without necessarily overlapping the existing spheres.

0032-5910/$-see front matter D 2005Elsevier B.V .All rights reserved.doi:10.1016/j.powtec.2005.04.055

*Corresponding author.

E-mail addresses:k.han@https://www.wendangku.net/doc/a816211409.html, (K.Han),

y.feng@https://www.wendangku.net/doc/a816211409.html, (Y .T.Feng),d.r.j.owen@https://www.wendangku.net/doc/a816211409.html, (D.R.J.Owen).

Powder Technology 155(2005)33–

41

https://www.wendangku.net/doc/a816211409.html,/locate/powtec

These difficulties significantly increase the algorithmic com-plexity.At the implementation level,a far larger number of different circumstances need to be dealt with properly when different shaped geometric domains are considered.

It should be noted that in many applications,the actual packing configurations are produced under various physical forces,especially the action of gravity.Consequently,a compressive packing direction is introduced to represent the realistic situation as closely as possible.

Based upon the above observation,this work attempts to develop a new numerical approach,termed the compression algorithm,to generate a random packing of spheres of various sizes within a given geometric domain.As will be detailed below,the new algorithm is again geometric based and employs several existing numerical techniques,partic-ularly the recent development of a very highly efficient contact search algorithm[8,10].This algorithm is charac-terized by its simplicity,high efficiency and applicability not only to the random packing of disks and spheres,but to other shaped particles.The novelty of the algorithm lies not in the proposal of those individual techniques employed but in the unique way that they are combined together to achieve a simple and effective sphere packing algorithm.To the authors’best knowledge,no such an algorithm or such a combination of those techniques used has been previously reported for the packing of spheres.

In the remainder of the paper,the algorithmic description of the compression algorithm is presented.The key issues of the algorithm are discussed in detail.Filling spheres of various sizes within different geometric domains are illustrated via examples.

2.Algorithmic description

Consider the problem of randomly filling a domain with spheres of different sizes.The sphere radius r is randomly generated by a prescribed distributional function,and the do-main X can be any arbitrary polyhedron or simple shape,such as a cube or cylinder.The compression direction V g is specified.

Starting with an initial packing configuration,the algorithm proposed is a two-step procedure:(1)compress the initial packing;and(2)refill the space remaining,and if successful,compress the spheres by using the techniques in (1).Repeat the procedure until the domain is full.

There are different ways to create an initial packing within a geometric domain.The spheres can be randomly distributed or regularly positioned.All the spheres are checked against each other so that no overlap exists.

It is highlighted that no real force is involved in the packing procedure,therefore,it is a geometric-based approach.

https://www.wendangku.net/doc/a816211409.html,pression of a given packing

The basic idea of compressing a given packing can be stated as:if a sphere’s immediate neighbours are known,it can be moved along the given compression direction to a new position in touch with the first neighbouring sphere(s)by assuming that the neighbours are temporarily static.In one iteration after all spheres have been repositioned,a more tightened packing should be achieved.Such iterations are repeated until no further compression is possible.Apparently,determining the dynamically changing neighbours of a sphere,which is accomplished by a search algorithm,is the key to the success of the algorithm.

There exist two slightly different options in searching for the neighbourhood information of a sphere.In the first option,the neighbour list of a sphere is built,then followed immediately by repositioning before proceeding to the next sphere.As a result,the neighbour lists of subsequent spheres may be affected by the new positions of the spheres already processed.In the second option,the neighbour lists of all the spheres are obtained altogether at the beginning of the iteration,and the spheres are then compressed sequentially.To ensure that the neighbour information obtained in the beginning remains valid for the whole iteration,a bounding box is assigned to each sphere, and the movement of any sphere must be confined within the bounding box.Note that such a requirement is also necessary for the first option to essentially define the F close_neighbours of a sphere and to compute the maximum allowable moving distance.Consequently,the determination of the neighbour information of spheres is equivalent to the contact detection among spheres repre-sented by axis aligned bounding boxes.

Since more efficient contact detection algorithms are available for handling the contact among all the objects than the contact of one object with others at a time,the second option is therefore chosen.As contact detection comprises a major proportion of the total CPU time involved in the complete packing procedure,it will be addressed separately in the following section.

https://www.wendangku.net/doc/a816211409.html,pression

Prior to any iteration,a bounding box with buffer zone is assigned to each sphere.Issues relevant to the bounding boxes will be discussed later.

Once the neighbour list is built,the next step is to evaluate the maximum moving distance of each sphere along the compression direction.For the convenience of illustration,a2D problem is taken as an example as shown in Fig.1.Assume that V g=(t x,t y)T is the given compression direction with its normal n g=(n x,n y)T;C is the sphere to be compressed,and T i(i=1;...,n t,where n t is the number of neighbours of sphere C)is one of C’s neighbour spheres.The maximum moving distance of C just touching T i can be calculated as:

d i?tà

????????????????????????????

r ctr tT

e2às2

q

K.Han et al./Powder Technology155(2005)33–41 34

with

t?n y x ctàn x y ct;s?t y x ctàt x y ct

x ct?x càx t;y ct?y cày t

where r c and r t are respectively the radius of sphere C and T i;and(x c,y c),(x t,y t)their centre coordinates.By looping over all the neighbouring spheres,the moving distance d c of sphere C should be the smallest value of d i,i.e.,

d c?min d iT

e;i?1;...;n t

Nevertheless,as mentioned earlier,the movement of the sphere must be confined within the bounding box,or in other words,d c must not be greater than the size of buffer zone b u:

d c?b u if d c>b u

Furthermore,the computed distance should be checked with the boundary of the packing domain.If overlap occurs, the distance should be recomputed based on the shape of the boundary so that the sphere will just touch the boundary. Clearly any shape of boundary can be accommodated in the above procedure.The new position of the sphere,X c,is then updated as

X c@X ctd c V g

Within one iteration,each sphere will be moved a distance between[0,b u]along the compression direction. The procedure is repeated until no further movement is possible,or a certain convergence criterion is met.Here a F potential energy_based convergency check is employed,in which the F height_h e and the F potential energy_p e of any sphere are respectively defined as

h e?V g I X càX0T

e

p e?h e r3where X c are the centre coordinates of the sphere,and X0the coordinates of a reference point.The total F potential energy_ P e of the system involving all the spheres can thus be calculated as

P e?~p e

The procedure is terminated if the difference of potential energy between two consecutive iterations is less than a given tolerance s:

j P c e=P p eà1j

where P e c and P e p are respectively the total F potential energy_ of the spheres at the current and previous iteration steps.

The above algorithm is defined as global compression, and is outlined in Box1.

2.1.2.Shaking

If spheres are compressed in the given compression direction only,a highly chained configuration oriented to the direction is rapidly developed.As a result,only a limited compression can be achieved.Such a phenomenon is due to the inability of the above compression procedure to

move Box1

Global compression algorithm

Fig.1.Maximum moving of sphere C(its original position is shown by the

dotted line).

K.Han et al./Powder Technology155(2005)33–4135

the spheres elsewhere other than in the compression direction only,and thus is equivalent to a situation where particles are highly sticky.A simple scheme to overcome this difficulty is to introduce for each sphere a randomly generated local compression direction V l,which is different from the global direction V g,having a component normal to V g as defined by

V l?a V gtb V n with V g I V n?0;jj V n jj?1 where the normal direction V n is randomly generated,and a Z[0,1]and b Z[0,bˉ]are also two uniformly distributed random variables in the intervals.The value of b defines the angle of the cone where any local compression direction lies (Fig.2),i.e.,the maximum proportion of the normal component in the local compression direction.With the randomly generated local compression direction,the global compression algorithm is similarly employed to undertake compression by simply replacing the global compression direction V g with the local direction V l.This is termed random compression.

With the strategy stated above,the packing density is significantly increased.It is observed,however,that local arches are often formed which prevent further increase of packing density.A well-known physical process,shaking, is conceptually incorporated into the packing procedure to break the arches.To do so,it is only necessary to perform several additional iterations with compression along randomly generated local directions V h,which are perpendicular to the global compression direction,and the remainder of the compression procedure is exactly the same as that in Box1.As this compression does not change the total F potential energy_of the system,the convergence check is not applicable,but the number of iterations is specified instead.The procedure is referred to as shaking.

Consequently,the compression algorithm is the combi-nation of the three different compression procedures:global compression,random compression and shaking,which is summarised in Box2.

2.2.Refilling

After an initial packing has been compressed,further steps are required,if necessary,to fill the void created from compression until the whole domain is full.There are different methods to achieve this goal.In this work,a strategy called sphere insertion is adopted.First of all,the domain lid L is defined,which is usually one of the domain boundary surfaces opposite to the compression direction. The idea is to generate a new sphere on the lid,and then drop it along the compression direction to touch the nearest existing sphere(s).Such an insertion procedure can also be accomplished by applying the above compression algorithm to this single sphere.The generation and dropping proce-dure is repeated until no space is left to allow the insertion of new spheres.This process creates a new batch of spheres, which can be compressed by using the compression algorithm in Box2.

The complete filling domain algorithm is summarised in Boxes3and4

.

v n

=1)

g

( =1, =0)

α

β

β

Fig.2.Random compression

directions.

Box2

Box3

K.Han et al./Powder Technology155(2005)33–41 36

3.Contact detection

As highlighted earlier,contact detection plays a central role in the compression algorithm.In fact,tests show that contact detection itself comprises up to 70%of the CPU requirement for a complete packing.It is therefore crucial to employ highly efficient search algorithms to achieve an effective overall packing performance.

There are a number of contact detection algorithms available,in which cell or grid based is the most popular and has been extensively used in various applications due to its simplicity and efficiency.In recent years,further progress has been made,leading to the development of the effective and robust algorithm,termed here the dynamic cell based search algorithm ,or the F D-cell _for short.

In [8],a so-called no binary search (NBS)contact detection algorithm is introduced with the efficiency of total detection time T proportional to the total number of discrete objects N :T ?O N eT

thus exhibiting a superior performance,particularly for large-scale problems,to general tree based search algo-rithms [9]which usually have a computational complexity of

T ?O N ln N eT

The NBS,however,is limited by its applicability to systems comprising objects of identical size.

The D-cell is primarily designed to overcome this limitation and therefore can be applied to systems compris-ing discrete objects of various sizes.The main aspects of the algorithm is outlined below and the details can be found in [10].For the convenience of illustration,a 2D case is considered although extension to 3D problems is straight forward.

Assume a system consists of N discrete objects,represented by N bounding boxes,of different sizes occupying a bounding space [x min ,y min ]?[x max ,y max ].In the preliminary step,the space is virtually subdivided into n x cells in the x direction and n y cells in the y direction:

n x ?Int x max àx min

x m

;

n y ?Int

y max ày min

y m

where x m and y m are respectively user specified cell sizes in the x and y directions.Then each discrete object is mapped onto an integer grid cell (i x ,i y )according to its lower bounding box coordinates (x l ,y l ):

i x ?Int x l àx min x m ;i y ?Int

y l ày min

y m Thus each object is assigned to one and only one cell.An

example is given in Fig.3,where for instance,disk 1is assigned to the cell (3,4),and disk 2is assigned to the cell (3,2).An object is said to be mapped to a particular row if

it

Box 4

Fig.3.Mapping of discrete objects onto cells and the linked list.

K.Han et al./Powder Technology 155(2005)33–4137

is mapped to any cell of that row.Equivalently,an object is said to be mapped to a particular column if it is mapped to any cell of that column.

The obvious way to represent the mapping is the use of a 2D array of size n x ?n y .This is extremely expensive in terms of RAM requirements,especially for loose packing where the total number of discrete objects may be much smaller than the total number of cells.Similar to the NBS,linked lists are used here to avoid empty cells so as to reduce memory require-ments significantly.All the objects are mapped first to the rows of cells and a singly connected list Y i y (i y =1,...,n y )is established to represent objects in the i y th row.Objects are mapped by looping over all objects in ascending numerical order.In Fig.3,list Y 3(for Row 3)is formed by placing disk 1onto the list,and then pushing it by disk 2,which is pushed by disk 5,lastly by disk 8.

The rows are processed sequentially starting from the first non-empty row,and the cells of the row are then handled from the left to the right.Suppose that a non-empty row (e.g.,Row 3)is the row being processed with the corresponding linked list Y i y (Y 3).As will be explained later,Y i y (Y 3)may have been updated at the previous step,with objects whose upper-bound y -coordinates y u are

greater than the lower-bound y -coordinate of the current row.The modified linked list Y 3is:Y 3?1;2;5;8;7;14f g

where disks 7and 14are migrated from previous row (Row 2)since their upper-bound y -coordinates are greater than the lower y -coordinate of Row 3.Then the objects in list Y i y (Y 3)are mapped onto cells of the row by building the linked list X i x (i x =1,...,n x )according to their integerised coordinates i x .A particular (X i x ,Y i y )list contains all objects with integerised coordinates i x and i y .For example,(X 4,Y 3)contains disks 1and 14,while (X 5,Y 3)is empty.In addition,an ordered list is maintained to provide a sequential list of non-empty cells in the row.

Starting from the first cell in the ordered list,the cells in the row are processed sequentially following the order in the list;therefore,empty cells can be avoided.For the first non-empty cell i 1,the corresponding list X i 1may comprise objects originally in the cell and those migrated from the previous row.Overlap checks are performed among the original objects and against the migrated ones.No checks are needed,however,for the updated objects since they have been checked in previous rows.Before proceeding to the next cell,all the objects in the current cell are examined so that those with larger upper-bound x -coordinates than the lower-bound x -coordinate of the next cell in the ordered list are migrated and appended to the object list of the next cell.Therefore,for objects in Row 3(Fig.3),disk 5will be migrated to cell 4and disk 1will be migrated to cell 5.Consequently,the objects in the new cell to be processed may have four different types of membership:&Type 00:objects originally in the cell

&Type 01:objects originally in the row and migrated from the previous cell

Fig.4.Overlap checks among different types of objects.

Fig.5.Different phases of filling a cylindrical domain with spheres.

K.Han et al./Powder Technology 155(2005)33–41

38

&Type10:objects originally in the previous rows and migrated directly to the cell;

&Type11:objects originally in the previous rows,first migrated to the previous cells and then migrated to the cell from the previous cell.

With each object being one of the above four types, overlap checks in the cell are performed based on the following three cases,as illustrated in Fig.4:

1)among objects of type00;

2)Type00objects against objects of other types;

3)Type01objects against type10objects.

The row is completely processed when the last non-empty cell is reached.Before going to the next row,all the objects in the row,including the original and migrated ones, are checked so that any object with a greater upper-bound y-coordinate than the lower-bound y-coordinate of the next row is migrated to the new row.

The following remarks are made concerning the D-cell algorithm.(1)Although the cell size can be different from the discrete object sizes,the performance of the algorithm is strongly dependent on the cell dimension selected.The

optimal cell size is not easy to determine since it may depend on several factors.Our experience suggests that taking the cell size as three times for2D problems or five times for3D problems of the average object size often yields a good performance.(2)Mapping discrete objects to the columns first and processing columns sequentially is an equivalent choice to processing rows sequentially as described above.(3)The memory requirement is O(N) unless the cell size is extremely small.The computational time T may be expressed as

T?O Nte

eT

where e represents the costs associated with the maintenance of various lists used in the algorithm,which is related to the cell dimension chosen.

The D-cell contact detection algorithm is summarised in Box5.4.Other issues

The buffer zone used to extend the bounding box of each sphere plays an important role in the algorithm.The size of buffer zone has two competing effects:(1)it directly controls the length of the neighbour list in contact detection.The larger the buffer zone is,the longer the neighbour list will be,or the more expensive the contact detection;(2)it controls the maximum moving distances of spheres,particularly at the early stage of compression.The larger the buffer zone is,the faster the packing will be,i.e.,fewer iterations are needed.Therefore,the choice of the buffer zone size should balance the contact detection cost and the total number of iterations.

The bounding boxes are usually extended with the same buffer zone size along each coordinate direction.However, since the spheres are moved in certain directions only at different compression stages,it is prudent to extend the bounding boxes along these directions.For instance,a buffer zone is added only in the global compression direction in the global compression phase;and in the directions perpendicular to the global compression direc-tion during shaking.Further improvement can be achieved through adaptive control of the buffer zone size for each sphere in each packing phase.As the moving distances of spheres become less and less with the increase of packing density,the buffer zone size can also be reduced accordingly to reduce the cost associated with the contact detection.

At the later stage of shaking down,a large proportion of spheres may hardly be moved.This F mobility_of spheres

Box5

The D-cell contact detection algorithm

K.Han et al./Powder Technology155(2005)33–4139

can be monitored easily.Clearly,any sphere that was not moved at the previous iteration can only be moved if at least one of its neighbours has been repositioned.By applying this strategy,local movement checks can be avoided for those immobile spheres.

It is also found beneficial to move spheres according to their heights in the global compression direction,i.e.,to process spheres starting from the bottom and then upward.This can be explained by the fact that for a sphere to be processed,it is more likely to be moved with a larger distance if the lower level spheres in the neighbour list are moved first.

5.Illustrations

The validation of the proposed compression algorithm is tested via numerical examples.

Fig.5(a)–(h)illustrate different phases of filling a cylindric domain with spheres.The global compression direction is downwards from top to bottom of the cylinder,and the sphere radius is chosen to be uniformly distributed within [0:01,0:02].The initial packed 3071spheres,as shown in Fig.5(a),are randomly positioned in the cylinder,which is compressed by implementing the algorithm in Box 2,leading to a tightened packing (Fig.5(b)).To fill the void space created following the compression,a new batch of spheres (7261)are gener-ated,with green colour in Fig.5(c),and then compressed (Fig.5(d)).The procedure is repeated until no space is left to allow the insertion of more spheres.The final configuration is presented in Fig.5(h),with a total number of 26,787spheres and 52.89%packing density.It takes 181seconds on a 1.4-GHz PC to complete the packing.

Fig.6(a)and (b)also show the final configurations of filling two polyhedral domains with a total number of 51,036and 10,280spheres uniformly distributed within the range of [0.02,0.05]and [0.05,0.01],and taking about 413and 112seconds respectively.

6.Conclusion

A geometric-based approach,compression algorithm,has been presented in this work for the random filling of geometric domains with spheres of various sizes.It is motivated by the idea of compression complemented by an efficient physical process to increase packing density-shaking.By specifying a global compression direction,the compression procedure is the combination of three different strategies:global compression,random compression and shaking.As contact detection comprises a major proportion of the computational time for a complete packing,a dynamic cell based search algorithm,or D-cell ,has been employed to achieve an efficient packing performance.The compression algorithm is characterised by its simplicity and high efficiency,which has been demonstrated through test examples.

Acknowledgement

This work is funded by the EPSRC of UK under grant No.GR/R87222/01.This support is gratefully acknowledged.

Appendix A

Nomenclature a ,b Two random variables N The total number of spheres,or discrete objects C Sphere to be compressed T i C ’s neighbour spheres n t The number of neighbour spheres of C O c ,O t Centres of spheres C and T R c ,R t Radii of spheres C and T

(x c ,y c ,z c );(x t ,y t ,z t )Coordinates of O C and O T

X c =(x c ,y c ,z c )T ;X i =(x i ,y i ,z i )T Vector coordinates of

spheres C and i

(a)

(b)

Fig.6.Final configurations of filling polyhedral domains.Sphere radii uniformly distributed within:(a)[0.02,0.05];(b)[0.05,0.01].

K.Han et al./Powder Technology 155(2005)33–41

40

X0Coordinates of a reference point

vˉg=(t x,t y,t z)T(Global)compression direction

vˉn=(n x,n y,n z)T;nˉg=(n x,n y,n z)T;Unit directions perpen-dicular to vˉg

vˉl=(t x,t y,t z)T Local compression direction

d i,d c Moving(compression)distanc

e o

f C

b u Buffer zone size

h e Height of a sphere

p e Potential energy of a sphere

P e,P e c,P e p Total potential energy and energy at the current and previous iterations

s Convergence tolerance

n r,M r Current and maximum rejection numbers

T Total computational cost of a search algorithm [x min,y min]?[x max,y max]The maximum bounding range of all discrete objects

x m,y m Cell sizes in x and y directions

n x,n y Total number of cells in x and y directions

nn x,nn y Numbers of non-empty cells in a row or column x l,y l Lower bounding box coordinates of a object

i x,i y Cell indices

X i

x ,Y i

y

Linked lists

References

[1]J.D.Bernal,A geometrical approach to the structure of liquids,Nature

183(1959)141–147.

[2]G.D.Scott,Packing of spheres,Nature188(1960)908–909.

[3]W.M.Visscher,M.Bolsterli,Random packing of equal and unequal

spheres in two and three dimensions,Nature239(1972)504–507.

[4]J.L.Finney,Fine structure in randomly packed,dense clusters of hard

spheres,Mater.Sci.Eng.23(1976)199–205.

[5]J.H.Conway,N.J.A.Sloane,Sphere Packings,Lattices and Groups,

Springer-Verlag,Berlin,Germany,1988.

[6]Y.T.Feng,K.Han,D.R.J.Owen,Filling domains with disks:an

advancing front approach,Int.J.Numer.Methods Eng.56(2003)

699–703.

[7]Y.T.Feng,K.Han,D.R.J.Owen,An advancing front packing of

polygons,ellipses and spheres,3rd Int.Conf.on Discrete Element

Methods,Santa Fe,New Mexico,USA—Sept23–25,2002.

[8]A.Munjiza,K.R.F.Andrews,NBS contact detection algorithm for

bodies of similar size,Int.J.Numer.Methods Eng.43(1998)

131–149.

[9]Y.T.Feng,D.R.J.Owen,An augmented spatial digital tree algorithm

for contact detection in computational mechanics,Int.J.Numer.

Methods Eng.55(2002)159–176.

[10]E.Perkins,J.R.Williams,CGrid:neighbor searching for many body

simulation,4th Int.Conf.on Analysis of Discontinuous Deformation,

Glasgow,UK—June6–8,2001.

K.Han et al./Powder Technology155(2005)33–4141

霍尼韦尔(Honeywell)智能家居系统解决方案

龙源期刊网 https://www.wendangku.net/doc/a816211409.html, 霍尼韦尔（Honeywell）智能家居系统解决方案 作者： 来源：《物联网技术》2012年第11期 摘要：Honeywell HRIS-1000系统是基于TCP/IP协议和Ethernet网络平台的全数字化智能家居平台。该平台集成了丰富的居住环境控制及安防功能，而且各种功能可以协调统一，有机融合。文中介绍了Honeywell HRIS-1000系统的主要组成和系统功能。 关键词：智能家居；霍尼韦尔；解决方案；系统功能 具有120多年历史的霍尼韦尔（Honeywell）是世界自动化控制技术的领导者。霍尼韦尔自1980年起推出智能家居解决方案以来，陆续推出高集成社区规模智能家居系统家庭网关，可视对讲系统、无线灯光控制系统等。迄今为止，韩国已有200 000多户住宅使用霍尼韦尔的智能家居产品，在亚太其他地区，霍尼韦尔的集成智能家居解决方案也开创了许多成功案例，包括中国首个及最大的顶级智能社区深圳红树西岸、华北首个全集成智慧住宅天津赛顿中心、华东首个高集成智慧住宅杭州东方润园和中东迪拜的Old Town Commercial Island，霍尼韦尔 的可视对讲系统也正被广泛采用……事实上，全球有超过一亿个家庭和500多万幢大型建筑在使用霍尼韦尔的楼宇和住宅产品与技术。 1 智能住宅 随着国民经济和科学技术水平的提高，特别是计算机技术、通信技术、网络技术、控制技术、信息技术的迅猛发展与提高，促使了家庭生活的现代化，衣食住行的舒适化，居住环境的安全化。这些高科技已经影响到人们生活的方方面面，改变了人们生活习惯，提高了人们的生活质量，人类技术发展的最终目的和方向是服务于生活所需，智能住宅也正是在这种形势下应运而生的。 智能住宅是将家庭中各种与信息相关的通讯设备、家用电器、环境调节设备和家庭保安装置等，通过有线或无线网络连接到一个家庭智能化系统上进行集中的监视和智能控制，同时可以支持远程的监控，实现信息化家庭事务管理，并保持这些家用设施与住宅环境的和谐与协调。 2 系统介绍 Honeywell HRIS-1000系统是基于TCP/IP协议和Ethernet网络平台的全数字化智能家居平台。在这个平台上集成了丰富的居住环境控制及安防功能，各种功能协调统一，有机融合。家庭网关是户内控制和网络协议转换的中心，利用家庭网关使所有可能的设备信息互通，实现环

霍尼韦尔安防监控系统方案.(DOC)

霍尼韦尔健康舒适家居安防监控系统 钻石山设计方案 工程编号：20150331 二零一五年三月

霍尼韦尔家居系统优点 1、提供符合每位客户需求的定制解决方案 2、根据多年来的家居系统的经验和技术发展，提供可靠的解决方案 3、同一品牌的家居系统集成确保质量可靠、服务周到 霍尼韦尔家居系统的价值观 智能由于先进信息技术的广泛应用，霍尼韦尔的客户享受着更高标准的生活方式。 安全霍尼韦尔为您提供了最全面防范的安全体系，如高科技集成防盗和防火系统，门禁系统等。 舒适霍尼韦尔通过使用其全球顶尖的自动化控制技术，在水处理、空气质量、温度、湿度、安防、燃气安全、灯光和家居智能等方面，为您提供全面舒适的居住环境。 一、家庭安防监控系统简介： 近年来，随着我国经济的迅速发展，城乡居民的生活水平有了显著的提高，尤其是城镇居民的居住条件不断改善，人们在解决了居住问题后，日益关心的是居住是否安全，人们在购房时，安全性是考察物业管理水平是否完善的一个重要条件。尤其是那些流窜作案的犯罪分子，往往选择居民小区作为攻击目标，入室盗窃、抢劫、杀人案件屡屡发生，以往靠小区报安以人防为主的防范措施已满足不了人们的要求。利用安全防范技术进行安全防范首先对犯罪分子有种威慑作用，使其不敢轻易作案。如小区的安防系统、门窗的开关报警器能及时发现犯罪分子的作案时间和地点，使其不敢轻易动手。 家庭安防应有两重含义，一是指生命安全，二是指财产安全。 传统家庭安防是安装被动红外探测器，其特点是安装在室内，因此对室内无人值守时，其防盗保护财产的作用能够实现。而新型智能家庭防盗报警系统在感应端通过磁头、红外、煤气、烟感、玻璃破碎等探测器来感应异常变化，当感应器感应到异常情况，就会自动报警并把警情发送至相关部门和人员的电话上，以便得到及时的处理，减少人员伤亡和财产损失。 目前，家庭监控的业务模式是IP摄像机+各类报警探测设备+平台接入。在这样的模式下，用户离开时布防，发生意外情况时触发报警，报警信息以短信形式及时发送给用户，同时启动警铃并联动IP摄像机进行预置位转动、抓拍图片、启动录像等程序。用户收到短信后通过PC或手机登录并实时浏览现场视频，确认后采取处置措施，用户回来后撤防。 而对于家庭安防系统的建立，现阶段还需要专业队伍的安装和服务。目前，家庭安防工程主要有房产开发商委托安防工程商来完成。业主购房后若是要进行装修，就需要工程商为

BAS系统在地铁环境控制中的应用及实现

BAS系统在地铁环境控制中的应用及实现 发布时间：2009-7-20 文章来源：本站 1 概述 广州地铁一号线共有14个地下车站、2个地面车站和一座地铁控制中心（OCC）大楼，全长18.6公里，采用了集散控制系统（DCS）对地铁全线环控设备及其它车站机电设备进行集中监控，由于引进了楼宇控制概念，地铁车站设备监控系统亦被称为BAS（Building Automation System）系统。广州地铁一号线采用美国CSI公司的I/NET2000系统对全线环控系统进行监控，并对全线车站的扶梯、给排水设备、应急电源进行监视报警。 2 BAS系统在地铁环控中的作用及功能 2.1. 地铁BAS系统在地铁环控中的主要作用： 控制全线车站及区间的环控及其它机电设备安全、高效、协调的运行，保证地铁车站及区间环境的良好舒适，产生最佳的节能效果，并在突发事件（如火灾）时指挥环控设备转向特定模式，为地铁乘车环境提供安全保证。 2.2. 广州地铁一号线BAS系统主要功能： (1) 监控并协调全线各车站及OCC大楼通风空调设备、冷水系统设备的运行。 (2) 监控并协调全线区间隧道通风系统设备的运行。 (3) 对车站机电设备故障进行报警，统计设备累积运行时间。 (4) 对全线环境参数（温、湿度）及水系统运行参数进行检测、分析及报警。 (5) 接收地铁防灾系统（FAS系统）火灾接收报警信息并触发BAS系统的灾害运行模式，控制环控设备按灾害模式运行。

(6) 通过与信号ATS接口接收区间堵车信息，控制相关环控设备执行相应命令。 (7) 紧急状况下，可通过车站模拟屏控制环控设备执行相关命令。 (8) 监视全线各站及隧道区间给排水、自动扶梯等机电设备的运行状态。 (9) 管理资料并定期打印报表。 (10) 与主时钟接口，保证BAS系统时钟同步。 3 BAS系统对环控设备的监控原理及内容： 3.1. 环控系统组成： 大系统——车站公共区（站厅/站台）通风空调系统； 小系统——车站设备用房通风空调系统； 水系统——地下站冷水机组系统； 隧道通风系统——执行隧道区间正常及紧急情况下通风排烟工况的环控子系统。 3.2. BAS系统监控点数的配置： 以陈家祠站为例，纳入BAS监控的环控设备总数约100台（包括风机、风阀和水系统设备等），环控监控总点数约430点（包括温湿度等参数检测约60点），车站监控点数分布情况如下： (1) 隧道通风系统：BAS系统对4台隧道风机及联动风阀、两台推力风机和组合风阀进行监视控制，监视风机过载故障报警信号，检测两端隧道入口温湿度，共计点数DO 20点、DI 28点，AI 8点 (2) 车站大通风空调系统：BAS系统对空调机、新风机、回排风机及联动风阀和调节风阀等设备进行监视控制，监视风机过载故障报警信号，检测新/排/混/送风及站厅/台温湿度，控制组合风柜出水二通阀开度来调节空调器送风温度，共计DO 44点、DI 72点，AI 30点、AO 4点

霍尼韦尔智能家居系统解决方案

霍尼韦尔智能家居系统

地区总经销商鸣迅智能科技有限公司

目录 一、公司背景┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅2 二、智能住宅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅3 三、系统介绍┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅5 四、系统功能┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅6 1．系统整体介绍‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐6 2．防盗报警‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐6 3．灯光窗帘控制‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐9 4．地板采暖控制‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐10 5．空调控制‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐11 6．远程视频监控‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐12 五、主要产品说明┅┅┅┅┅┅┅┅┅┅┅┅┅┅13 1．控制面板‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐14 2．智能家居主机‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐15 3．智能控制模块‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐16 六、产品优势┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅19 1．品牌的优势‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐19 2．高度集成化的系统‐‐‐‐‐‐‐‐‐‐‐‐‐‐19 3．开放系统，国际标准的通讯协议‐‐‐‐‐‐‐‐20 4．稳定的产品质量‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐20 七、具体案┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅20 八、案例介绍┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅20 九、售后服务承诺┅┅┅┅┅┅┅┅┅┅┅┅┅24 1．售后服务主导思想坚持质量第一，用户至上的精神维护本公司的声誉，确保工程项目及产品售后服务发挥其应有的效能‐24

2．售后服务围‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐24 3．实施办法‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐24 一、公司背景 霍尼韦尔（Honeywell）公司是一家年销售额达300 亿美元，在多元化科技和制造业领域占据世界领导地位的跨国公司。在全球，其业务涉及航空产品及服务、住宅及楼宇控制和工业控制技术、自动化产品、特种化学、纤维、塑料以及电子和先进材料等领域。霍尼韦尔公司在全球95 个拥有10.8 万员工，总部设在美国新泽西州莫里斯镇。在纽约、伦敦和芝加哥太平洋证券市场的交易代码为HON。为道琼斯工业指数的30 家构成公司之一，也是“标准普尔500 指数”的组成部分。霍尼韦尔具有百年的历史，今天的霍尼韦尔一如既往地把创造一个安全、舒适、节能、高效、创新的人类生存空间作为Honeywell 品牌的第一承诺。霍尼韦尔公司以诚信的态度、优质的产品、精湛的服务和客户至上的原则，一步一个脚印地在中国市场辛勤耕耘、拓展。作为一家多元化的跨国公司，霍尼韦尔正在源源不断地将其各个部门的顶尖技术和产品带到中国。

地铁BAS系统现场网络结构的说明V1

关于地铁BAS系统现场级网络应用的说明 1、概述 地铁BAS系统作为综合监控系统的重要组成部分承担着地下车站机电设备监控以及紧急情况下防灾救灾的重责。由于地下车站机电设备分布广泛，因此BAS系统核心控制器及远程IO之间一般通过网络通信的形式连接。随着城市轨道交通技术的发展，国内外地铁环境与设备监控系统已经走过了各站分离的阶段，进入了全线组网的新阶段，设备监控多采用分散控制、集中管理的系统模式。目前BAS系统现场级网络主要有全总线和工业以太网两种实现形式。 由于现场总线技术的各种标准之间转换困难、系统集成存在各种壁垒等种种制约性，而相对的工业以太网的种种优势，随着全球工业自动化技术的不断进步，造成了BAS系统网络正在从现场总线向工业以太网方向发展的趋势。 2、工业以太网与现场总线比较 目前国内城市轨道交通BAS系统普遍采用PLC设备，是一个基于网络的自动化系统，涉及多种通信及网络技术，如用于装置控制层的现场总线技术。而由于现场总线标准存在12种之多，如何统一现场总线标准经过了16年的标准大战，最终没有形成一个统一的标准，多标准等于无标准，因此无论是最终用户还是制造商，普遍都在关注现场总线技术的发展动态，寻求高性能低成本的方案。以太网技术由于其开放性、稳定性和可靠性，在全球范围取得了巨大成功，因此如何对以太网技术进行改进，使

其适合应用于工业控制领域的数字通信，已成为业内近些年内的热门研究方向，很多人都寄希望于现场总线技术在以太网技术的基础上达成统一，改变目前多标准并存的现状，同时用以太网统一工业控制网络的各个层次，实现真正的无缝信息集成。BAS系统网络也随着工业以太网的发展，逐渐实现装置控制层设备由采用现场总线改变为工业以太网技术。 1） BAS系统采用工业以太网方案对比传统的总线方案具有以下优点： 传统双现场总线方案中，车站两端冗余PLC各自负责一端的BAS系统设备。对于车站内需要联动运行的部分设备，如正常模式下分布在车站不同端的风机、风阀联动、火灾模式下的两端空调系统联动等均需要两端的冗余PLC之间首先相互联动和确认设备状态到位后才能执行下一步动作。在常规地铁设计中，车站两端的冗余PLC虽然采用了热备方式，配置了两块背板、两块CPU、两块电源等，但所有的模块均放置在同一房间甚至同一面控制柜内，当房间内发生火灾或电源故障，容易引起冗余PLC整体故障。而一端的冗余PLC一旦退出服务，则另一端的冗余PLC则可能因为联锁动作失败而导致系统整体瘫痪。若采用光纤环网方式连接两端冗余PLC，若一端冗余PLC发生整体性故障退出服务，系统将立即切换到另一端的一套冗余PLC上继续工作，保证系统在极端恶劣的情况下能正常运行，中央和车站下达的指令能迅速传达到现场设备。 传统双现场总线方案中，双总线均采用平行布线方式，两条总线紧靠着

南京地铁BAS系统设计与应用

地铁BAS系统设计与应用 楼宇智能化系统所涉及的容众多。采用智能化系统分散管理。BAS系统利用计算机编程及网络通信技术，对这些设备的测量控制点进行集中管理和自动监测，对减少运行、操作、维护人力，保持设备的正常运转。地铁BAS（Building Automation System）本着“安全、可靠、节能”的原则进行设计，将现代科技的计算机及网络技术结合机电设备自动化控制原理，以专门的地铁环境通风空调及防灾处理等理论为基础的自动化控制系统，利用分布式微机监控系统对地铁车站及区间隧道的空调通风、给排水、照明、电梯、自动扶梯、导向标识等机电设备进行全面的运行管理与控制，在发生火灾或列车阻塞等事故情况时，能够及时迅速地进入防灾运行模式，根据火灾报警系统发送的着火点信息或列车自动控制系统发送的阻塞点信息自动调度送风和排风，进行通风排烟，引导人员疏散，极提高地铁运营的智能化和安全性。BAS可通过采用前反馈、后反馈众多调控形式进行实时在线运行与自动控制，并将在保证地铁热环境控制要求前提下，实现设备自动、稳定、安全、节能的运行。关键词PLC 楼宇自动化通风空调系统 目次 1 概述1 2 地铁1号线BAS系统监控对象及功能3 2.1 设计原则3

2.2 地铁1号线环境与设备监控对象4 2.3 BAS系统主要功能4 2.4 BAS系统的接口6 3 地铁1号线BAS系统的软件体系7 3.1 BAS系统软件的组成7 3.2 地铁BAS系统采用的第三方软件8 3.2.1 环境优化控制软件9 3.2.2 BAS与FAS通讯软件9 3.2.3 故障管理软件9 4 地铁BAS系统构成及网络结构9 4.1 BAS系统的构成9

地铁BAS系统

地铁BAS系统 2003年5月，国家质量监督检验检疫总局和建设部，联合发布了国家标准——《GB 50157-2003地铁设计规范》，标准中正式命名“环境与设备监控系统，Building Automation System（BAS）”，并对其定义为：“对地铁建筑物内的环境与空气条件、通风、给排水、照明、乘客导向、自动扶梯及电梯、屏蔽门、防淹门等建筑设备和系统进行集中监视、控制和管理的系统”。 基本功能： 1.机电设备监控 具有中央和车站二级监控功能； BAS控制命令应能分别从中央工作站、车站工作站和车站紧急控制盘（IBP）人工发布或由程序自动判定执行，并具有越级控制功能，以及所需的各种控制手段； 对设备操作的优先级遵循人工高于自动的原则； 具备注册和权限设定功能。 2.执行防灾及阻塞模式功能 能接收FAS系统车站火灾信息，执行车站防烟、排烟模式； 能接收列车区间停车位置信号，根据列车火灾部位信息，执行隧道防排烟模式； 能接收列车区间阻塞信息，执行阻塞通风模式； 能监控车站逃生指示系统和应急照明系统； 能监视各排水泵房危险水位。 3.环境监控与节能运行管理功能 通过对环境参数的检测，对能耗进行统计分析，控制通风、空调设备优化运行，通过地铁整体环境的舒适度，降低能源消耗。 4.环境和设备管理功能 能对车站环境等参数进行统计； 能对设备的运行状况进行统计，据此优化设备的运行，实施维护管理趋势预告，提高设备管理效率。 地铁BAS监控内容： 正常运营模式的判定及转换； 消防排烟模式和列车阻塞模式的联动； 设备顺序启停； 风路和水路的联锁保护； 大功率设备启停的延时配合； 主、备设备运行时间平衡； 车站公共区和重要设备房的温度调节； 节能控制； 运行时间、故障停机、启停、故障次数等统计； 配置数据接口以获取冷水机组和水系统相关信息； 若冷水机组带有联动控制功能，则空调水系统冷冻水泵、冷却水泵、冷却塔、风机、电动蝶阀的控制程序由冷水机组承担，BAS仅控制冷水机组的投切、监测空调系统的参数和状态、冷量实时运算、记录及累计。

BAS系统在地铁环境控制中的应用及实现

1 概述 广州地铁一号线共有14个地下车站、2个地面车站和一座地铁控制中心（OCC）大楼，全长18.6公里，采用了集散控制系统（DCS）对地铁全线环控设备及其它车站机电设备进行集中监控，由于引进了楼宇控制概念，地铁车站设备监控系统亦被称为BAS（Building Automation System）系统。广州地铁一号线采用美国CSI公司的I/NET2000系统对全线环控系统进行监控，并对全线车站的扶梯、给排水设备、应急电源进行监视报警。 2 BAS系统在地铁环控中的作用及功能 2.1. 地铁BAS系统在地铁环控中的主要作用： 控制全线车站及区间的环控及其它机电设备安全、高效、协调的运行，保证地铁车站及区间环境的良好舒适，产生最佳的节能效果，并在突发事件（如火灾）时指挥环控设备转向特定模式，为地铁乘车环境提供安全保证。 2.2. 广州地铁一号线BAS系统主要功能： (1) 监控并协调全线各车站及OCC大楼通风空调设备、冷水系统设备的运行。 (2) 监控并协调全线区间隧道通风系统设备的运行。 (3) 对车站机电设备故障进行报警，统计设备累积运行时间。 (4) 对全线环境参数（温、湿度）及水系统运行参数进行检测、分析及报警。 (5) 接收地铁防灾系统（FAS系统）火灾接收报警信息并触发BAS系统的 灾害运行模式，控制环控设备按灾害模式运行。 (6) 通过与信号ATS接口接收区间堵车信息，控制相关环控设备执行相应命令。 (7) 紧急状况下，可通过车站模拟屏控制环控设备执行相关命令。

(8) 监视全线各站及隧道区间给排水、自动扶梯等机电设备的运行状态。 (9) 管理资料并定期打印报表。 (10) 与主时钟接口，保证BAS系统时钟同步。 3 BAS系统对环控设备的监控原理及内容： 3.1. 环控系统组成： 大系统——车站公共区（站厅/站台）通风空调系统； 小系统——车站设备用房通风空调系统； 水系统——地下站冷水机组系统； 隧道通风系统——执行隧道区间正常及紧急情况下通风排烟工况的环控子 系统。 3.2. BAS系统监控点数的配置： 以陈家祠站为例，纳入BAS监控的环控设备总数约100台（包括风机、风阀和水系统设备等），环控监控总点数约430点（包括温湿度等参数检测约60点），车站监控点数分布情况如下： (1) 隧道通风系统：BAS系统对4台隧道风机及联动风阀、两台推力风机和组合风阀进行监视控制，监视风机过载故障报警信号，检测两端隧道入口温湿度，共计点数DO 20点、DI 28点，AI 8点 (2) 车站大通风空调系统：BAS系统对空调机、新风机、回排风机及联动 风阀和调节风阀等设备进行监视控制，监视风机过载故障报警信号，检测新/排/混/送风及站厅/台温湿度，控制组合风柜出水二通阀开度来调节空调器送风温度，共计DO 44点、DI 72点，AI 30点、AO 4点 (3) 车站小通风空调系统：BAS系统对空调机、送/排风机及联动阀、调节阀监视控制，检测设备/管理用房温湿度，控制小空调器出水二通阀开度来调节相关设备房的温度，共计DO 41点、DI 41点，AI 17点、AO 3点

霍尼韦尔智能家居系统解决方案

霍尼韦尔智能家居系统上海地区总经销商上海全宜智能科技有限公司

目录 一、公司背景 ------------------------------------------------------------------------------ 3 二、智能住宅 ------------------------------------------------------------------------------ 4 三、系统介绍 ------------------------------------------------------------------------------ 5 四、系统功能 ------------------------------------------------------------------------------ 6 1．系统整体介绍 (6) 2．防盗报警 (6) 3．灯光窗帘控制： (10) 4．地板采暖控制： (11) 5．空调控制： (12) 6．远程视频监控： (121) 五、主要产品说明 ---------------------------------------------------------------------- 14 1、控制面板 (14) 2、智能家居主机 (16) 3、智能控制模块 (17) 六、产品优势 ---------------------------------------------------------------------------- 20 1、品牌的优势： (20) 2、高度集成化的系统 (20) 3、开放系统，国际标准的通讯协议 (21) 4、稳定的产品质量 (21) 七、具体方案 ---------------------------------------------------------------------------- 22 八、案例介绍 ---------------------------------------------------------------------------- 22 九、售后服务承诺 ---------------------------------------------------------------------- 26 1.售后服务主导思想坚持质量第一，用户至上的精神，维护本公 司的声誉，确保工程项目及产品售后服务发挥其应有的效能。 26 2.售后服务范围 (26) 3.实施办法 (26)

浅谈城市轨道交通BAS系统的发展

浅谈城市轨道交通BAS系统的发展 发表时间：2011-12-28T13:33:37.657Z 来源：《时代报告》2011年11月下期供稿作者：陶汉卿[导读] 我国城市轨道交通系统迅速引入了基于计算机技术、自动控制技术和网络通信技术的各类自动化系统。 陶汉卿 (柳州铁道职业技术学院，广西柳州 545007) 中图分类号：U29 文献标识码：A 文章编码：1003-2738（2011）11-0022-01摘要：介绍现阶段BAS系统的结构和构成，探讨BAS系统的标准化设计、系统优化方法及系统评价方法，以摆脱其在发展过程中所面临的 技术问题，实现可持续发展。关键词：城市轨道交通；BAS系统；问题；发展一、引言 随着我国城市轨道交通的规模化、高速化发展，我国城市轨道交通系统迅速引入了基于计算机技术、自动控制技术和网络通信技术的各类自动化系统，大量采用国际先进水平的现代化机电设备，其中城市轨道交通环境与设备监控系统(BAS系统)就是其中之一，具有体系复杂、技术含量高、专业面广、设备维护困难的特点，并且需要根据业务需求不断地进行更新改造。我国发布的《GB 50157-2003地铁设计规范》中正式将该系统命名为“BAS，环境与设备监控系统”，并对其定义为：“是对地铁建筑物内的环境与空气条件、通风、给排水、照明、乘客导向、自动扶梯及电梯、屏蔽门、防淹门等建筑设备和系统进行集中监视、控制和管理的系统”。城市轨道交通BAS系统是一个典型的集成开放系统，是确保城市轨道交通系统安全、快捷、准点、有效地运行的关键工艺系统，是城市轨道交通中不可缺少的一个重要组成部分。 二、BAS系统现有的系统结构与构成 BAS系统从系统组成而言包括中心BAS系统、车辆段BAS系统和车站BAS系统，完整的BAS系统或完整的BAS功能系统是一个以骨干网为基础的、地理上分散的、分层分布式系统结构的大型SCADA系统，从逻辑上讲，硬件系统包括纵向3个层次。（一）中央级监控系统。主要位于OCC，由中央实时服务器、中央历史服务器、操作员工作站、工程师工作站、打印设备、网络设备、大屏幕或模拟显示设备等计算机及网络硬件构成，软件则包括操作系统、大型数据库、系统应用软件、应用软件开发与维护平台、网管软件其它辅助软件等。（二）车站级监控系统。车站级监控系统位于车站，以车站监控工作站、PLC控制为基础，具体包括车站监控局域网、打印机、后备操作盘等设备。（三）现场控制级设备。位于车站各就地监控点或数据采集点，具体包括各类传感器、执行器、远程I/O模块、接口模块或装置等。 BAS系统在横向呈现分布式的集散型结构，包括两个方面：各个车站的BAS系统因为车站沿城市轨道交通线路呈地理上分布式结构，因此整个BAS也是以车站BAS为单位的地理上分散的SCADA系统，另外在车站，根据设计规范的要求，车站BAS由多个控制器和统一的监控设备构成一个集散型系统（DCS）。（四）软件结构。软件结构包括数据结构层、数据处理层和人机接口层。管理和处理各种数据、接口控制以及提供信息显示和操作界面。（五）BAS系统的应用。 BAS系统的结构分为有中心功能的结构、无中心功能的结构和混合结构。有中心功能的BAS系统是一种完全独立的系统结构，较为传统和经典，目前建成或在建线路的 BAS 系统大多采用这种结构方式，如南京地铁 1号线，天津地铁，广州地铁2号线等；无中心功能的BAS系统是一种不完整的结构形式，这种结构形式的BAS是以车站为单位的一个个相对独立的系统，如广州地铁3号线，北京地铁5号线；混合式BAS系统既要在车站和综合监控系统接口，同时又要通过地铁骨干网形成一个较完整的BAS系统，如广州地铁4号线。 三、BAS系统在当前发展中所面临的问题城市轨道交通自动化程度将会越来越高，同样地，BAS系统将会更加复杂，监控的设备更多，系统集成的需求更大，技术含量更高，专业面更广，设备维护更加困难。目前，BAS系统存在以下问题：（1）系统与设备、设备与设备之间的控制集成成功率不高，相关系统结合“接口”界面如通信协议、网络构架的标准化、统一性不够；（2）运营管理水平跟不上，没有充分进行运行优化；（3）缺少正确有效的城市轨道交通BAS系统的评估方法，限制了城市轨道交通BAS系统研究更好地开展。 四、BAS系统的发展事物总是在矛盾中不断向前发展的。为了有效地解决城市轨道交通BAS系统在当前发展过程中存在的问题，BAS系统具有以下的发展趋势，从而实现BAS系统的可持续发展。（一）标准化设计。 BAS系统是一个对若干设备进行控制和监控管理、并基于“设备”、面向乘客服务的系统，其开放的集成系统构架，已经为构建城市轨道交通综合自动化系统奠定了基础，如果淡化专业系统概念，扩大BAS外延，即演变成一个综合自动化IAS系统，该系统相对于BAS系统而言，只是增加了多项专业功能和服务而已。因此，面向“设备”监控和管理的思路，基于开放的BAS集成系统构成，构建城市轨道交通综合自动化系统，已成为目前的发展趋势。集成监控平台的搭建也是BAS系统标准化设计的一个条件，监控平台应形态合理，提供强大的应用开发接口，数据组织和展现方式应符合城市轨道交通监控的习惯和特点，支持不同方式的硬件集成环境和软件配置形态，运用冗余、容错、自恢复等技术充分保证系统的稳定运行。 （二）系统优化。

地铁BAS系统网络介绍—工业以太网方式

地铁BAS系统网络构架介绍----工业以太网方式 ●BAS系统介绍： 地铁BAS系统对地铁各个车站及停车场、车辆段的暖通空调系统、给排水、低压配电与动力照明系统、电梯系统、车站事故照明电源等车站设备进行全面、有效地进行自动化监控及管理，确保设备处于安全、可靠、高效、节能的最佳运行状态，从而提供一个舒适的乘车环境，并能在火灾或阻塞等灾害状态下，更好地协调车站设备的运行，充分发挥各种设备应有的作用，保证乘客的安全和设备的正常运行。 关键词：BAS --- 环境与设备监控系统 FAS ---火灾自动报警系统 HMI ---人机界面 ISCS---综合监控系统 PLC ---可编程序控制器 UPS ---不间断电源 ZPLC--- 专用PLC：特指各区间的水泵房、风机， 线路外侧的冷冻站内设置的PLC。 维修工作站 --BAS 的车站级，作为BAS 的维修操作终端。 WINCC--- 西门子监控系统软件 IBP --- Integrated Backup Panel（综合后备盘） ●BAS系统站级网构架（2层网络） 第一层网络（工业太网络）：系统根据车站建筑形式分为南端与北端BAS子系统。车站的环境与设备监控系统网分二层布置，第一层为站级系统网络，采用工业以太网，担负BAS与ISCS,南北端PLC 间、维修工作站同南北端PLC的数据交换。一般采用网管型工业交换机配置为冗余的双环以太网。 第二层网络（现场总线）：第二层为现场设备级网络，采用专业工业现场总控冗余工业控制网，担负BAS控制器与BAS现场设备的数据交换。 网络构架图

注：BAS系统采用双网双设备冗余，但对站级工业以太网络需要接入设备并不多，主要是车站南北两端的PLC、站级维护工作站、ISCS FEP等。一般情况下，一个站6台左右2多模光口、6个电口网管型工业交换机可以满足需求。

霍尼韦尔智能家居系统解决方案-公开概要

霍尼韦尔智能家居系统HRIS-1000系列智能家居主 机 黄浦湾一期解决方案 目录 公司背景UUU ------------------------------------------------- 3 UUU 、 ------------------------------------------------ 4 UUU二、智能住宅UUU ------------------------------------------------ 5 UUU三、系统介绍UUU ------------------------------------------------ 6 UUU四、系统功能UUU ?系统整体介绍UUU (6 UUU1 .防盗报警UUU (6 UUU2

.灯光窗帘控制 :UUU (9 UUU3 .地板采暖控制 :UUU (10 UUU4 .空调控制 :UUU (11 UUU5 .远程视频监控 :UUU ........................................... 1U2 UUU6 ------------------------------------------ 13 UUU 五、主要产品说明 UUU 、控制面板 UUU (14 UUU1 、智能家居主机 UUU (15 UUU2 、智能控制模块 UUU (16 UUU3 ---------------------------------------------- 19 UUU 六、产品优势 UUU 、品牌的优势 :UUU (19 UUU1 、高度集成化的系统 UUU (19

BAS系统的内容及功能

BAS系统 地铁BAS系统： 2003年5月，国家质量监督检验检疫总局和建设部，联合发布了国家标准——《G B 50157-2003地铁设计规范》，标准中正式命名“环境与设备监控系统，Building Aut omation System（BAS）”，并对其定义为：“对地铁建筑物内的环境与空气条件、通风、给排水、照明、乘客导向、自动扶梯及电梯、屏蔽门、防淹门等建筑设备和系统进行集中监视、控制盒管理的系统”。 基本功能： 1.机电设备监控 具有中央和车站二级监控功能； BAS控制命令应能分别从中央工作站、车站工作站和车站紧急控制盘（IBP）人工发布或由程序自动判定执行，并具有越级控制功能，以及所需的各种控制手段； 对设备操作的优先级遵循人工高于自动的原则； 具备注册和权限设定功能。 2.执行防灾及阻塞模式功能 能接收FAS系统车站火灾信息，执行车站防烟、排烟模式； 能接收列车区间停车位置信号，根据列车火灾部位信息，执行隧道防排烟模式； 能接收列车区间阻塞信息，执行阻塞通风模式； 能监控车站逃生指示系统和应急照明系统； 能监视各排水泵房危险水位。 3.环境监控与节能运行管理功能 通过对环境参数的检测，对能耗进行统计分析，控制通风、空调设备优化运行，通过地铁整体环境的舒适度，降低能源消耗。 4.环境和设备管理功能 能对车站环境等参数进行统计； 能对设备的运行状况进行统计，据此优化设备的运行，实施维护管理趋势预告，提高设备管理效率。 地铁BAS监控内容： 正常运营模式的判定及转换； 消防排烟模式和列车阻塞模式的联动； 设备顺序启停； 风路和水路的联锁保护； 大功率设备启停的延时配合； 主、备设备运行时间平衡； 车站公共区和重要设备房的温度调节； 节能控制；

霍尼韦尔智能家居系统解决方案

霍尼韦尔智能家居系统山东地区总经销商山东鸣迅智能科技有限公司

目录 一、公司背景┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅2 二、智能住宅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅3 三、系统介绍┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅5 四、系统功能┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅6 1．系统整体介绍‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐6 2．防盗报警‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐6 3．灯光窗帘控制‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐9 4．地板采暖控制‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐10 5．空调控制‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐11 6．远程视频监控‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐12 五、主要产品说明┅┅┅┅┅┅┅┅┅┅┅┅┅┅13 1．控制面板‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐14 2．智能家居主机‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐15 3．智能控制模块‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐16 六、产品优势┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅19 1．品牌的优势‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐19 2．高度集成化的系统‐‐‐‐‐‐‐‐‐‐‐‐‐‐19 3．开放系统，国际标准的通讯协议‐‐‐‐‐‐‐‐20 4．稳定的产品质量‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐20 七、具体方案┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅20 八、案例介绍┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅20 九、售后服务承诺┅┅┅┅┅┅┅┅┅┅┅┅┅24 1．售后服务主导思想坚持质量第一，用户至上的精神维护本公司的声誉，确保工程项目及产品售后服务发挥其应有的效能‐24 2．售后服务范围‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐24 3．实施办法‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐24