AU679944B2 - A structure design method and system - Google Patents

Description

Regulation 3.2

AUSTRALIA

Patents Act 1990 COMPLETE SPECIFICATION FOR A STANDARD PATENT

(ORIGINAL)

p.

p P. l p.

p pe p p p *D b P C

.P

Name of Applicant: TELSTRA CORPORATION LIMITED, A.C.N. 051 775 556, of 242 Exhibition Street, Melbourne 3000, Victoria, Australia Actual Inventors: Herman Lucas FERRA, Adam KOWALCZYK and Teck Hoe TAN Address for Service: DAVIES COLLISON CAVE, Patent Attorneys, of 1 Little Collins Street, Melbourne 3000, Victoria, Australia Invention Title: "A STRUCTURE DESIGN METHOD AND SYSTEM" Details of Associated Provisional Application No: PL6503/92 The following statement is a full description of this invention, including the best method of performing it known to us: -1- I I A STRUCTURE DESIGN METHOD AND SYSTEM The present invention relates to a structure design method and system.

Structures, such as towers and masts, for supporting equipment, such as transmission antennas, are presently designed by structural engineers on a relatively ad hoc basis. Whilst conventional structural enginecring design techniques are employed to calculate probable loads in various weather conditions the techniques are purely theoretical and the actual loads experienced by a structure with various combinations of equipment attached thereto may be significantly different when the structure is erected at the predetermined site. Therefore the design and construction of the structures has, in the past, been performed solely by experts who have relied on a combination of basic theoretical rules and practical experience to provide sound structures. The same experts are also consulted whenever further equipment needs to be added to a structure.

For a telecommunications industry to rely solely on a relatively small number of 20 experts for communications structure design is inefficient, and so it is desired to provide a method and system which would allow any party wishing to add equipment to a structure to obtain an indication as tc the feasibility of the addition.

In accordance with the present invention there is provided a structure design 25 method using a computer graphics processing system, said method comprising: generating at least one view of a structure; graphically selecting a position on said view for an object; displaying said object on said structure at said position; determining load parameters for said structure with said object in said position, where said load parameters represent loads on selected parts of said structure; and 4 zdetermining on the basis of said load parameters, and displaying, at least one indicator of the soundness of said structure.

P:\OPERDBW52658.93 12/5/97 -3- *cc *c S

S

S

*c S *5 c..

c. S c c *c S Re S Re S. S c.

Re etcc The present invention also provides a structure design method, comprising: selecting a position on a structure for an object; determining load attributes for said structure with said object in said position, where said load attributes represent loads on selected segments of said structure; and executing a neural network algorithm on the basis of said attributes to obtain at least one indicator of the soundness of said structure.

The present invention further provides a structure design system comprising: means for generating at least one view of a structure; 0 means for graphically selecting a position on said view for an object; means for displaying said object on said structure at said position; means for determining load parameters for said structure with said object in said position, where said load parameters represent loads on selected parts of said structure; and 5 means for determining on the basis of said load parameters, and displaying, at least one indicator of the soundness of said structure.

The present invention also provides a structure design system comprising: means for selecting a position on a structure for an object; 0 means for determining load attributes for said structure with said object in said position, where said load attributes represent loads on selected segments of said structure; and means for executing a neural network algorithm on the basis of said load attributes to obtain an indicator of the soundness of said structure.

Preferred embodiments of the present invention are hereinafter described, by way of example only, with reference to the accompanying drawings, wherein: Figure 1 is a display produced by a graphical user interface of a first preferred embodiment of a structure design system; 0 Figure 2 is a diagram of a neural network structure used by the system; Figure 3 is a main display produced by a second preferred embodiment of a structure design system; C- C.

Su S JN* P:\OPER\DBW2658,93 12/5I -4- Figure 4 is a zoom view display produced by the structure design system of Figure 3; Figure 5 is an antenna selection display produced by the structure design system of Figure 3; Figure 6 is a main display produced by a third preferred embodiment of a structure design system; Figure 7 is an antenna selection display produced by the structure design system of Figure 6; Figure 8 is a region display produced by the structure design system of Figure 6; and Figure 9 is a terrain display produced by the structure design system of Figure 6.

A first structure design system is used for the design of guyed masts and employs a neural network algorithm to determine a confidence value when different 15 telecommunications antennas are positioned on a guyed mast. The guyed mast structure has in the past posed the most problems for communications structure design experts. A *guyed mast which is 1 metre wide and a 100 metres high has a relatively low tolerance to the addition of telecommunications equipment, such as microwave links and radio transmission antennas, and the structural engineering considerations are complex and 20 involved.

.Pi- P \OPER\DBW\52658.93.12/5/97

I--Y

The guyed mast design system is implemented in software apd can be executed qropblc l on an IBM T compatible PC. The system includes a -phNsIuser interface which produces a visual display as shown in Figure 1 when running under Microsoft" Windows

T

The display includes a representation of the guyed mast 2 with the antennas which have been positioned on the mast. A current antennas window 6 lists the types of antennas which have been added, their height on the mast 2, and their orientation in degrees about the mast 2. An add antenna window 8 allows the user of the system to select an antenna type from a list of antennas in an antenna selection box 10. Using scroll bars 12 and 14 the height and orientation of the selected antenna can be chosen as desired. When the type, height, orientation have been selected as required, the antenna can be added to the mast 2 by selecting an add antenna button 16. The antenna is then automatically displayed on the mast 2. As the antenna is added, the computer program executed by the guyed mast design system uses structural ergineering equations to determine the loads exhibited on the guyed mast which correspond to the attributes 15 required to execute a neural network algorithm employed by the program. The neural network algorithm of the system is used to determine a confidence value concerning the stability and soundness of the mast 2 with all of the chosen antennas, including the additional antenna, and the confidence value is displayed in a confidence window 18.

The confidence value is displayed as a percentage which represents the likely stability of the structure 2 in a selected foundation 20, such as mass concrete, and subject to a selected wind speed 22 and direction. Current antennas can be selected using the current antennas window 6 and deleted using a delete antenna button 24 as desired, and the confidence level displayed in the window 18 is affected automatically.

0 The system is particularly advantageous as it does not require a user to answer a S* laborious set of technical questions in order to produce the confidence level, as is the case with most expert systems. All that is required is for an antenna to be selected and positioned on the mast 2. The use of the neural network algorithm to generate the confidence level reduces complexity, eliminates calculation steps, and enhances the speed of the system. The algorithm is able to produce a relatively accurate confidence level without executing complex structural engineering design steps, and instead relies on a data set of previously constructed structures from which the algorithm is derived.

93122Zp:opcAdbw.PLA53.9Z5 IL~ P I -6- The algorithm used may be represented by a network 30, as shown in Figure 2.

As an initial input, the network 30 includes raw inputs 32 which may be the loads experienced by various segments of the guyed mast 2. From the raw inputs 32 are derived a set of predetermined attributes 34 which are chosen as being important for a determination in respect of whether the structure is sound. For example, the load at a particular segment may be assigned a number of threshold levels and those levels are then associated with respective attributes. If the load on that segment falls above a particular threshold then the corresponding attribute may be set high. The attributes are applied to hidden units 36 where they may be logically compared. The comparison performed in this instance is an AND operation. From the hidden units, the results are summed by at least one adder 38 after respective weights are applied to the outputs submitted to the adder 38. A bias unit 40 also applies a bias value to the adder 38 so as to produce a result which can be applied to a threshold test 42 to give a correct positive or negative result. A number of adders 38 may be employed depending on the number of results or outputs required, in which case the hidden units 36 may each have more than one output and respective weight.

The neural network algorithm used in the guyed mast design system is developed from a set of erected radio mast data to which a' neural network training process is 20 applieu. The original radio mast data used is listed in the accompanying Appendix 1.

The object of the training process is to produce an algorithm or rule to determine whether the mast structure is safe and sound. For example, in the simple case where an erected mast is subject to a load as follows: Load Safe/Unsafe 40 Safe Safe Unsafe Unsafe The rule in this case would simply be that a load greater than 50 is unsafe for the structure. In considering more attributes, the mast may be divided into two segments, top and bottom, and the data set may be as follows: 93lpopci'db.PrU .9Z.6 I M -7- Bottom Load Top Load Safe/Unsafe 20 Safe 10 Safe 10 Unsafe 60 20 Unsafe 15 Unsafe The rule may then be (Bottom Load 2 Top Load 60) ensures a sound structure.

The training process used needs to handle a large number of attributes over an extensive data set and be able to select any of those attributes which are necessary to determine whether the mast structure is safe or unsound.

The list in Appendix 1 was produced by ensuring the data for each mast was complete for its site, wind, height and type fields and then serial numbers where added.

The data was then placed in a form which could be readily accessed by a program written in C that is used to recode the data so as to place it in a form as shown in Appendix 2.

The recoded form is in a format which can be used by a learning algorithm, described hereinafter, and simplifies the representation for guy size configuration and antenna loads.

The numbered fields for each record are initial attributes and are described below in Table 1.

For example, field 5 represents the guyed mast configuration. Only four guy configurations were used in the data set, and approximately 80% were the standard configuration. Therefore the list of guy sizes could be replaced by the label of field to simplify the task comparing similar cases. Listing the antenna loads in the manner shown in Table 1 for attributes 6 to 22 also resolved an additional complexity. It is 'difficult to compare two cases involving different combinations of antennas at different heights when listing antennas on a table one at a time. To resolve this a numerical load is assigned to each antenna, and the mast is divided into a number of segments along its length, and the total load in each segment is summed. For example, [100, 96] represents loads at heights from 96 metres to 100 metres inclusively, while [96, 74] represents loads at heights from 74 metres up to but not including 96 metres. This allows slightly 931222,p-pcdbw.PL6503.92.7 I I i r

~I_

-8different antenna configurations to be considered similarly by the learning program.

Three different segmentations have been used to describe each mast, which allows the learning algorithm to select the most informative. To have realistic learning tasks with the relatively small data set, the data set was put into two separate cases according to foundation type, tfirteen cases of masts with bored pier (BP) and 47 cases with mass concrete (MC) foundations. Furthermore the data set only contains six examples with non-standard guy configuration split into three different types, and since this number was too small to learn anything empirically for such configurations, they were converted to "insufficient standard configuration" cases to simplify the task into simply deciding whether the standard guy configuration was sufficient or not, "OK" being represented by 1 and "Not OK" or non-standard guy configuration being represented by 0 as shown in Table 1.

0.

go o a a 932p:\opczbbwPLS5O3.928 I i O I~IIIIIL Usl -9- TABLE 1. Input attributes for learning algorithm ATTRIBUTE MEANING 1 Serial number of case, starting from 1 2 Status of case (1 OK standard guy configuration, 0 not OK or non-standard guy configuration) 3 Always set to 1 4 Foundation type (0 mass concrete, 1 bored pier) Guy configuration (0 (14, 13, 13, 11, 10), standard 1 (16, 13, 13, 11, 2 (14, 14, 13, 11, 3 (14, 13, 14, 11, 6 Total load (multiplied by 100) in segment of tower [100, 96] 7 Total load (multiplied by 100) in segment of tower [96, 74] 8 Total load (multiplied by 100) in segment of tower [74, 54] 9 Total load (multiplied by 100) in segment of tower [54, 36] Total load (multiplied by 100) in segment of tower [36, 18] 11 Total load (multiplied by 100) in segment of tower [18, 0] 12 Total load (multiplied by 100) in segment of tower [100, 15 13 Total load (multiplied by 100) in segment of tower [85, 64] 14 Total load (multiplied by 100) in segment of tower [64, Total load (multiplied by 100) in segment of tower [45, 27] 16 Total load (multiplied by 100) in segment of tower [27, 0] 17 Total load (multiplied by 100) in segment of tower [96, 94] 20 18 Total load (multiplied by 100) in segment of tower [94, 92] 2 19 Total load (multiplied by 100) in segment of tower [92, S. 20 Total load (multiplied by 100) in segment of tower [90, 88] 21 Total load (multiplied by 100) in segment of tower [88, 86] 22 Total load (multiplied by 100) in segment of tower [86, 84] The load attributes of fields 6 to 22 were quantised according to the thresholds shown in Table 2 below so as to provide a binary vector input for the learning algorithm.

The thresholds were chosen so as to partition uniformly the major values for each attribute.

9312p:\opadbwJ.P3M.92,9 II-- rl I0 10 TABLE 2. Thresholds for quantisation of input loads Field Number Numerical values of thresholds Thresh.

6 13 100 200 500 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 7 13 100 200 500 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 8 4 3000 400000 6000 9 0 2 100 200 11 0 12 13 100 200 500 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 13 13 100 200 500 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 14 0 2 00 200 16 0 17 3 100 200 500 18 3 100 200 500 19 13 100 200 500 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 3 500 4000 4000 21 3 600 4000 6000 22 0 Neural network or classification rules were developed on the basis of the quantised unit loads using the learning algorithm described in detail in A. Kowalczyk and H.L. Ferra, "Developing Higher Order Networks with Empirically Selected Units", IEEE Transactions on Neural Networks, Vol. 5 No. 4, July 1994. The algorithm used is Algorithm 2 discussed on pages 4 and 5 of the paper, where the fields and threshold values are inputted as x, x x, to produce the weights w for selected terms, or hidden units, t. The algorithm is also discussed in A. Kowalczyk, G. Aumann, H.L. Ferra, and J. Cybulski, "Associative Mappings with Positive Bounded Coefficients", T. Kohonen et al., eds., Artificial Neural Networks, Proc. Inter. Conf. on Art. Neural Net. (ICANN-91) Espoo (Finland), 1991, North-Holland, Amsterdam, 1991, and A. Kowalczyk, H. Ferra, and K. Gardiner, "Discovering Production Rules with Higher Order Neural Networks: A Case Study", Machine Learning: Proc. of the Eight International Workshop (ML91), L.A. Birbaum and G.C. Collins, eds., Morgan Kaufman, 1991. The algorithm attempts sc o o a 940524,pApedtubrwPL6503.924O

I

11to fit a polynomial with a relatively small number of terms to the training data set so as to give an accurate numerical result that corresponds to the data set. Accuracy is maintained with a small number of terms by using a number of algebraic heuristics that are the essence of the algorithm. The algorithm was executed on a Sun Microsystems workstation and the training time required to produce the classification rules a few seconds.

The rules produced from the quantised guyed mast data set of Appendix 2 are shown in 'Noles 3, 4 and 5. The tables list the critical attributes that need to be considered, which each comprise a load field and threshold level, and specify the weight to be applied to each attribute. The rules correspond to the network 30 described previously with reference to Figure 2 and may be conveniently viewed as a score table, where the final score is summed at the adder 38 after different weights are attached to different combinations of load segments.

TABLE 3. Rule 1 for classification of MC cases eg o o ,,25 o o eooo oooo e o NO. FIELD THRESHOLD WEIGHT 0 bias 0.8 1 6 500 0.2 2 6 8000 -0.8 3 13 100 -0.6 4 13 5000 0.6 5 6 7 4000 7000 940524,p:\Aope\dbPL553M9Z11 I 12 TABLE 4. Rule 2 for classification of MC cases NO. FIELD THRESHOLD WEIGHT 0 bias 1 6 100 2 6 5000 3 6 7000 4 6 8000 7 4000 6 7 5000 7 7 7000 8 8 4000 9 10 100 12 500 11 1.2 5000 12 12 10000 13 13 5000 14 19 6000 21 600 16 6 6000 17 6 9000 18 12 6000 19 12 7000 a a.

93122p:\opczdw.PL65MAS12 C I -13 TABLE 5. Rule 3 for classification of MC cases NO. FIELD THRESHOLD WEIGHT 0 bias 1 6 100 2 6 5000 3 6 8000 4 7 4000 7 5000 6 7 7000 7 8 4000 8 12 500 9 12 5000 13 5000 11 19 6000 15 12 21 600 For example, with reference to Tablet, the first entry is the bias field 40 and the initial score is 0.8. To continue to evaluate the score for a given case the prescribed fields and thresholds need to be checked. If the value of a listed field, field 6 say, is 20 greater than or equal to the listed threshold, 500, then the attached weight is added to the 3 score, 0.2. In the case of combined fields and thresholds, as for item 5 of Table the results needs to be logically ANDed before the weight 1.0, can be added to the score.

The final score is compared with the threshold 42, which is set at 0.5. If the final score is less than 0.5, the structure is classified as unsafe (not OK), and otherwise the structure 25 is considered sound A percentage confidence value can be obtained from the resultant sum. Considering cases 14 and 25 in the database the loads in each field are as follows:

S

S

1 2345 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 14 1 1 0 05198 4996 0 0 0 0 10194 0 0 0 0 0 0 0 0 0 0 0 1 0 0 12899 6881 0 0 0 0 19235 5450 0 0 0 0 6336 0 0 0

LI

93l222.p: ,AoMbWm.PL6503.9Z13 r 14- Thus for case 14 the score is 0.8 0.2 0.0 0.0 0.0 0.0 1.0 0.5. Hence, the final decision is which agrees with the database. For case 25 we obtain: 0.8 0.2 0.8 0.6 0. 0.0 -0.4 0.5. Thus the final classification is "Not OK", which again agrees with the database.

The algorithm of Table 3 classifies correctly all but three structures of the 47 structures with MC foundations in the original data set. For the alternative rules developed, the rule of Table 4 also only misclassifies three structures, but has the advantage of having integer weights for enhanced processing. The rule of Table misclassifies only one structure for which the final score was approximately 0.5, the case being number 31, Blakely River.

For the thirteen BP foundation structures, the following simple rule proved accurate for all cases: if field 6 is less than 700 then structure is so ad, otherwise it is not sound.

A second structure design system for guyed masts has also been developed which is a more sophisticated version of the first structure design system. The main display generated by the system, as shown in Figure 3, is part of a graphical user interface which allows the user to rotate the view of the mast 2 using rotate view buttons 44 and 46. By "clicking" on one of the buttons 44 and 46 using a pointing device, such as a mouse, the display of the guyed mast 2 can be rotated about its axis, and accordingly its orientation changed. The degree of change is reflected in a compass display 48 which shows the orientation of a grid reference point 50 of the mast 2 relative to true or magnetic north 52. The display of the mast 2 and the antennas 4 which have been selected and positioned on the mast 2 is more graphically accurate than the display of the first design system, and a zooming facility is provided for enlarging the display of a selected area of the guyed mast 2, as shown in Figure 4. This assists in providing accurate or fine positioning of the antennas 4 which, once placed on the mast 2, can be selected and "dragged" to a new position using the pointing device. The procedure for adding an antenna to the mast 2 has also been enhanced in that the procedure involves simply selecting a position by clicking on the mast using the pointing device, which produces 931222,popcdbw.PL603.92,14

I

I an ADD antenna dialogue box 54, as shown in Figure 5. The type of antenna can be selected using the selection box 56 and then the polarisation selected using a polarisation selection box 58 (obscured in Figure 5 by box 54). The orientation and height of the antenna on the mast can also be selected using height and orientation selection boxes and 57, respectively. On adding a new antenna 4, or moving an existing antenna 4, a neural network algorithm of the system determines a confidence value which is displayed in a confidence box In addition to displaying the confidence value, the system displays each guy 62 of the mast 2 in a colour which represents the nature of the tension in the guy 62. For example, a black guy indicates that the tension experienced by the guy is within acceptable limits, a purple guy indicates that the tension is reaching a high level which a designer may not wish to maintain in the guy 62, and a red guy indicates that the tension of the guy 62 will cause it to fail or it needs to be replaced by stronger guy.

The neural network used by the second design system to generate the confidence value and the state of each of the guys 62 has the same structure as that of the network described previously with reference to Figure 2, except the hidden units 36 instead of performing an AND logic operation on their respective inputs, instead perform "fuzzy" AND logic on the inputs. The fuzzy AND operation performed by a hidden unit 36 involves applying a ramp function between two thresholds to each attribute input to the hidden unit. For each attribute input, a threshold and a ramp slope are provided which define a ramp between a minimum threshold corresponding to a bit value of zero and a maximum threshold corresponding to a bit value of one. If the attribute input falls between the minimum and maximum thresholds, then the hidden unit 36 determines a 'value for that attribute, for example 0.2 or 0.8 depending on the slope between the minimum and maximum thresholds. If a hidden unit 36 has more than one input, then the fuzzy AND logic operation selects the minimum of each ramp function value determined for each input and the minimum constitutes the output of the hidden unit 36.

For the second design system, the neural network algorithm needs to generate up to nine different outputs, one relating to the confidence value for the entire mast 2, and eight representing confidence values for each of the guys 62 of the mast 2. A 150 metre mast 9312p:opaO 4bwXPd503.9215 e-I

I

-16has eight guys 62, whereas the 100 metre mast displayed in Figure 3 has five guys 62.

Therefore up to nine adders 38 are used and nine different weights are applied to each output of a hidden unit 36 to generate nine respective outputs for the adders 38. The use of fuzzy logic by the neural network provides more accurate confidence values and improves the transition between confidence values as antennas are moved along the mast 2.

The data set used to train and develop the neural network for a second design system is similar to that provided in Appendix 1 but is somewhat more extensive in that it includes about 100 existing guyed mast constructions from which structural measurements have been taken.

The neural network for the second system is established using the same training algorithm discussed previously which provides strict thresholds for each of the input 15 attributes, and weights for the hidden unit outputs based on the strict thresholds and the assumption that the hidden units are AND gates. Ramp slopes are allocated to each of the thresholds in an ad hoc manner and then the weights of the hidden unit outputs adjusted using singular value decomposition techniques to take into account the now fuzzy AND output of each of the hidden units 36.

S• The attributes considered are described below in Table 6 which are more extensive than those considered in Table 1 for the first design system in that they now include guy loads and loads on sub-segments of the mast segments considered in Table 1.

93122Zp\ pczpbwJL6503.92.16

I

I

17 TABLE 6. Network attribute inputs INPUT DESCRIPTION 1 mast height in [m] 2 mast type 3 foundation code, 0 for bored piers, 1 for massed concrete 4 guy configuration, e.g. 0 for standard configuration 12 approximate guy loading in kN divided by guy safe working load [kN] 13 28 load distribution a.* .oo o o.

o ot a To determine an approximate guy loading the system performs the following procedure. If an antenna 4 is located between two guys 62 then its load is distributed between two guys in proportion to the inverse of its distance from the guy's location.

Then the load is divided by the cosine of the guy angle with the horizontal direction.

Next predetermined initial tensions due to wind acting on the mast only are added and 15 the result is divided by the safe working load of the guy of a given diameter. The result is subtracted from 1, multiplied by 1000 and rounded to the nearest integer.

For the 16 load distribution inputs each of the 4 top segments of the mast, delineated by the attach points 63 of the guys 62, is divided into 4 sub-segments. If an 20 antenna is located in between two points of a division between sub-segments then the load corresponding to it is distributed between these two points, according to the inverse of the distances from these points.

The neural network algorithm used by the second design system is defined by the 25 contents of a data file 100MCBP7.TRM provided in Appendix 3. The files lists the configuration of 62 hidden units of the network from the bias unit 40 and the first hidden unit to the 61s e hidden units up to number 41 have one input, units 42 to 60 have two attribute inputs and unit 61 has three attribute inputs, which denotes the order of the hidden units as indicated next to the hidden unit number in the file. The record for each hidden unit indicates which attribute input is to be sent thereto, the threshold and slope which define the ramp function for the unit and the six weight values to be applied to the output of the hidden unit. For example, for the first hidden unit, 93122Zp-\op=dbsw.PL6O3.ZI7 i -I

YP~Y__

18attribute input number 4 is to be applied, the threshold is 125 and the slope is 20. The six weight values for a 100 metre mast 2 are listed one after the other underneath the slope value. For the 61st hidden unit, attribute inputs 4, 5 and 6 are to be applied thereto, the thresholds are 60, 0 and 0, respectively, and the slopes are 20, 20 and respectively.

A third structure design system is used for designing antenna poles 70 which are up to 30 metres high, as shown in Figure 6. The pole 70 includes a number of pole sections 72 which form the pole 70, and a head frame 74 mounted at the top of the pole 70. The pole system, similar to the guyed mast systems described previously, allows antennas to be selectively mounted on the head frame 74 or the pole sections 72 by using a pointing device to "click" on a desired position of the head frame 74 or the pole sections 72 in the displays provided. The main display of the system, as shown in Figure 6, provides a side view 76 of the pole, a plan view 78 of the head frame and a side view 15 80 of the head frame 74, all of which can be rotated using the rotate view buttons 44 and 46 about the axis of the pole, and then the orientation displayed in a compass box 48, as discussed previously for the second design system. A zoom view can also be selected for a part of the pole '70 as for the second design system.

Selecting an antenna position causes the system to produce an antenna selection •dialogue box 82, as shown in Figure 7, where the type of antenna technology to be employed can be selected in a technology box 86, which provides a choice of antenna types from which a selection can be made in a scroll box 84. A plurality of the selected antenna can also be selected to be mounted on all of the sectors corresponding to a selected position on the head frame 74. As each antenna 88 is selectively mounted on ithe pole 70, the system generates a confidence value which is displayed in a confidence box 60 indicativLe of the confidence in the stability of the pole structure. Pole sections 72 which are determined to be subject to failure are displayed in a red colour, as shown for the bottom pole section 72 in Figure 6. A neural network is not required to generate confidence values for the pole 70 or the pole sections 72 as structurally the pole is a linear cantilever which is only subject to linear stresses, Therefore the structural calculations to determine the stresses imparted on the pole 70 are relatively simple, such 93lp:\opctdbwPL6503.92,l I i; e II -M 19that the confidence values can be provided rapidly in the interactive environment of the structure design system using standard structure engineering calculations. This, however, is not the case with the guyed mast structures, as discussed previously, which involve consideration of complicated non-linear stresses.

Various characteristics of the pole 70 can be varied using selections available in an edit menu and include pole type, head frame type, foundation, and region and terrain in which the pole 70 is mounted. Selecting the region option causes a map display as shown in Figure 8, to be generated which shows in different colours zones classified according to average wind speed. For example, red may indicate a severe tropical cyclone zone, purple may indicate a tropical cyclone zone, yellow may indicate an intermediate wind speed strength zone and green may indicate a normal wind strength zone. The type of region in which the pole 70 is likely to be mounted can then be selected by clicking on a corresponding position in the display 90. The type of zone in which the pole 70 is located is then taken into account when determining the confidence values. Similarly, selecting the terrain menu option causes generation of a terrain display 92, as shown in Figure 9, from which different types of terrain can be selected and then S* taken into account in determination of the confidence values.

The structure design systems can also generate displays of the radiation patterns produced by selected antennas when mounted on a guyed mast 2 or a pole 70 to determine radiation levels and any disadvantageous effects they may have. This feature is particularly advantageous for use in maintenance of existing masts 2 and poles 70, as well as for new installations. Once the antenna configurations for a mast 2 or pole has been finally determined, the design systems can then be used to produce detailed specifications for the construction and installation of the mast 2 or pole The structure design systems discussed above provide an interactive environment where parties without any expertise in structural engineering can graphically select antennas for a guyed mast 2 or pole 70, arrange the antennas as desired, and select characteristics, such as mounting region and terrain, which may effect the structure. The systems then provide the designer with an immediate indication as to whether the chosen 931222,p:ope\dbw ,P1 03.92,19 ill ~Y L 20 structure configuration is satisfactorily stable, by displaying confidence values for the structures. Various detailed views of the structure are also available which facilitates accurate placement of antennas. Although developed specifically for antenna mounting structures, the systems can be applied to any other structures to which loads may be added.

Throughout this specification and the claims which follow, unless the context requires otherwise, the word "comprise", or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated integer or group of integers but not the exclusion of any other integer or group of integers.

:*0 :0.

0* 00 93122Zp-ropuWbw.PL6503.92,20 r 21 APPENDIX I

I

1% I jiia Ra.jctMast Iaih 0 0 *0 ~0 0 0* *0 0 Site Wind Type HeightS/M Guy sizes Fund Stat. Ant-type 1 Santos 2 Ross 3 Ross 4 Ross 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D M 14 13 13 11 10 MC OK 14 13 13 11 10 MC OK 14 13 13 11 10 MC No 2 KP213-501 DU6 3 1500MHZOMNI KP6-15 NP 6-15 5 1500MHZOMNI 1(26-15 1(26-15 KP210-17 KPIO0-17 16 13 13 11 FourUMT 6 FourM-T 7 ThurlooD 8 ThurlooD 9 Santos StAnn 11 Sorrento 12 Warbrec.

100 45D S 100 45D S 100 45D S 100 45D S 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 '14 13 13 11 14 13 13 11 14 13 13 11 10 MC OK 5 1500MHZ_-OMNI 1(26-15 1(26-15 KP210-17 KPI10-17 10 mC OK 2 ATW8L1HTO KP213-50 1 10 MC OK 3 RFSSU1-8C 1(213-501

OMNI

10 mc OK 2 ATWBLI-TO KP213-50 1 10 MC OK 3 ATW8L1HTO KP213-50 1

OMNI

10 MC OK 2 1(213-501 DU6 10 MC ON 3 KP213-501 DU 6 1(213-501 10 MC ON 3 KP213-501 DUI6 NP 13-501 10 BP No 3 124528-2 GKA3 8 GKA38 Ant._boc Orient.

100 73 100 0 100 0 99 160 99 100 0 99 160 99 96 160 96 100 0 99 160 99 96 160 96 100 270 98 42 100 270 98 42 9 C 225 100 270 98 117 100 270 98 117 90 225 100 73 100 0 100 100 0 80 336 100 156 100 0 100 86 100 0 98 101 98 288 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 13 Warbrec. 45 14 Warbrec. 45 100 45D M 100 45D S 16 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 Nejouis 16 Haughton 17 Sun-Hill 45 100 45D S 45 100 45D S 45 100 45D S 18 Diamant. 45 19 Diarnant. 45 10 BP OK 3 124528-2 GKA3 8 GKA38 10 BP OK 3 124528-2 GKA38 GKA38 10 BP OK 2 DU6 GKA38 10 BP OK 2 DU6 GKA38 10 mc OK 4 GKA38-15 DU6 YU15-4 YU15-2 10 mc OK 1 KP13-9 10 MG No 4 KP13-9 YV11-2 YV11-2 YV9 10 MC OK 4 KP13-9 YV11-2 YV11-2 YV9 10 mc OK 4 YV11-2 KP13-9 YV11-2 YV9 10 BP OK 2 KP13-9 100 0 98 101 98 288 100 0 98 101 88 288 100 0 98 108 100 0 98 100 83 100 0 100 353 100 216 90 43 90 43 88 178 77 178 30 179 90 43 88 178 77 178 30 179 92 178 90 43 88 17 8 30 179 100 223 100 45D S 100 45D S 131 131 11 131 131 11 Diamant. 45 100 45D M 141 141 13 11 21 Diamant. 45 100 45D S 141 131 13 11 22 Urbana 23 Urbana 24 Adavale Balootha 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 KP13-9 90 27 10 BP OK 3 YV9 100 156 KP13-9 100 223 KP13-9 90 27 10 mc OK 3 1500 MHZ OMNI 100 0 KP13-15 98 134 KP13-15 90 357 10 mc No 5 KP13-15 100 161 KP13-15 100 284 KP13-15 90 348 11dBi-Omni 100 0

S

S* S S 26 Balootha 27 Breada.

28 Ballynure 29 Ballynure Ballynure 31 BuckleyR 32 Bulgroo 33 Goombie 34 CrystalB 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 16 13 13 11 YV11-2 80 53 10 MC OK 5 KP10-15 100 161 KP10-15 100 284 KP13-15 84 348 lldBi Omni 100 0 YV11-2 80 53 10 MC OK 3 KP10-15 90 8.4 GKA38-15 100 195.2 124528-2 Omni 100 0 10 MC OK 3 KP10-15 100 241 GKA38-1415N 90 345 124528-2 Omni 100 0 10 MC No 3 KP10-15 100 241 GKA38-1415 80 345 124528-2 Omni 100 0 10 MC OK 6 KP10-15 100 241 124528-2_Omni 100 0 GKA38-1415N 90 345 90 0 85 0 YV11-2 85 283 10 MC OK 3 124528-2 100 0 1 GKA38-1415N 90 343.6

M

GKA38-1415N 100 197.6 10 MC OK 2 DU6 Omni 100 0 GKA38 17-19 100 8 10 MC OK 2 DU6 Omni 100 0 GKA38 17-19 100 298 10 MC OK 3 KP10-501 100 88 124528-2_Omni 100 0 YV9 100 0 10 MC OK 4 KP10-501 100 88 124528-2_Omni 100 0 YV11-4 98 246.2 YU15-2 95 197.5 10 MC OK 3 1.5GHZ_Omni 100 0 KP10-15 98 34 KP13-15 70 177 10 BP OK 3 KP13-15 100 74 CrystalB 36 Cobbrum 100 45D S 45 100 45D S 37 Eyerah 45 100 45D M 00 0 *C S. 0 0~ 38 Eyerah 39 Eyerah Gilmore 41 Gilberton 42 Gilberton 43 Headingly 44 Headingly 'Headingly 46 Headingly 47 HooleyLag 48 HooleyLag 49 HooleyLag HooleyLag 51 HooleyLag 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D M 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 14 11 14 13 13 11 14 13 13 11 14 13 13 11 14, 13 13 11 14 13 13 11 14 13 13 11 KP13-501 100 223 8dBi Omni 100 0 10 BP No 3 KP13-15 100 74 KP13-501 100 223 8dBi Omni 100 0 10 BP OK 3 KP13-15 100 74 KP13-501 90 223 8dBi Omni 100 0 10 MC OK 2 KP13-15 98 177 DU6 Omni 100 0 10 MC OK 2 DU6 Omni 100 0 KP10-15 90 61 10 MC No 2 DU6 Omni 100 0 KP10-15 80 61 10 MC OK 3 124528-1 100 0 GKA38-1415N 85 338.8 GKA38-1415N 100 202.7 10 MC No 3 124528-1_Omni 100 0 GYA38-1415N 90 338.8 GKA38-1415N 100 202.7 10 MC OK 3 124528-1 Omni 100 0 GKA38-1415N 70 338.8 GKA38-1415N 10C 202.7 L 10 MC No 3 124528-1_Omni 100 0 GKA38-1415N 70 338.8 GKA38-1415N 100 202.7 10 MC No 2 124528-2 Omni 100 0 KP10-15 65 118 10 MC OK 2 124528-2 Omni 100 0 KP10-15 100 118 10 MC OK 2 124528-2 Omni 100 0 KP10-15 100 118 10 BP OK 5 124528-2 Omni 100 0 KP10-15 100 118 90 217 YU15-2 92 356 YU15-2 94 186 10 MC OK 5 124528-2 Omni 100 0 KP10-15 100 118 1 0 *0 0 4 52 HooleyLag 45 100 45D S 53 Macksland 54 Lyndhurst Moontah 100 45D S 100 45D S 45 100 45D M 56 Moontah 57 Moontah 58 Petford 59 Petford Springy.

45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 45 100 45D S 14 13 13 11 14 13 13 11 14 13 13 11 16 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 14 13 13 11 10 BP OK YU15-2 94 186 YU15-2 92 356 YU15 90 217 10 MC OK 4 GKA38-15 90 278.3 124528-2_0mni 100 0 KP10-15 100 65.5 KP10-15 90 156.2 10 BP OK 2 DV6_Omni 100 0 KP10-15 100 0 10 MC OK 4 124528-2 Omni 100 0 GKA38-1415N 100 94.9 GKA38-1415N 90 22.7 KP10-15 90 197.1 10 MC OK 4 124528-2 0mni 100 0 GKA38-1415N 100 94.9 SGKA38-1415N 87 22.7 KP10-15 65 197.1 10 MC No 4 124528-2_0mni 100 0 GKA38-1415N 100 94.9 GKA38-1415N 90 22.7 KP10-15 90 197.1 10 MC OK 5 KP10-501 100 257 124528-2_0mni 100 0 YV11 100 316 YV11 100 144 100 140 10 MC OK 3 KP10-501 100 257 124528-2_0mni 100 0 YV11-2 94 140.2 10 MC OK 2 GKA38-15 100 336.1 DU6 Omni 100 0 90 217 YUI5-2 92 356 YU15-2 94 186 6 124528-2_0mni 100 0 KP10-15 100 118 98 211 27 APPENDIX 2

S

S S

S.

S

SS S

SS

S.

S. S S S S S 55 Entf'e l Radid ast: a-t: for LVar.iI?'* Alg .ithm 7 8 0 i 2. 13 14 15 6* *2 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 12345 3763 3545 10927 8895 4263 3954 4263 4263 3763 3763 7299 10194 10194 5198 5223 5223 5823 0 0 0 0 4327 4445 6563 12899 7609 5198 3893 3893 3893 5198 5223 5223 2534 3507 3918 10099 0 0 0 0 0 227 0 227 0 3536 0 0 0 4996 0 0 0 4327 5417 5417 5417 4327 4327 6336 6881 6881 3691 4996 4996 5741 4996 0 0 0 200 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6336 0 3763 3545 10927 8895 4263 4181 4263 4490 3763 3763 7299 10194 10194 10194 5223 5223 5823 4327 4872 4872 5417 8654 8772 12899 19235 7609 8889 8889 3893 9634 10194 5223 5223 2534 3707 3918 10099 0 0 0 0 0 0 0 0 0 3536 0 0 0 0 0 0 0 0 545 545 0 0 0 0 545 6881 0 0 4996 0 0 0 0 0 0 6336 0 0 0 0 0 0 227 0 227 0 0 0 0 0 0 0 0 0 4327 4327 4327 4327 4327 4327 6336 6336 0 3691 4996 0 5096 4996 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4996 0 0 0 0 545 545 545 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6336 0 0 0 645 0 0 0 0 0 0 0 12 34 5 6 10099 6563 6563 227 227 5113 5113 5113 5113 202 3893 3893 3893 3893 3993 3893 4191 5198 5198 5198 3062 2416 5223 7 8 0 3536 0 3691 3691 4996 4996 0 0 0 0 0 500 500 500 8687 0 8687 4996 8687 0 545 0 0 0 0 0 0 0 0 4996 4996 3691 0 0 0 0 0 0 0 0 3691 0 0 0 0 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 12 10099 10099 6563 3918 227 10109 10109 5113 5113 202 3893 3893 4393 4393 4493 12580 4191 13885 10194 13885 3062 2961 5223 V 13 0 0 0 0 3691 0 0 4996 4996 3691 0 0 0 0 0 0 0 0 3691 0 0 0 0 14 15 16 17 18 19 20 21 22 0 0 0 0 0 0 0 0 0 0 0 0 200 200 200 0 0 0 0 0 0 545 0 0 3536 0 3691 0 0 4996 0 0 0 0 0 100 100 100 8687 0 8687 0 86 B7 0 0 0 0 0 0 0 0 4996 0 0 0 0 0 0 0 0 0 0 0 0 4996 0 0 0 0 ss(lL~I~ r~ 30 o BAPPENDIX 3

S

S. S

S*

S

S*

S. S

S

S. S

SS

*s.

S. S S S S rI 31 100MCBP7.TRM Actual Network "NN.FileName" "NN. NumberOfTerms" "NN.NumberOfInputs" "NN. NumberOfOutputs" "NN. MaxOrder" "NN. FuzzyThreshods" 0 0 -1 "NNI.trm" 62 28 9 4

YY.-.?I

0 0 0 0 0.000000 0.000000 0.000000 0.000000 0.902535 0.980388 0.966018 0.992356 1.001694 1.000476 0.000000 0.000 000 0.000000 3. S. -1 -1 -1 05 PO§558 .000000 0.000000 -1 -1 -1 0.000000 0.000000 0.000000 -0.®g1836 04 9676 -cVD1484 -04090440 0.%0Q562 0.000000 0 2 1 4 -1 -1 -1 0 -1 -1 -1 20.000000 0.000000 0.000000 0.000000 -0.146467 -0.136228 -0.003178 -0.001482 -0.000023 0.000000 0.000000 -0.001665 0.000000 L 4 20.000000 -0.138912 0.000000 -1 -1 -1 -1 -1 -1 0.000000 0.000000 0.000000 -0.157392 0.007820 0.006198 0.000980 0.001643 0.000000 0.0 a a. at a a a. a. a a a a.

a. a a a.

a a 00000 4 1 4 -1 -40 -1 -1 -1 20.000000 0.000000 0.000000 0.000000 -0.257310 -0.244441 -0.022545 -0.007469 -0.001612 -0.001932 0.000000 0.000000 0.000000 5 1 000000 6 1 000000 4 -60 20.000000 -0.287258 0.000000 4 -100 20.000000 -0.114830 0.000000 -1 -1 -1 -1 -1 -1 0.000000 0.000000 0.000000 -0.289500 0.007141 0.005834 -0.001299 0.001333 0.000000 0.

-1 -1 -1 0.000000 0.000000 0.000000 -0.155980 -0.006769 0.001654 0.000056 0.000016 0.000000 0.

7 1 5 290 20.000000 -0.024665 000000 0.000000 8 1 0 20.000000 -0.421713 .000000 0.000000 9 1 -1 -1 -1 -1 -1 -1 0.000000 0.000000 0.000000 0.000215 -0.019532 0.033441 0.004715 -0.001793 0.000000 0.

-1 -1 -1 0.000000 0.000000 0.000000 -0.025094 -0.231893 0.010283 -0.001950 0.004421 0.000000 0 -1 -1 32 "NN. FileName" "NN. NumberOfTerms" "NN. NumberOf Inputs" "NN. NumberOf Output s" "NN. MaxOrder" "NN. FuzzyThreshods" 0 0 -1 -1 0 0 0.000000 0.000000 0.902535 0.980388 000 0.000000 NN1.trm" 62 28 n=I 9 m 4 1I 0 0 0.000000 0.000000 0.966018 0.992356 1.001694 1.000476 0.000000 0.000 ii 4 -1 -1 -I 125 -1 -I -1 20.000000 0.000000 0.000000 0.000000 -0.005558 -0.001836 0.003676 -0.001484 -0.000440 0.000562 0.000000 0 .000000 0.000000 2 1 4 -1 -1 -1 0 -1 -1 -1 20.000000 0.000000 0.000000 0.000000 -0.146467 -0.136228 -0.003178 -0.001482 -0.000023 -0.001665 0.000000 0.000000 0.000000 20.000000 -0.138912 0.000000 00000 a a.

a *Oaa a. a a a 9* a.

a a a. a.

a a a a. a a. a a a a.

a.

a a. 0* .a a. a a 4 1 4 -40 20.000000 -0.257310 0.000000 0.000000 -1 -1 -1 0.000000 0.000000 0.000000 -0.157392 0.007820 0.006198 0.000980 0.001643 0.000000 0.0 -1 -1 -1 1 -1 -1 0.000000 0.000000 0.000000 -0.244441 -0.022545 -0.007469 -0.001612 -0.001932 0.000000 -1 -1 -1 -1 -1 -1 0.000000 0.000000 0.000000 -0.289500 0.007141 0.005834 -0.001299 0.001333 0.000000 0.

5 1 000000 6 1 000000 4 -60 20.000000 -0.287258 0.000000 4 -100 20.000000 -0.114830 0.000000 -1 -1 -1 0.000000 0.000000 0.000000 -0.155980 -0.006769 0.001654 0.000056 0.000016 -1 -1 -1 -1 -1 -1 0.000000 0.000000 0.000000 0.000215 -0.019532 0.033441 0.004715 -0.001793 7 1' 5 290 20.000000 -0.024665 000000 0.000000 8 1 5 0 20.000000 -0.421713 .000000 0.000000 9 1 0.000000 0.

0.000000 0.

-1 -1 -i -1 -I -1 0.000000 0.000000 0.000000 -0.025094 -0.231893 0.010283 -0.001950 0.004421 0.000000 0

I

33 20.000000 0.000000 0.000000 0.000000 -0.281627 0.008580 -0.306463 -0.019457 0.000738 -0.000978 0.000000 0 .000000 0.000000 1 20.000000 -0.025176 0.000000 0.000000 11 1 5 20.000000 -0.046652 0.000000 0.000000 12 1 5 20.000000 0. 123392 000000 0.000000 13 1 -100 20.000000 0.026220 0 00000 0.000000 0.000000 0.000000 0.000000 -0.000534 -0.117457 -0.015376 -0.004619 -0.002410 0.000000 -1 -1 -1 -1 -1 -1 0.000000 0.000000 0.000000 0.011638 -0.044566 -0.023122 -0.001129 -0.000575 0.000000 -1 -1 -1 -1 -1 -1 0.000000 0.000000 0.000000 0.014915 -0.270185 0.03PEU27 0.003630 -0.000847 0.000000 0.

0.000000 0.000000 0.000000 .025234 0.013640 -0.005269 0.001814 -0.001312 0.000000 0.0 14 1 19 0000 -0.0307 0.000000 0.0000 1 20.0000 5832 0.000000 0.0000 6 -1 -1 '0 -1 -1 -1 00 0.000000 0.000000 0.000000 '57 -0.001977 -0.013387 -0.005810 -0.018397 0.014515 0.000000 00 6 -1 -1 -1 0 -1 -1 -1 00 0.000000 0.000000 0.000000 95 -0.001799 -0.012516 -0.266764 -0.010834 0.004730 0.000000 '00 16 1 6 -20 20.000000 -0.082543 0.000000 -1 -1 -1 0.000000 0.000000 0.000000 -0.003670 0.013494 -0.194041 0.006574 0.001717 0.000000 0.

000000

C.

C

C

C

C.

C

CC

C

17 1 6 20 20.000000 094965 0.000000 0.000000 18 1 6 -40 000000 031900 000000 0.000000 19 1 6 60 20 .000000 025164 .000000 0.000000 -1 -1 -1 0.000000 0.000000 0.000000 -0.000090 -0.033696 -0.170521 -0.016672 -0.001685 0.000000 -1 -1 -1 -1 -1 -1 0.000000 0.000000 0.000000 0.002537 0.005560 -0.187912 0.001805 -0.001724 0.000000 0.

-1 -1 -1 -1 -1 -1 0.000000 0.000000 0.000000 -0.003219 0.021411 -0.048921 -0.004951 0.003147 0.000000 0 -1 -1 m 1 6 20.000000 0.000000 -0.037331 0.002219 000000 0.000000 21 1 7 210 -1 20.000000 0.000000 0.017237 -0.006130 00000 0.000000 22 1 7 -1 0.000000 0.023632 0.00c000 0.00354 8 -34 -1 -1 0. 000 000 0.(X54927 -0.031237 0.029910 0.000000 0.0 20.0000C 6397 0.000000 0.0000 23 1 0 1 100 0.000000 40 0.040927 100 7 -1 0 -1 00 0.000000 0.000000 0.000000 -0.017325 -0.001212 -0.396810 -0.016261 0.000000 -2 20.0000 1038 0. 00000 1 -1I 0.000000 0.000000 87 0 -0.043335 0.014254 0.021837 -0.135992 -0.021819 0.000000 0 0000( 24 1 00000 1 0 7 20.000000 052109 0.000000 7 20.000000 004298 P0 0.000000 7 20.000000 039959 0 0.000000 0.000000 0.000003 0.000000 0.014250 0.000000 0.000000 0.004451 0.000112 -0.412555 0.017759 0.000000 0.0 0.000000 0.000000 -0.013954 -0.003283 -0.047406 0.012215 0.000000 0 C C .00000 26 1 .00000 27 1 0.000000 0.000000 0.000000 -0.023298 0.003152 -0.014570 0.003040 -0.008669 0.000000 0 7 1- 1- 20.000000 0.000000 0.000000 0.000000 1 0.027394 0.021187 0.013184 0.005909 -0.002567 -0.004025 0.000000 0.0 00000 0.000000 28 1 7 -1 -1 -1 -100 -1 -1 -1 20.000000 0.000000 0.000000 0.000000 0.000726 0.006855 0.000861 0.019564 -0.002274 0.004486 0.000000 0.00 0000 0.000000 29 1 8 -1 -1 -1 240 -1 -1 -1 20.000000 0.000000 0.000000 0.000000 -0.195520 -0.031194 0.004967 0.004384 -0.035082 -0.020882 0.000000 0 .000000 0.000000 1 8 -1 -1 -1 0 -1 -1 -1 20.000000 0.000000 0.000000 0.000000 iffim 35 -0.058815 .000000 0.000000 31 1 8 -0.013180 0.004158 -0.001897 0.006446 -0.201051 0.000000 0 -1 -1 -1 -1 0.000000 0.000000 0.000000 0.037843 0.004253 0.003218 -0.039887 -0.237175 0.000000 0.

20.000000 -0.345100 0.000000 000000 32 1 8 1 20.000000 -0.241067 0.000000 0.000000 33 1 17 207 20.000000 -0.045604 000000 0.000000 1 1 1 0.000000 0.000000 0.000000 -0.009819 -0.006110 -0.005240 0.067835 -0.543042 0.000000 0.000000 0.000000 0.000000 0.181476 -0.229089 0.017633 0.000835 -0.001100 0.000000 0.

-1 -1 -1 34 000 0000 36 1 4 -153 -1 -1 -1 20.000000 0.000000 0.000000 0.000000 0.027281 0.006783 0.016089 0.000276 0.001233 0.000241 0.000000 0.000 0.000000 '00 :1 72 -1 20.000000 0.000000 0.079724 -0.000297 0. 000000 0. 000000 0.181472 0.000000 -0.120276 0.016444 -0.017421 0.000000 0.

-1 4 -1

C

C.

C

-292 1 1 1 20.000000 0.000000 0.000000 0.000000 0.020588 0.011305 -0.006482 0.006577 -0.002496 0.002101 0.000000 0.0 00000 0.000000 37 1 13 -1 -1 -1 49 -1 -1 -1 20.000000 0.000000 0.000000 0.000000 -0.058726 -0.064459 0.013944 -0.006254 0.001328 -0.001642 0.000000 0 .000000 0.000000 38 1 21 -1 -1 -1 31 -1 -1 -1 20.000000 0.000000 0.000000 0.000000 0.054844 -0.006796 0.010438 0.004292 0.003425 -0.003027 0.000000 0.0 C 00000 0.000000 39 1 26 -1

CC

S C 64 -1 20.000000 0.000000 0.028129 -0.000128 0.000000 4 -1 0.000000 0.000000 -0.001435 0.008569 -0.040980 0.021211 0.000000 0.

000000 1 222 -1 -1 -1 20.000000 0.000000 0.000000 0.000000 0.036950 0.010057 0.029664 0.000241 0.000756 -0.001726 0.000000 0.00 0000 0.000000 41 1 1118 -1 -1 -1 36 22 20.000000 -0.068718 .000000 0.000000 42 2 4 1 1 1 0.000000 0.000000 0.000000 -0.023574 -0.120596 0.054417 -0.005021 0.005822 0.000000 0 5 -1 -1 0 -1 -1 20.000000 0.000000 0.000000 -0.095622 0.050282 0.006068 -0.000455 -0.001230 0.000000 0 125 000000 -0.019673 0.000000 .0000 43 00000 44 00000I 000 00( 46 00 2 5 -1I 0 0 -1I 1 20.000000 20.000000 0.000000 0.000000 0.2G9888 -0.019504 0.033387 0.000295 -0.002955 0.001744 0.000000 0.0 0. 000000 2 4 5 -1 -1 0 -20 -1 -1 20.000000 20.000000 0.000000 0.000000 -0.222127 -0.310780 0.015010 -0.000330 0.001850 0.000510 0.000000 0.

0 0.000000 2 4 7 -1 -1 0 -20 -1 -1 20.000000 20.000000 0.000000 0.000000 0.603040 0.011419 -0.012069 -0.036309 0.009127 -0.000481 0.000000 0.

0 0.000000 2 4 5 -1 -1 0 -1 -1 20.000000 20.000000 0.000000 0.000000 0.066214 0.016655 -0.036876 0.008730 0.002273 -0.000114 0.000000 0.0 0.000000 00000 0* 47 2 4 5 -1 -1I 40 0 -1 1 20.000000 20.000000 0.000000 0.000000 -0.,212462 -0.125190 -0.153539 -0.020741 -0.005842 0.001536 0.000000 0.000000 0.000000 48 2 4 5 -1 -1 -20 -1 -1 20.000000 20.000000 0.000000 0.000000 0.463497 0.214961 0.194325 0.018562 -0.001916 0.001650 0.000000 0.00 0000 0.000000 49 2 000000 2 4 5 -1 -1 0 -1 -1 20.000000-20.000000 0.000000 0.000000 0.292789 0.209601 -0.016262 -0.009676 0.007821 -0.001557 0.000000 0.

0.000000 4 6 -1 -1 60 0 -1 -1 20.000000 20.000000 0.000000 0.000000 0.566254 -0.000661 -0.033536 -0.037984 -0.002780 0.004360 0.000000 0 .000000 0.000000 51 2 4 18 -1 -1 -100 22 -1 -1 20.000000 20.000000 0.000000 0.000000 0.154345 0.133059 0.110050 -0.018144 0.003491 -0.005471 0.000000 0.0 di 37 00000 0.000000 52 2 5 6 -1- 0 190 -1- 20.000000 20.000000 0.000000 0.000000 0.084166 0.005838 0.087854 -0.036686 -0.003331 -0.053072 0.000000 0.

0.000000 5 6 -1 -1 00000( 53 2 00000 54 2 0 0 1 -3.

20.000000 20.000000 0.000000 0.000000 0.438476 0.022064 -0.298321 0.009238 -0.007736 0.010899 0.000000 0.0 0. 000000 7 -1I 0 0 -1I 1 20.000000 2.0.000000 0.000000 0.000000 0.531164 -0.012162 0.001975 0.041324 0.046132 0.000338 0.000000 0.00 0000 0.000000 2 5 8 -1 -1 0 240 -1 -1 20.000000 20.000000 0.000000 0.000000 0.420257 0.023717 0.038876 -0.001744 -0.000929 -0.007202 0.000000 0.

000000 0.000000 56 2 5 13 -1 -1 0 49 -1 -1 20.000000 20.000000 0.000000 0.000000 0.190897 -0.036611 0.315546 -0.021311 -0.000081 -0.001931 0.000000 0 .000000 0.000000 57 2 5 6 -1 -1 0 -1 -1 20.000000 20.000000 0.000000 0.000000 0.210740 -0.021293 0.289143 0.064554 0.011696 -0.022475 0.000000 0.0 00000 0.000000 58 2 5 8 -1 -1

S

5 *5 0* 0* S S

S

-40 20 -1 -1 20.000000 20.000000 0.000000 0.000000 0.361750 -0.007866 -0.072834 -0.007084 0.005051 0.007792 0" 00000 0.

0.000000 000000 59 2 0 0 -1 -1 20.000000 20.000000 0.000000 0.000000 0.598403 -0.002589 0.015726 0.059572 0.035527 0.013882 0.000000 0.00 0000 0.000000 2 7 8 -1 -1 0 20 -1 -1 20.000000 20.000000 0.000000 0.000000 0.511323 -0.013453 0.006293 -0.009084 0.070978 0.0217'.1 0.000000 0.0 00000 0.000000 61 3 4 5 6 -1 0 0 -1 20.000000 20.000000 20.000000 0.000000 -0.497975 -0.006015 0.106628 0.035354 0.012729 -0.009044 0.000000 0.

000000 0.000000

I

Claims (39)

1. A structure design method using a computer graphics processing system, said method comprising: generating at least one view of a structure; graphically selecting a position on said view for an object; displaying said object on said structure at said posit'on; determining load parameters for said structure with said object in said position, where said load parameters represent loads on selected parts of said structure; and determining on the basis of said load parameters, and displaying, at least one indicator of the soundness of said structure.

2. A structure design method as claimed in claim 1, including graphically selecting said object from a plurality of predetermined objects.

3. A structure design method as claimed in claim 1 or 2, including selecting characteristics of said structure which affect determination of said load parameters.

4. A structure design method as claimed in claim 3, wherein said characteristics 20 include mounting surface, foundation and wind speed at location.

A structure design method as claimed in any one of the preceding claims, including rotating said view. 25

6. A structure design method as claimed in any one of the preceding claims, including selecting part of said view and generating an enlarged display of said part.

7. A structure design method as claimed in any one of the preceding claims, including automatically generating a specification for the construction of said structure.

8. A structure design method as claimed in any one of the preceding claims, wherein said object is an antenna. P:\OPER\D\W\52658.93 12/5197 -39-

9. A structure design method as claimed in claim 8, including generating a radiation pattern display for said antennae on said structure.

A structure design method as claimed in any one of the preceding claims, wherein determining said at least one indicator includes executing a neural network algorithm on the basis of said load parameters which comprise load attributes.

11. A structure design method as claimed in claim 10, wherein said algorithm is derived from a neural network training process performed on a data set for constructed structures.

12. A structure design method as claimed in claim 10 or 11, wherein executing said algorithm includes: applying a predetermined threshold to each of said load attributes to generate 15 threshold attributes; performing a logic operation on selected ones of said threshold attributes to generate logic outputs; applying predetermined weights to said logic outputs to generate weighted outputs; and 20 generating said at least one indicator on the basis of selected combinations of said weighted outputs. V

13. A structure design method as claimed in claim 12, wherein said logic operation is an AND operation.

14. A structure design method as claimed in claim 10 or 11, wherein executing said algorithm includes performing a fuzzy logic operation. A structure design method as claimed in claim 12, wherein said logic operation is a fuzzy AND operation and a predetermined slope is applied to each of said load 97 -1 attributes.

P:\OPERDB\V52658.93 l"5/97 I III I

16. A structure design method, comprising: selecting a position on a structure for an object; determining load attributes for said structure with said object in said position, where said load attributes represent loads on selected segments of said structure; and executing a r':ial n,.;rwork algorithm on the basis of said attributes to obtain at least one indicator of the soundness of said structure.

17. A structure design method as claimed in claim 16, including selecting said object from a plurality of predetermined objects.

18. A structure design method as claimed in claim 16 or 17, wherein said algorithm is derived from a neural network training process performed on a data set for evaluated structures. 99 15

19. A structure design method as claimed in claim 16, 17 or 18, wherein said posiaon *selecting step is performed using a graphical user interface, which, on completion of said selecting step, displays said at least one indicator using said interface.

A structure design system comprising: 20 means for generating at least one view of a structure; means for graphically selecting a position on said ,iew for ai object; means for displaying said object on said structure at said position; means for determining load parameters for said structure with said object in said position, where said load parameters represent loads on selected parts of said structure; 25 and *means for determining on the basis of said load parameters, and displaying, at least one indicator of the soundness of said structure.

21. A structure design system as claimed in claim 20, including means for graphically selecting said object from a plurality of predetermined objects. SP.\OPER\DBWU2658.93 12/5I97 I -41-

22. A structure design system as claimed in claim 20 or 21, including means for selecting characteristics of said structure which affect determination of said load parameters.

23. A structure design system as claimed in claim 22, wherein said characteristics include mounting surface, foundation and wind speed at location.

24. A structure design system as claimed in any one of claims 20 to 23, including means for rotating said view.

A structure design system as claimed in any one of claims 20 to 24, including means for selecting part of said view and generating an enlarged display of said part.

.26. A structure design system as claimed in any one of claims 20 to 25, including 15 means for automatically generating a specification for the construction of said structure. 0

27. A structure design system as claimed in any one of claims 20 to 26, wherein said object is an antenna. 20

28. A structure design system as claimed in claim 27, including means generating a radiation pattern display for said antenna on said structure.

29. A structure design method as claimed in any one of claims 20 to 28, wherein said indicator determining means includes means for executing a neural network algorithm on 25 the basis of said load parameters which comprise load attributes.

A structure design method as claimed in claim 29, wherein said algorithmn is derived from a neural network training process performed on a data set for evaluated structures.

31. A structure design method as claimed in claim 29 or 30, wherein said indicator determining means includes: P:\OPERDBW5265893. 12/5/97 ~-~II 42 means for applying a predetermined threshold to each of said load attributes to generate threshold attributes; means for performing a logic operation on selected ones of said threshold attributes to generate logic outputs; means for applying predetermined weights to said logic outputs to generate weighted outputs; and means for generating said at least one indicator on the basis of selected combinations of said weighted outputs.

32. A structure design method as claimed in claim 31, wherein said logic operation is an AND operation.

33. A structure design method as claimed in claim 29 or 30, wherein said indicator determining means performs a fuzzy logic operation. S. S.

34. A structure design method as claimed in claim 31, wherein said logic operation is a fuzzy AND operation and said indicator determining means includes means for applying a predetermined slope to each of said load attributes. S 20

35. A structure design system comprising: means for selecting a position on a structure for an object; e, e. means for determining load attributes for said structure with said object in said position, where said load attributes represent loads on selected segments of said structure; and 25 means for executing a neural network algorithm on the basis of said load attributes •to obtain an indicator of the soundness of said structure.

36. A structure design system as claimed in claim 35, including means for selecting said object from a plurality of predetermined objects.

37. A structure design system as claimed in claim 35 or 36, wherein said algorithm is derived from a neural network training process performed on a data set for evaluated structures. P.\OPER\DBW\52658.93 12/5/97 I I Y -43-

38. A structure design system as claimed in claims 35, 36 or 37, wherein said position selecting means comprises a graphical user interface of said system, which displays said indicator.

39. A structure design method substantially as hereinbefore described with reference to the accompanying drawings. A structure design system substantially as hereinbefore described with reference to the accompanying drawings. e Ce *e DAVIES COLLISON CAVE 1 DATED this 12th day of May, 1997 TELSTRA CORPORATION LIMITED By its Patent Attorneys DAVIES COLLISON CAVE r 'S P:\OP1R\DBW\52658.93- 12/5/91 I