Professional Documents
Culture Documents
23•05
ACCEPTANCE TESTING
of Digital Control Systems for L. UYTTERSPROT (Belgium)
L. GRAINE (France)
23 • 05
H. TIMMERMAN (The Netherlands)
G. SCANU (Italy)
© Alstom (France)
unit to unit, e.g. for control and protec-
tion cubicles, a bay simulator and a test
set for analogue signals. The other units
– mimic board, remote control cubicle,
etc. – require elementary test means
capable to read or impose the exchanged
information. Of course, the link between
23 • 05
the remote control cubicle and the dis-
patching centre requires other test
means.
● Provision for progress manage- conditions shall be executed to obtain internal data version, system configura-
ment, remedy management and higher quality assurance and reliability. tion, swap file area and process priority,
version management shall be made These tests shall be carried out in accor- etc.
in the test specification. dance with test specifications. If any test
● Initial state of the DCS shall be item differs from the test specifications,
stated for each test item. the situation shall be documented. 5.4 Site acceptance test -
● Environmental requirements for The method for processing problems New control system
the tests. or questions relating to a test item shall be
23 • 05
● Spot tests and new or existing mentioned in the test specification. If soft- The purpose of the site test is not to
type-tests. ware changes or modifications are neces- repeat factory acceptance test but to con-
sary, the issue shall be clearly understood trol the operation of the DCS in reality, i.e.
➤ Test environment by the test team and the related design files connected to all equipment such as pri-
To create a test environment for the shall be checked by designated people. The mary equipment, protective relays, tele-
DCS as close as possible to the real sub- tests after software changes or modifica- control.
station, it is preferable to assemble opera- tions shall cover all items related with the The test items shall be focused on
tor console, mimic board and telecontrol modified item. Analysis of the results shall power system operation and application,
equipment at the factory even if they are include measured degree of performance, not on software design or manufacturing.
under separate contracts. If this is not pos- the decision whether accepted or not, and At this stage software is considered as a
sible, for the telecontrol equipment for will mention tester name, approver name black box.
example, we recommend to use a simu- and date. It is recommended for future sys- It is important to have a test document
lator instead. tem development to record for each item specifying the following items:
System configuration, equipment and the test process details such as period, ➤ organisation
tools which will be used for the acceptance work-days, problems encountered, soft- ➤ test procedure and environment
test are specified. ware modification, etc. ➤ test items and schedule
Test review will take place at the end
5.3.3.Organisation of each phase, i.e. after combined tests, 5.4.1 Organisation
Test execution and test management total system tests and user’s witness tests, The responsibility of the project leader
organisations are described in a document if applicable. This test review will be placed concerns the test document, the organisa-
mentioning the distribution of the work under the responsibility of the nominated tion of the test team, the execution of the
between user and manufacturer, the per- person. The test review will be part of the tests and the transfer of the tested system
son who is responsible for the factory acceptance test report. to the user. The test team includes experts
acceptance tests and the person from the in DCS, in protections, in primary equip-
user’s side who has authority for approval. 5.3.6 Management ment, in telecontrol. This test team may if
The management plan concerns safety needed be strengthened with experts from
5.3.4 Schedule during factory acceptance tests, test the manufacturer, depending on the test
It is necessary to consider the follow- progress, problem processing and system team’s expertise and on the results of fac-
ing critical periods in the schedule : environment. It will also organise the tory acceptance test. Another possibility is
➤ Possible periods for data amend- meetings during test periods. Safety man- to call the manufacturer for particular
ments and software changes that may be agement covers safety at the factory related tests. The test team execute the tests
requested by the user. to the use of temporary equipment. according to the test document.
➤ Possible periods for software Progress management monitors and con-
improvements following faults detected trols actual progress compared to planned. 5.4.2 Test procedure and environ-
during the tests. Problem management determines the ment
➤ Periods for analysis and assessment methods for problem documenting and The tests are executed on site, i.e. in
of the acceptance tests. processing. the substation. The test procedure depends
It also addresses the problems of ques- on the status of the substation. In case of
5.3.5 Test process and review tionable specifications or changes of spec- a replacement of an existing control sys-
Tests shall be executed by using ade- ifications which arise during tests. tem, the switching of the primary equip-
quate tools to verify elementary and com- System environmental management ment will be subject to authorisation by
pound functions in short period effi- deals with history of change for hardware the grid operator. Similarly, precautions
ciently. Continuous working tests, and software platform. Software platform must be taken when executing tests involv-
combined tests and tests outside normal covers application program, facility data, ing telecontrol.
The team executes the tests accord- between primary installation, DCS and The problem is different if the addition is
ing to the test document, and tests results telecontrol. made by the user. This requires that the user
are recorded in a test file. Faults have to be Even in case of a spare bay some re- should have the complete hardware and
repaired as soon as possible and the items engineering of the DCS may be necessary software description of the DCS in order to
tested again. due to evolution or modification of the evaluate the consequences of the addition
When all items are successfully tested, specifications. of the bay. In this case, he must use his own
the user will take over the new installation reference DCS.
and put it in service. This last step con- 5.5.3 Newly added bay Another aspect requires close atten-
23 • 05
cludes the work of the test team. In this case nothing has been foreseen tion : it is necessary to verify that the design
for the additional bay in the DCS – no constraints of the DCS - e.g. maximum
5.4.3 Test items and schedule engineering, no equipment, no test. The number of bays, data - are still met when
The test document contains a list of addition of a bay requires additional hard- the new bay is added.
test items. As factory acceptance tests have ware, adaptation of the existing software
already been performed it is not necessary and extension of the database.
to repeat all those tests on site. Random Furthermore, this addition has also 5.6 Modification, addition
tests are sufficient. However, it is very repercussions on substation level compo- of a function
important to perform all test items related nents and functions of the DCS – e.g. MMI,
to the working of the DCS in its real envi- telecontrol interface – and even on bay level Different cases arise : bay-related func-
ronment, i.e. connected to the H.V. equip- functions of existing bays – e.g. substation tion or substation-related function, new
ment, telecontrol, … The second objective interlocking for which additional infor- function or activation of an existing func-
of the site tests is to control the coherence mation has to be exchanged between bays. tion in the software library. The methods
between primary and secondary systems The addition of functions will be studied of testing will depend on the actual case.
and telecontrol. further. Which tests should be done ? The In case of an entirely new function,
answer to this question depends on the con- tests have to be carried out on a test sys-
sequences for the DCS of the addition of tem. This test system must be identical to
5.5 Addition of a bay the new bay. If the database alone has been the existing DCS, this means same hard-
5.5.1 Overview modified, site tests are sufficient because ware and same software. For bay-related
Two cases arise : an additional bay has the existing bays remain unchanged. Con- functions the test system is composed of a
already been planned for in the DCS, that versely, in case of software adaptations, the bay unit and a central unit. For substation-
bay being an already installed spare one, entire DCS should be tested. This is obvi- related functions the test system consists
or in the other case the building of a new ously impossible with the existing DCS and of a central unit and a number of bay
bay is decided. In the first case, we suppose the only option left is that of performing units. After these tests, the new functions
that the engineering for the (spare) bay has factory acceptance tests on a reference DCS. can be implemented in the real
already been completed, all equipment is
installed and all tests have been done.
© BEL Engineering (Belgium)
DCS, where, depending on the local pos- the existing hardware, software and Inputs/Outputs of the simulator is pro-
sibilities, some control tests can be database. It is the manufacturer’s respon- portional to the number of bays. Techni-
repeated. sibility to manage the various hardware, cally, tests at the substation level are better
In case of activation of an existing system software and application software than tests at the bay level because the DCS
function from the software library, we can versions. will be tested in working conditions simi-
start from the hypothesis that the function What about tests ? From the definition lar to the reality. This is particularly impor-
has already been tested during the factory of system software and application soft- tant for avalanche testing (representing
acceptance tests or, at least that the man- ware, it follows that a new release affects enormous status changes) and for testing
23 • 05
ufacturer has completely tested the func- the whole DCS. This entails that new tests functions common to several bays, for
tion on a similar DCS. It results that com- have to be carried out on a test system example interlocking, automatism… Also,
plete tests on a reference DCS are no identical to the existing DCS. the cost, the size, the complexity of con-
longer necessary and the new function can necting the simulator increases propor-
be directly implemented in the real DCS, tionally to the number of Inputs/Outputs.
for which only some control tests are 5.8 Modification, addition The problems related to configuration of
needed. The same reasoning applies to the of a data (database/wiring) the simulator to make it consistent with the
modification of the configuration of an H.V. installation must not be overlooked.
existing function. Problems encountered with modifi- These different drawbacks have resulted in
cation/addition of a data are generally substation simulators not being used much
related to the wiring. in the case of conventional control systems.
5.7 Software update When the wiring is changed a test has
to be performed to check the modification Question n° 2 :
There are three kinds of software in a from one point before it to one point after Simulation of the primary equipment
DCS : it. or only reading and imposing inputs/out-
➤ system software, including firm- When only the database is changed, puts? This choice determines the func-
ware keyboarding should be checked, com- tionality included in the simulator. In the
➤ application software pleted by a software test or a site test if the second case - reading and imposing
➤ database. data is used for a function. inputs/outputs - there is no functionality
in the simulator. The simulation of the pri-
The system software is the core of the mary equipment requires functionality
hardware (computer systems and periph- such as switch on and switch off with cor-
eral equipment). The basic functionality 6. Implications on test responding feedback signals and realistic
of the DCS is roughly dictated by the hard- equipment response time, energy monitoring, simu-
ware and the system software (real-time lation of abnormalities
processing, database management, I/O Having seen what tests to do in dif-
processing). ferent cases – factory and site acceptance Question n° 3 :
The application software is an extra tests, new DCS or modification of an exist- Necessity of sophisticated tests like
shell around the core. This software allows ing DCS – we can now analyse the required measurement of response time of the
the DCS to execute the functions described test equipment. DCS, simulation of a high quantity of
in item 2. It contains a library of functions events, …?
which can be configured for each specific The following questions will help us This choice may influence the tech-
application. This application software to select the most appropriate test equip- nology used to build the simulator : it is
mostly contains user-friendly language ment for each application. easier to generate an avalanche test with
allowing the user to develop specific func- a digital simulator than with a relay based
tions, logic such as automatic switching Some answers depend on the experi- simulator.
sequences, interlocking… ence of the user with DCS, others on the
Finally, the database is the result of the terms of the contract between the user and Question n° 4 :
configuration of the DCS for a specific the Manufacturer. What do we test? (See figure 1)
substation. ➤ DCS only
Software updates apply to the system Question no. 1 : ➤ DCS and remote control
software as well as to application software. Test at substation level or at bay level? ➤ all secondary equipment in the sub-
A new release of system software of appli- This choice determines the size of the station (except protection relay, auxilia-
cation software must be compatible with simulator because the number of ries)
23 • 05
imposing inputs
We recommend to analyse carefully DCS and remote same as “DCS only” same as “DCS only”
the problem of overall testing of DCS from control
the very start of engineering because it has Secondary equipment At the interface between primary Simulation of primary equipment
impacts on engineering, on test equip- and secondary equipment Or only reading outputs and
ment, on project planning, on the terms imposing inputs
of the contract between the Manufacturer Secondary equipment same as “Secondary equipment” same as “Secondary equipment”
and the user. and remote control