Failure risk assessment in water network in terms of planning renewals – a case study of the exemplary water supply system

Task of water supply system (WSS) is to provide consumers with required amount of water having proper quality, necessary pressure, according to the valid standards, price acceptable by consumers and at time convenient for them. At present the main problem of water supply is not only water shortage in the developing regions but also the operation of existing old water network that often is in a bad technical condition. Functioning of an old water network with pipe age from 50 to even 100 years is characterized by the high failure frequency. The proper operation of urban WSS should take into account the minimization of water losses, as well as issues connected with system operation and safety reliability (Nowacka et al. 2016).

The important problem which occurs in many urban water-pipe networks are also considerably oversized diameters in water-pipe network which results in decreasing water flow velocity and as a consequence causes the deterioration of water quality in water network (Friedl & Fuchs-Hanusch 2012; Kabir et al. 2015).

A crucial role in the procedures of the failure analysis and prediction plays the gathering of proper failure records, as well as opinions and assessments of experts (e.g. operators) and specialists (e.g. researchers).

Being consumers of drinking water we deal with the issue of WSS safety every day. There are many studies on WSS reliability and safety (Kleiner et al. 2001; Kutylowska & Orlowska-Szostak 2016; Pietrucha-Urbanik 2016). The division of water consumers into five categories was established (Kwietniewski et al. 1993), which takes into account the required quantity of water supply Q_w compared to the total nominal water demand Q_n. The solution presented in (Twort et al. 2000) takes into account the technical and social aspects of the required level of reliability of the system and subsystems of rural water supply, as defined by the minimum failure rate obtained on the basis of operational data (Ondrejka Harbulakova & Markovic 2015).

Task of WSS is to ensure the continuity of water supply to consumers including such criteria as system reliability and socially acceptable level of water price, taking into account the requirements of public safety, protection of the natural water environment and quality life standard (Rak 2009; Kaźmierczak & Wdowikowski 2016).

The result of the water supply network failure can be a lack or reduction of water supply for a given number of consumers or decrease in the pressure below the required, especially for those who live in the upper floors of buildings. Some examples of major water pipe breaks influencing many recipients in the recent years were:

• - the water main break in 2014 at the campus of the University of California, which caused flood, a five meter hole in the street and a nine meter water column, 30 thousands m³ of water flowed from the 93 year old steel pipe,

• - fourteen major water line breaks occurred in August 2014 in Texas, caused the water towers to be emptied, due to high temperature and age of water pipes, which were up to 100 years old and reached the end of their lifetime.

The indicators of reliability of physico-chemical quality of drinking water and the concept of tolerable risk were introduced in the thesis (Rak 2009). WSS has its own characteristics, its various subsystems perform different functions and interacting with each other form an integral whole (ISO 24512:2007). Standards of the World Health Organization (WHO) concerning the sanitary quality of drinking water have been applicable since the 1950s. In 1958 the WHO presented the first publication on that subject entitled ‘International standards for drinking water’. According to the standards given by the WHO drinking water should not only fulfill the consumers safety criteria but also meet the organoleptic requirements, however, from a technical requirements point of view, water should be supplied to consumers in a suitable amount and with a suitable pressure in accordance with the applicable standards. In 2004, the third edition of the guidelines for drinking water quality (Guidelines for Drinking-Water Quality) published by the WHO provided guidance for the development of the so-called Water Safety Plan, which aim is to obtain the requirements concerning critical infrastructure protection. The water safety plans should also be included in the standards concerning water supply for customers. In the mentioned standards the failure risk analysis of water network is recommended to be performed in order to fulfill safety and reliability requirements for WSS.

The procedures of correct WSS designing, constructing and operating should be performed on the basis of the detailed failure analysis (Tchorzewska-Cieslak et al. 2012; Pietrucha-Urbanik 2015).

The failure risk analysis of water supply network should be performed in the following stages (Kleiner et al. 2001; Tabesh et al. 2009; Wang & Chen 2016):

• - determining the type of WSS,

• - determining the critical value of the failure rate of the water supply network,

• - specifying the type of safety associated with the operation of the water supply network,

• - determining the critical value of risk levels.

The proposed analysis can provide the basis for comprehensive risk management of failure analysis implemented in the water safety plans and also in decision-making processes. The aim of this paper is to propose a new scheme for operational assessment related to technical functioning of the system, on the example of the examined WSS from seven year observation period.

The WSS operation is characterized by the changeable pressure and flow parameters, which change amount of water used by consumers (Kleiner et al. 2001; Vloerbergh & Blokker 2010; Garcia-Mora et al. 2015).

Water network consists of mains, distributional pipe and water supply connections together with the particular fittings as, for example, check valves, hydrants and flow meters.

Failures which occur in water-pipe network and fittings have random character and can be caused by the events connected with groundwork, water-pipe technical state, errors at mounting or sudden temperature changes. Such situations cause the difficulty in performing the analysis.

During the WSS functioning various failures can occur causing water losses and they can be a reason for the secondary contamination of water in water network, which is a serious threat to consumers safety.

Such situations very often cause high failure frequency in water network, through, according to (Tabesh & Saber 2012; Ondrejka Harbulakova et al. 2015):

• - incorrectly assumed concept of network structure (network in open or mixed system),

• - wrongly chosen network operating hydraulic conditions,

• - too high working pressure,

• - lack of cut-off and control fittings protecting against water hammer.

Frequently the failures in the water network concern:

• - pipes, e.g. cracks, corrosion,

• - connections, e.g. leaks,

• - fittings, e.g. damage of hydrants.

• - pressure measurement in water-pipe network, water-pipe network fittings inspection (maintenance or removal), expansion of water network and construction of new connections, water-pipe network failure repairs,

• - pipes renovation (inside surface is covered with cement mortar or epoxide mortar, flexible lining),

• - pipes reconstruction (pipes relining, compact pipe (U-liner), plastic pipes are put into pipe being repaired),

• - pipes renewal (using the trench and trenchless method, removing or leaving old pipe).

The significant problem concerning subsystem operation is gathering and archiving statistical data of failures. To perform this activity the developed database and computer systems (e.g. SCADA) should be used. Data of failures should contain information about failure date, detailed data of water-pipe network type (main, distributional, water connections), pipe material, diameter and age, working pressure, as well as the data concerning water-pipe network localization, ground conditions, depth of foundation, time of failure repair, the possible cause of failures and their consequences (Berardi et al. 2014).

Reliability analysis of the WSS treated as an integral whole is, in practice, very difficult because it requires a lot of detailed information concerning particular subsystems and their elements, as well as an analysis of failure dependence (sometimes failure in one subsystem can cause failure of the next one, the so called domino effect), so the particular subsystems are often analyzed separately and even the elements of the particular subsystems are analyzed separately (Rak 2009; Gheisi & Naser 2014; Taeho et al. 2014). The water-pipe network working life is not always a cause of high failure frequency of the network, there are some examples of water-pipe networks made of grey cast iron having approximately 80 or even 100 years of working life which are in good technical condition but there are also much younger networks which should be already renewed. Therefore the failure rate of each segment should be taken into account (Studziński & Pietrucha-Urbanik 2015).

The failure analysis is inseparably connected with the maintenance of network operational reliability and helps to provide consumers with proper quality water.

The examined WSS was built in the mid-seventies of the twentieth century. It is placed in the eastern Poland in the city inhabited by 20 thousand people. The water supply network is developed in the whole territory of the municipality and about 95% of the total population benefit from WSS, the rest of the population uses water from their own household wells.

Water for the residents of the municipality is derived from six intakes (drilled wells located in the municipality), the commune has three water treatment plants, additionally there is an emergency intake with a distinct water treatment station. The WSS is composed of the main section of the distribution network of 1,500 m (Ø500), 3 rings of the distributional network, straight sections of the distribution network, a separate distribution network for one of the intakes and the numerous water supply connection.

The length of the distribution network is almost 190 kilometers increased by an average of about 2.7 km per year. Length of water supply connections in 2008 was 179.52 km and within ten years increased by 25.7 kilometers, the average annual increase in the length of water supply connections was about 4.5 km.

The failure rates used in the functioning analysis of the WSS are:

• - the failure rate λ(t) [number of failures·year (day)⁻¹] or [number of failures·a⁻¹·km⁻¹]. This indicator is used in the analysis and assessment of the WSS failure and calculated as the average value of the damage intensity of pipes, connectors and fittings. It is calculated as the total number of failures in the time interval by the number of analyzed elements or for linear elements their length L [km] and time of observation:

  

  2

  where n(t, t + Δt) is the total number of failures in the time interval Δt and L is the length of water network.
• - Mean Time Between Failures MTBF [d], which is the expected value defining operating time (ability of the system (or its components)) between two consecutive failures:

  

  3
• ­ Mean Time To Repair MTTR [h] describes the value of time from the moment of failure until re-enable water flow in the damaged section of the water supply network:

  

  4

  where T_et is an expected time for repair [h] and T_r is a real time of repair [h].
• ­ the repair rate μ(t) [number of repairs·a(h)⁻¹] determines the number of failures repaired per time unit, it can be determined as the reverse of the mean repair time:

  

  5
• ­ the average availability indicator R(t):

  

  6
• ­ Average Water Not Supplied AWNS [m³·d^–1·customer^–1]:

  

  7

Presented indicators can be helpful in description of proper WSS functioning for both the operator and customer in terms of its reliability and safety.

In order to determine what kind of failure has the great impact on the failure rate in Table 1 the number of failures depending on the cause and the failure rate were presented.

In case of unsealing, the great influence has the precision of pipelines connections (e.g. the appropriate welding of PE pipes, the proper implementation of the weld connection in the steel pipes) and connections of these pipelines with water supply fittings. Mechanical damages are usually caused by ground works near lying pipes, while the longitudinal and transverse cracking is caused by a wrong foundation of the pipeline or by its placement along the busy streets, also by inappropriate pressure. Failure caused by freezing is the most common cause of damage occurring during the winter. Very low temperatures and lack of required minimum pipe depth, especially where water flow is relatively small (in the end of the network, water connections), cause the damage due to freezing of water in the pipes.

Most failures occurred during the summer, the cause of such situation is that during the holiday season water demand increases and risk of pipe failure is higher. Despite recommendations for reducing tap water consumption in summer (due to lower groundwater levels), during this period the residents consume too much water for watering greenery and washing cars. In January the number of failures was lower than the average by 40.7%, in February by 47.3%, also in September the number of failures was significantly different from the average value (by 47.3%) The large increase compared to the average value occurred in August, 197.8%. The analyzed data indicate that the WSS is characterized by high seasonality, the number of failures increases in March and decreases in September.

The median value of the indicator of water not delivered to customers during the year to the number of customers connected to the water supply for the examined period of time was 0.85 m³ d^–1·customer^–1 and is within the range specified for other comparable systems of water supply, from 0.53 to 0.91 m³ d^–1·customer^–1.

In case of the distribution network the repair time depends on the pipe diameter in which the failure occurs, such diameters are larger than the water supply connections diameters, what usually causes the use of heavy equipment. The number of distributional failures lasting more than one working day is less than in the case of water supply connections, usually MTTR equals 8 h and the repair rate μ = 0.1251·h⁻¹.

The analysis shows that the median value of the availability indicator R(t)_med for distributional network is 0.962477 and for water supply connections 0.958814.

From the beginning of the observation period the decreasing trend of the water supply availability occurs, which in case of this indicator shows negative situation in the WSS. To reduce the failure indicators the investment and modernization projects should be performed through replacement, development and modernization of the existing WSS.

The detailed analysis of failures showed that some interventions did not require closing the water pipe, which in consequence did not result in a lack or limited water supply to consumers.

The precise definition of operating states of the WSS has a significant impact on the analysis of reliability and safety of the system. A new scheme is proposed, which defines the operating conditions in WSS (Table 2). The proposed criteria and categories are developed on the basis of waterworks practice and failure analysis performed in different water supply systems.

In case of the tolerable category, the system can be operated under no special conditions, the operators should perform inspections and conduct failure analysis. In comparison, when the unacceptable category of WSS occurs, the system should not be operated and immediate action should be initiated. The availability indicator R(t) in the last year of observation for the examined water supply in case of the distributional pipes equals 0.95824598, for the water connections R(t) = 0.95744681 and for the water network fittings R(t) = 0.958146731, each of these values respond to the controlled criteria proposed in Table 2 (for the water network supplying more than 2,000 and less than 200,000 recipients). In this case the system operator should take some action to modernize the pipes having the highest failure occurrence. The modernization of water supply network will significantly increase its reliability reducing the cost of continuous repairs and will adapt the WSS to new tasks as to cover the increased demand for water in continuously developing areas.

For the risk assessment, the indicators of the probability of undesirable events occurrence and the indicators related to the time duration of the individual operating states, should be used.

In the presented example of the failure analysis for the city in eastern Poland, based on the operation data, the number of failures in the water-pipe network, depending on the type of failure, in the successive years, has been analyzed and it has been found that the most often failures occurred in the water supply connections. The most common cause of failures in the water supply connections was cracking, with the failure rate at the level of 0.131 number of failures·a⁻¹ km⁻¹. The calculated mean failure rates for the particular type of network compared with the criteria values do not exceed the recommended values presented in Table 1. However, it is suggested to perform detailed analysis of the particular segments of the water-pipe network in order to modernize them. It is also worth to note the considerably high failure frequency in case of the fittings (including fire-protective hydrants), which suggests performing the inspections and possible replacement.

The WSS is characterized by a continuous operation, its reliable and safe operation has a direct impact on the quality of water consumers lives.

Analysis of water network reliability, as well as a precise database of operation information of the system have a significant impact on the correctness of performed analysis and the final result of the reliability analysis. Therefore to perform the proper analysis the failure database consisting of the failure protocol should be developed.

Failures in the WSS do not occur without a cause, often appear in a chain of undesirable events, they are also a result of making wrong decisions and poor management, resulting in a negative impact on the WSS operation. To improve the present situation of the WSS functioning and reduce or regulate the pressure in the water supply network, monitoring of the water network should be provided. Also modernization of valves in nodes in the distributional pipe and modernization of the existing WSS, including the Active Leak Detection, should be implemented.

Risk and failure analyses in the WSS should be standard during system operation and may be considered a tool to support decisions made in the process of proper management.

The presented issue is intended to draw attention to the need for further improvement and the standardization of criteria for the analysis and risk assessment in the WSS.