An automated monitoring system for surveillance

3 downloads 0 Views 2MB Size Report
building, and the energy expert should manually obtain them. For this reason, within .... A Kamstrup Multical 602 is an all-purpose energy calculator for heat ...
An automated monitoring system for surveillance and KPI calculation Álvaro Corredera, Andrés Macía, Roberto Sanz, José L. Hernández Fundación CARTIF, Energy Division Boecillo (Valladolid), Spain [alvcor, andmac, robsan, josher]@cartif.es

Abstract— Monitoring is one of the key points in any retrofitting project because, thanks to its results, both the current and improved status can be analysed. Within this context, not only monitoring is important, but also consistency and coherency of data, as well as aggregated information is required. In that sense, surveillance methods and Key Performance Indicators (KPIs) play a pivotal role. Thus, the project iNSPiRe develops an automated monitoring system able to store daily information, as well as automatically calculate KPIs by means of database technologies and determines alarms at different levels. With the obtained results, iNSPiRE aims at conceiving, developing and demonstrating Systemic Renovation Packages, through the innovative integration of envelope technologies, energy generation (including Renewable energy systems (RES) integration), energy distribution, lighting and comfort management systems into deep energy renovation of buildings, both in the residential and tertiary sectors. Then, the combination of the automated monitoring framework and the renovation packages, the goal is to achieve energy consumption for the building lower than 50 kWh/m²/year. Keywords—Monitoring system; Key Performance Indicator; stored procedures; surveillance; iNSPiRe FP7

I.

INTRODUCTION

Nowadays, the energy efficiency is one of the key points due to the climate change. In this way, the European Commission is aware of this concerns and a new directive has been launched with the aim of reducing the energy consumption out of 20% before 2020 [1]. On the other hand, building stock in Europe is the main contributor for the energy use [2] with amounts up to 70% of electricity consumption and 40% of the total energy. Thus, one of the latest decisions from the European Commission has been to drastically reduce the foot-print and CO2 emissions with a new plan. As stated, most of the energy consumption in Europe is due to heating and cooling used for domestic, tertiary and industrial purposes [3]. The major energy production is because of burning fossil fuels which impact primarily to the negative environmental aspects with the higher CO2 emissions. In this way, the aforementioned European directive [1] encourages the substitution of these fuels by the integration of renewable energies to drive systems for heating and cooling. This solution would reduce the greenhouse gas emissions and the dependence on energy import. Aligned with the directive and the new needs around Europe for reaching the 20/20/20 target, the project iNSPiRe [3] aims at “conceiving, developing and demonstrating

Systemic Renovation Packages, through the innovative integration of envelope technologies, energy generation (including RES integration), energy distribution, lighting and comfort management systems into deep energy renovation of buildings, both in the residential and tertiary sectors”. For achieving these goals, multifunctional industrialized renovation kits are developed, manufactured and installed in the demonstrators. This integration would lead to major cumulative energy savings and higher efficiency in order to reach an overall primary energy consumption lower than 50 kWh/m²year.

Nevertheless, a first step before the implementation of the energy efficient solutions is the monitoring [4]. This phase is pivotal so as to understand the status of the buildings and determine the requirements in terms of renovation aspects. Monitoring is one of the most extended diagnosis techniques, but it is important to carry it out in an efficient manner by means of gathering data from those systems that affect the energy consumption and end-user comfort [5]. In this way, iNSPiRe has developed a monitoring system based on available commercial product, but through an energy performance perspective to feasibly measure the energy consumption and comfort parameters at dwelling level [6]. Initial tools, such as FAR ECHO® [7], developed by a project partners, have been widely used for the storage and visualization of data by means of a complete set of services. Apart from this system, additional monitoring platforms have been analysed according to the project requirements [6]. However, all of them present the lack of automation in terms of calculating Key Performance Indicators (KPIs), which give the real performance of the building, and the energy expert should manually obtain them. For this reason, within iNSPiRe, a novel approach has been deployed whose objective is to allow end-user to visualize monitoring data, as well as KPIs, from data stored in databases in an automatic way without the need of the human intervention. The project demonstrates this novel approach with regard to the monitoring system in two demonstrators located in Spain and Germany (i.e. different climates). Then, ad-hoc monitoring systems have been defined for each site where a set of dwellings are strategically selected to measure performance, as depicted in section II. Additionally, a multi-level surveillance service is deployed which automatically generates alarms when communication is lost and/or measurement values are out of range, among others. This service is widely described in section III that also includes high-level surveillance services,

named Fault Detection. Besides, a procedure to automatically calculate a set of KPIs is explained in section IV. These KPIs take the monitoring information for rendering some works that outcome in values which give feedback from the building with respect to the energy performance and help the designers at time of designing the renovation project. Finally, the conclusions and future lines from this study are extracted in section V. II.

DEFINITION OF THE MONITORING SYSTEM

Roughly speaking, a monitoring system is composed of sensors, gateways and data acquisition system. Those components are included in a network, which can have different architectures. Monitoring system architecture specifies the connection between different components of the installation: sensors, actuators and controllers. The main aim of the monitoring system installed and commissioned is to analyse the behaviour of the building in terms of energy as well as the comfort and behaviour of the end-users. In this respect, a thorough study of the building conditions and systems is needed, in order to set the parameters to be monitored [8]. Conceptually, in iNSPiRE project, a multiple level monitoring framework has been defined, as illustrated in Fig. 1 [6]. First of all, the upper level is the case study itself where the hardware is deployed and data loggers are created. These might be locally stored through different options, i.e. local databases in the building or csv files via mail or FTP server. In the middle, the Internet connection is required for assuring communication. Finally, lower level represents data acquisition systems which gather the information from the case study, parse data and store them into the server database.

Fig. 1. Monitoring scheme

Apart from the information from sensors, static data such as building geometry (air volume, area, façade) are required for the calculation of some KPIs [6][9] . Another adjacent feature related to the monitoring system is the recording frequency, selected in this case to 15 minutes. This frequency is the basis for the further indicators’ calculation which is determined in different frequencies (yearly, monthly, daily or time series).

After the identification of the monitoring requirements and considering buildings and facilities, the place for the meters is determined. The selection of the location has followed a methodology based on multiple criteria: applicability of the measurement; needs of local visualization, maintenance and optimization of sensor position, among others. Having the location in mind, diverse wired and wireless devices and various network topologies have been determined with the aim of reducing the complexity of cabling, assuring speed of transmission, decreasing vulnerability and minimise the rate of errors, such as illustrated in Fig. 2 [6]. These main topologies are centralized, decentralized, distributed and hybrid/mixed, although, iNSPiRE follows the last one.

Fig. 2. Monitoring topology schema

Additionally, the partitioning of the energy fluxes within the building allows an accurate distinction of the energy carriers, the systems and the different loads inside the building. The delivered energy carriers include in general fossil, district and electrical energies (gas, oil, biomass, district heat, district cooling, and electricity) as well as environmental energy (soil, ground water, extract air, waste heat and solar radiation). The consumed energy inside the building represents the loads occurring in the zones including the circuits, i.e. heating, domestic hot water, cooling, lighting and other electrical uses. The systems provide the energy need and - if necessary include a transformation process. The media to be measured define the type of the required sensors, i.e. calorimetric counters (heat/cold meter) or electrical meter. A. Variables to be monitored A standard set of monitoring devices for the two demo sites has been selected in order to speed up the design and the installation phase [6]. Indoor comfort, weather conditions and thermal/electrical consumptions have to be assessed in order to produce weekly and monthly reports for functional areas, to schedule an action plan and optimize with a fine-tuning the control systems of the thermal/electric existing plants.  Indoor comfort

Eltaco FCO2TF63 series is a wireless indoor sensor which measures CO2, relative humidity and temperature. A wireless technology has been selected in order to reduce as much as possible the interference for the occupants due to the installation. For this reason, a wireless base station is needed so as to get signals from the comfort sensors, which are EnOcean

[10] compliant, and transfer it to a gateway via LonWorks protocol [11]. A wireless transceiver has been selected for this purpose. In the residential demo case, one receiver per dwelling has been installed in order to acquire data from a set of indoor comfort sensors. In the office building, one receiver was sufficient.  Thermal and electricity consumption

The thermal consumptions are due to heating loads and domestic hot water preparation purposes. The thermal energy flows monitoring requires a heat meter installed along the pipelines. A Kamstrup Multical 602 is an all-purpose energy calculator for heat together with a ultrasonic sensor and with 2 wired temperature sensors. This meter saves energy data on a yearly, monthly, daily and hourly basis, which provides the operations manager with a complete performance analysis. The meter can be fitted with LonWorks [10] communication. The Countis series electricity meter by Socomec is the modular active energy meter selected for the three demo cases. The pulse output is gathered by a LonWorks [10] pulse memory module that redirects the information into the LonWorksnetwork.

B. Database definition The last point is the definition of the database for persistently storing data, making them available in further analysis. Fig. 4 illustrates a simplified picture of the database schema where the main entities are depicted. Firstly, sensor table contains the list of sensors installed into the demonstrator, while sensortype is the list of types of sensors (e.g. temperature). Then, the monitoring data is gathered within the sensorhistdatalog table that is related to the sensor tables (i.e. one data sample is always measured by a sensor). Similar to this table, sensorlastdatalog is filled though a trigger for keeping a copy of the last data, which speeds up visualization because the number of registers are reduced. Additionally, the strategy for the KPIs is identical than sensors, although, in this case, these values are obtained by means of automatic triggers and stored procedures, as it will be described below.

 Weather conditions

Outdoor temperature, relative humidity and direct and diffuse solar radiation are assessed through a Warema weather station. A Warema sensor is used to record temperature and humidity, while two Deltaohm pyrometers are used to gather diffuse and total solar radiation. The weather station is connected to a Warema Lonse III (LonWorks [11] sensor unit), enabling the integration into the LonWorks-network.

Having the sensors selected, the next step is the design of the monitoring network which follows the schema represented in Fig. 3 [6]. In this way, a methodology where the ways to deliver energy and the renewable energies are split, as well as the demand side (energy needs). Once the energy sources are identified, then, the distribution network is connected by lines. In the example, gas boiler feeds the heating plant in order to give heating output. Afterwards, bearing in mind the energy distribution systems, the sensors are established in the point of interest, as happens with the gas meter. In this manner, the sensor network is easily designed.

Fig. 4. Database schema

Last but not least, the fields name and name_lon are the ones used for the automatic storage of data. One Java process runs periodically to retrieve data in one of the aforementioned ways every day. Then, taking into account that the data logger gathers data in the LonWorks format [11], name_lon contains the LonWorks name of the data-point. Thus, the process maps the variable from the database and the network itself so as to store the information in the histdatalog table. The other field, name, is used for having a human-readable name in the visualization platform. III.

Fig. 3. Monitoring network desgin

ALARMS AND SURVEILLANCE

In this section, the alarms to supervise the operation of the systems and highlight the performance and inefficiencies of the building are described. Four different alarm levels have been defined [12], depending on the components launching the error signal, as well as how the fault affects the other components of the system. To clarify this organization, the alarm levels are the following: level 0 includes all the faults in the communication system; level 1 includes the faults of the sensors; level 2 includes the faults of complex equipment (machines, e.g. the heat pump) and, in general, all the equipment that is not performing only the function of a sensor; and level 3 includes system faults, all the abnormal situation caused by a bad

interaction between the equipment that makes up the system [12].

In our architecture, the alarm is generated by the alarm generator in the system components (if possible) or by the alarm detector in the system supervisor. The alarm is then sent to the FAR-Echo system via the iLON server. The FAR-Echo system will process the fault, classifying, prioritizing and recording it. Eventually, it takes actions like displaying them on screen or sending emails to registered users, as shown in Fig. 4 [12].

temperatures, mass flow, electric consumption and thermal consumption. With these alarms, the plant manager or operator will know when the equipment is working under performance or where there will be dimension problem, management and identify inefficiencies. Table I shows an example of the level 2 alarms implemented on a solar thermal system, with the mathematical rules to trigger each alarm. Indicator Primary flow Heat exchanger

Fig. 5. Alarm management

A. Level 0: Communication Faults Level 0 aims to collect any problem in communication between the monitoring and data storage systems, as well as in the web platform / display system. If any error is detected, the system shows on screen a message with a description of the problem, e.g. “Warning: ftp server connection dropped” and notifies via email the people in charge of solving the specific issue. Upon solution of the problem, the alarm is removed from the screen and the alarm logger update. Any other problem regarding the communication of data between systems, as for instance a problem accessing to the control network or missing data in a downloaded data file, fall in this category and should be recognized by the alarm generator system.

B. Level 1: Sensors fault The alarm generator/detector classifies as level 1 all the malfunctions in the measurements of the physical signals. In general this is achieved by defining upper and lower thresholds to each measurement and using them to define one or more validation rule for the data point. When the data readings of the given sensors are outside the limits, the alarm is triggered. At this level an email is sent to the people responsible for network system maintenance

C. Level 2: Equipment faults The alarms implemented at this level have the objective of reporting anomalous conditions regarding the equipment such as pumps, valves, heat pump and/or energy storage systems. Often such conditions are anomalous operating conditions which may cause either damage to the component or waste energy. In such case it is necessary to restore normal operative conditions after such an alarm is raised. The alarms are usually generated when the equipment is working outside the operative range with respect of one or more variables such as

TABLE I. LEVEL 2 SOLAR THERMAL FAULT DETECTION Equation

ŋ ŋ

ŋ

Rule 0.95